Front cover

FICON Planning and Implementation Guide

Topologies, concepts, and terminology

Planning, implementation, and migration guidance Realistic examples and scenarios

Bill White Wolfgang Fries Brian Hatfield Michal Holenia Dennis Ng Ewerson Palacio Ren Petry

ibm.com/redbooks

International Technical Support Organization FICON Planning and Implementation Guide September 2009

SG24-6497-02

Note: Before using this information and the product it supports, read the information in Notices on page ix.

Third Edition (September 2009) This edition applies to FICON features defined as CHPID type FC, supporting native FICON, High Performance FICON for System z (zHPF), and FICON Channel-to-Channel (CTC) on IBM System z10 and System z9 servers.

Copyright International Business Machines Corporation 2005, 2006, 2009. All rights reserved. Note to U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.

Contents
Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .x Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi The team who wrote this book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Become a published author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Part 1. Understanding FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Chapter 1. Introduction to FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1 Basic Fibre Channel terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1.1 Node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.2 Port . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.3 Switched fabric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.4 FC link . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.1.5 World Wide Name . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.2 System z FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.2.1 FICON advantages over ESCON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.2.2 FICON operating modes and topologies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.2.3 Terms used with FICON Directors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.2.4 Terms used with the Input/Output architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Chapter 2. System z FICON technical description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Using the FICON architecture for I/O operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 FICON initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 FICON I/O request . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3 Command mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.4 Transport mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.5 Modified Indirect Data Address Word . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.6 Transport Indirect Data Address Word (TIDAW). . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.7 Open exchange. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.8 Buffer-to-buffer credit usage in FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.9 Extended distance FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 System z FICON feature support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chapter 3. FICON Director technical description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 The role of the FICON Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1 Switched point-to-point configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2 Cascaded FICON Director configuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3 Basic components of a FICON Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.4 Basic functions of a FICON Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Qualified FICON Directors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 IBM System Storage SAN b-type family components. . . . . . . . . . . . . . . . . . . . . . 3.2.2 IBM System Storage SAN b-type family functions . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3 Cisco MDS 9500 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4 Functions of the Cisco MDS 9500 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 14 14 16 19 20 22 22 23 23 24 25 33 34 34 35 36 37 44 45 47 52 53

Part 2. Planning the FICON environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

iii

Chapter 4. Planning the FICON environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Structured approach for planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.1 Migrating from ESCON to FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.2 Moving to a high bandwidth environment (FICON Express8) . . . . . . . . . . . . . . . . 4.4.3 Migrating from a single site to a multi-site environment . . . . . . . . . . . . . . . . . . . . 4.4.4 Implementing a new FICON environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5 Topologies and supported distances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.1 Point-to-point. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.2 Switched point-to-point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.3 Cascaded FICON Directors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5.4 Extended distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6 Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.1 Intermix fabric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.2 Fabric security. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.3 High integrity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.4 Zoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7 Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7.1 Command-line interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7.2 Element management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7.3 Fabric management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7.4 Storage management initiative specification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7.5 System z management for FICON Directors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8 Virtualization and availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8.1 System z . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8.2 Control unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8.3 FICON Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9 Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.1 Frame pacing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.2 Extended distance FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.3 Multiple allegiance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.4 Parallel Access Volume and HyperPAV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.5 Modified Indirect Data Address Word . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.6 High Performance FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.7 Bandwidth management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.8 Traffic management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9.9 Evaluation tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.10 Prerequisites and interoperability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.11 Physical connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

59 60 62 63 63 64 65 66 66 67 68 68 68 70 71 72 73 74 74 75 75 76 76 76 77 78 78 81 82 83 84 85 86 86 87 88 89 91 91 92 93

Part 3. Configuring the FICON environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Chapter 5. Configuring a point-to-point topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 5.1 Establishing a point-to-point topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 5.1.1 Description of our environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 5.1.2 Tasks and checklist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 5.1.3 Getting started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Chapter 6. Configuring a switched point-to-point topology . . . . . . . . . . . . . . . . . . . . 6.1 Establishing a switched point-to-point topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.1 Description of our environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1.2 Tasks and checklist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
FICON Planning and Implementation Guide

115 116 116 117

6.1.3 Getting started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Chapter 7. Configuring a cascaded FICON Director topology . . . . . . . . . . . . . . . . . . 7.1 Establishing a cascaded FICON Director topology . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1.1 Description of our environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1.2 Tasks and checklist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1.3 Getting started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chapter 8. Configuring FICON Directors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1 Configuration overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1.1 Configuration flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1.2 FICON Director management connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Installing and using Data Center Fabric Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.1 Installing Data Center Fabric Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.2 Using the DCFM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 Setting up a FICON Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.1 Changing the IP addresses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.2 Enabling features (optional) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.3 Setting up a logical switch (optional). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.4 Configuring the Domain ID and Insistent Domain ID . . . . . . . . . . . . . . . . . . . . . 8.3.5 Setting up PBR, IOD, and DLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.6 Enabling the Control Unit Port . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.7 Changing port type and speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.8 Changing buffer credits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.9 Setting up the Allow/Prohibit Matrix (optional) . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.10 Setting up zoning (optional) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.11 Configuring Port Fencing (optional) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Setting up cascaded FICON Directors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4.1 Setting up Inter-Switch Links . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4.2 Setting up a high integrity fabric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.5 FICON Directors in an extended distance environment . . . . . . . . . . . . . . . . . . . . . . . 8.6 FICON Directors in an intermixed environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.7 Channel swap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.8 Backing up Director configuration data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.9 Backing up DCFM configuration data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 134 134 135 137 151 152 152 153 154 155 161 164 164 169 170 173 175 177 179 183 184 185 188 190 190 192 196 197 197 197 198

Part 4. Managing the FICON environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Chapter 9. Monitoring the FICON environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1 System Activity Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 Resource Measurement Facility monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3 Introduction to performance monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4 Introduction to Resource Measurement Facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4.1 Data gathering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4.2 RMF reporting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5 RMF example reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.1 DASD Activity Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.2 I/O Queueing Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.3 Channel Path Activity Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.4 FICON Director Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.5 ESS Link Statistics Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.6 General performance guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.7 Tape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.6 DCFM performance monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Contents

201 202 203 203 204 205 205 206 206 207 209 210 211 212 213 213 v

9.6.1 End-to-End Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 9.6.2 Real-Time Monitor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 Chapter 10. Debugging FICON problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.1 Preparing for problem determination activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.1.1 Using the D M - Display system Matrix command . . . . . . . . . . . . . . . . . . . . . . 10.1.2 Creating a CONFIG member . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.1.3 Most common I/O-related problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 Problem determination approach for FICON. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3 Environmental Record, Editing, and Printing program . . . . . . . . . . . . . . . . . . . . . . . 10.4 FICON link incident reporting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5 FICON Purge Path Extended . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.6 Helpful z/OS commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.7 Node descriptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.7.1 View node descriptors from DCFM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.8 DCFM logs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.9 DCFM data collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.10 Common z/OS FICON error message - interpretation . . . . . . . . . . . . . . . . . . . . . . 217 218 218 218 219 220 222 222 222 225 226 226 228 229 230

Part 5. Appendixes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 Appendix A. Example: planning workflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 Appendix B. Configuration worksheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 Appendix C. Configuration and definition tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hardware Configuration Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hardware Configuration Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CHPID Mapping Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I/O Configuration Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IOCDS statements and keywords used for FICON. . . . . . . . . . . . . . . . . . . . . . . . . . . . 255 256 256 256 256 257

Appendix D. Configuring the DS8000 for FICON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259 Getting started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 DS8000 port layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 Appendix E. Using HMC and SE for problem determination information . . . . . . . . . HMC and SE versions and user IDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Logging on to the HMC and SE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Displaying System z server CPC details. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Finding a physical resource on a System z server . . . . . . . . . . . . . . . . . . . . . . . . . . . . Displaying individual FICON channel information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Displaying the Analyze Channel Information panels. . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix F. Useful z/OS commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Using z/OS commands for problem determination. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Displaying system status using D M=CPU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Displaying additional z/OS information using D IPLINFO . . . . . . . . . . . . . . . . . . . . . . . Displaying the I/O configuration using D IOS,CONFIG. . . . . . . . . . . . . . . . . . . . . . . . . Displaying HSA usage using D IOS,CONFIG(HSA) . . . . . . . . . . . . . . . . . . . . . . . . . . . Display units command D U . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D U,,,dddd,1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D M=DEV(dddd) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DEVSERV command - DS P,dddd,n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 266 266 266 268 270 272 287 288 288 290 290 291 292 292 296 300

vi

FICON Planning and Implementation Guide

Appendix G. Adding FICON CTC connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305 Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Online resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . How to get Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Help from IBM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 307 307 309 309 309

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311

Contents

vii

viii

FICON Planning and Implementation Guide

Notices
This information was developed for products and services offered in the U.S.A. IBM may not offer the products, services, or features discussed in this document in other countries. Consult your local IBM representative for information on the products and services currently available in your area. Any reference to an IBM product, program, or service is not intended to state or imply that only that IBM product, program, or service may be used. Any functionally equivalent product, program, or service that does not infringe any IBM intellectual property right may be used instead. However, it is the user's responsibility to evaluate and verify the operation of any non-IBM product, program, or service. IBM may have patents or pending patent applications covering subject matter described in this document. The furnishing of this document does not give you any license to these patents. You can send license inquiries, in writing, to: IBM Director of Licensing, IBM Corporation, North Castle Drive, Armonk, NY 10504-1785 U.S.A. The following paragraph does not apply to the United Kingdom or any other country where such provisions are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THIS PUBLICATION "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Some states do not allow disclaimer of express or implied warranties in certain transactions, therefore, this statement may not apply to you. This information could include technical inaccuracies or typographical errors. Changes are periodically made to the information herein; these changes will be incorporated in new editions of the publication. IBM may make improvements and/or changes in the product(s) and/or the program(s) described in this publication at any time without notice. Any references in this information to non-IBM Web sites are provided for convenience only and do not in any manner serve as an endorsement of those Web sites. The materials at those Web sites are not part of the materials for this IBM product and use of those Web sites is at your own risk. IBM may use or distribute any of the information you supply in any way it believes appropriate without incurring any obligation to you. Information concerning non-IBM products was obtained from the suppliers of those products, their published announcements or other publicly available sources. IBM has not tested those products and cannot confirm the accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products. This information contains examples of data and reports used in daily business operations. To illustrate them as completely as possible, the examples include the names of individuals, companies, brands, and products. All of these names are fictitious and any similarity to the names and addresses used by an actual business enterprise is entirely coincidental. COPYRIGHT LICENSE: This information contains sample application programs in source language, which illustrate programming techniques on various operating platforms. You may copy, modify, and distribute these sample programs in any form without payment to IBM, for the purposes of developing, using, marketing or distributing application programs conforming to the application programming interface for the operating platform for which the sample programs are written. These examples have not been thoroughly tested under all conditions. IBM, therefore, cannot guarantee or imply reliability, serviceability, or function of these programs.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

ix

Trademarks
IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business Machines Corporation in the United States, other countries, or both. These and other IBM trademarked terms are marked on their first occurrence in this information with the appropriate symbol ( or ), indicating US registered or common law trademarks owned by IBM at the time this information was published. Such trademarks may also be registered or common law trademarks in other countries. A current list of IBM trademarks is available on the Web at http://www.ibm.com/legal/copytrade.shtml The following terms are trademarks of the International Business Machines Corporation in the United States, other countries, or both:
DB2 DS8000 ESCON FICON GDPS Geographically Dispersed Parallel Sysplex HyperSwap IBM Parallel Sysplex PR/SM Redbooks Redbooks (logo) Resource Link System Storage System z10 System z9 System z Tivoli z/Architecture z/OS z/VM z/VSE z9

The following terms are trademarks of other companies: Disk Magic, and the IntelliMagic logo are trademarks of IntelliMagic BV in the United States, other countries, or both. Snapshot, NOW, and the NetApp logo are trademarks or registered trademarks of NetApp, Inc. in the U.S. and other countries. SAP, and SAP logos are trademarks or registered trademarks of SAP AG in Germany and in several other countries. Java, and all Java-based trademarks are trademarks of Sun Microsystems, Inc. in the United States, other countries, or both. Active Directory, Convergence, Excel, Microsoft, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both. Linux is a trademark of Linus Torvalds in the United States, other countries, or both. Other company, product, or service names may be trademarks or service marks of others.

FICON Planning and Implementation Guide

Preface
This IBM Redbooks publication covers the planning, implementation, and management of IBM System z FICON environments. It discusses the FICON and Fibre Channel architectures, terminology, and supported topologies. The book focuses on the hardware installation and the software definitions that are needed to provide connectivity to FICON environments.You will find configuration examples required to support FICON control units, FICON Channel-to-Channel (FCTC), and FICON Directors. It also discusses utilities and commands that are useful for monitoring and managing the FICON environment. The target audience for this document includes IT Architects, data center planners, SAN administrators, and system programmers who plan for and configure FICON environments. The reader is expected to have a basic understanding of IBM System z10 and IBM System z9 hardware, HCD or IOCP, as well as a broad understanding of the Fibre Channel and FICON architectures.

The team who wrote this book

This book was produced by a team of specialists from around the world working at the International Technical Support Organization, Poughkeepsie Center. Bill White is a Project Leader and Senior Networking and Connectivity Architect at the International Technical Support Organization, Poughkeepsie Center. Wolfgang Fries is a Senior Consultant in the System z Support Center in Germany. He spent several years in the European Support Center in Montpellier, France, providing international hardware support for System z servers. Wolfgang has 32 years of experience in supporting large System z customers. His areas of expertise include System z servers and connectivity. Brian Hatfield is a Certified Consulting Learning Specialist who works for the IBM Systems and Technology Group in Atlanta, Georgia. He has more than 30 years of experience in the IBM mainframe environment and began his career as a Large System Customer Engineer in Southern California. Brian has been in the Education field for the past 17 years, and he currently develops and delivers technical training for the System z environment. Michal Holenia is a Senior Network IT Architect and Integrated Communication Services Leader in IBM Slovakia. He has 11 years of field experience with IBM, including working with Networking and WDM products. He is a Cisco Certified Internetwork Expert, ADVA Certified Engineer, and Certified Information Systems Security Professional. Michal provides design, implementation, and technical support for LAN, WAN, SAN, MAN and Private Optical Network WDM solutions. Dennis Ng is a Consulting IT Specialist with IBM North America. He has 34 years of IBM experience in Large Systems Technical Support. Dennis is currently working at the Washington Systems Center in Gaithersburg, Maryland, focusing on FICON architecture and FICON supported products. His areas of expertise include Systems z hardware and FICON/ESCON connectivity technical support. Ewerson Palacio is an IBM Senior Technical Staff Member (STSM) and a Certified Consulting IT Specialist for Large Systems in Brazil. He has more than 35 years of
Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

xi

experience with IBM Large Systems. His areas of expertise include System z server technical and customer support. Ewerson has been a System z hardware Top Gun course designer, developer, and instructor for the last five generations of IBM high-end servers. Rene Petry is a Support Center Specialist in the System z Support Center in Germany. Since serving a three-year apprenticeship at IBM in Mainz, Rene has worked for nine years as an Account Customer Engineer for a large banking customer in Germany. His areas of expertise include System z servers, FICON, Fibre Channel Directors, IBM Storage products (disk and tape), and fiber optic infrastructures. Thanks to the following people for their contributions to this project: Bob Haimowitz International Technical Support Organization, Poughkeepsie Center Connie Beuselinck IBM System z Product Planning, Poughkeepsie Charlie Hubert, Brian Jacobs, Sam Mercier IBM Vendor Solutions Connectivity (VSC) Lab, Poughkeepsie Lou Ricci IBM Systems Software Development, Poughkeepsie Jack Consoli Systems Engineer, Brocade Communications Systems, Inc. Thanks to the authors of the previous editions of this book. Authors of the first edition, FICON Implementation Guide, published in February 2005, were: Hans-Peter Eckam, IBM Germany Iain Neville, IBM UK Authors of the second edition, FICON Implementation Guide, published in January 2006, were: Hans-Peter Eckam, IBM Germany Wolfgang Fries, IBM Germany Iain Neville, IBM UK

Become a published author

Join us for a two- to six-week residency program! Help write a book dealing with specific products or solutions, while getting hands-on experience with leading-edge technologies. You will have the opportunity to team with IBM technical professionals, Business Partners, and Clients. Your efforts will help increase product acceptance and customer satisfaction. As a bonus, you will develop a network of contacts in IBM development labs, and increase your productivity and marketability. Find out more about the residency program, browse the residency index, and apply online at: ibm.com/redbooks/residencies.html

xii

FICON Planning and Implementation Guide

Comments welcome
Your comments are important to us! We want our books to be as helpful as possible. Send us your comments about this book or other IBM Redbooks publications in one of the following ways: Use the online Contact us review Redbooks form found at: ibm.com/redbooks Send your comments in an e-mail to: [email protected] Mail your comments to: IBM Corporation, International Technical Support Organization Dept. HYTD Mail Station P099 2455 South Road Poughkeepsie, NY 12601-5400

Preface

xiii

xiv

FICON Planning and Implementation Guide

Part 1

Part

Understanding FICON
This part introduces FICON and explains how it is exploited by the System z channel, the FICON Director, and the control unit. It also provides information about the FICON and Fibre Channel architectures and their uses in System z environments.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

FICON Planning and Implementation Guide

Chapter 1.

Introduction to FICON
The term Fibre Connection (FICON) represents the architecture as defined by the InterNational Committee of Information Technology Standards (INCITS), and published as ANSI standards. FICON also represents the names of the various System z server I/O features. This chapter discusses the basic Fibre Channel (FC) and FICON terminology, as well as System z FICON support, benefits, operating modes, and topologies. Throughout this chapter we use the term FICON to refer to FICON Express8, FICON Express4, FICON Express2, and FICON Express, except when the function being described is applicable to a specific FICON feature type.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

1.1 Basic Fibre Channel terminology

This section describes the Fibre Channel (FC) architecture and explains general terms that are used both in the FC and FICON environments. One defining characteristic of Fibre Channel is that it allows many existing, well-known, and long-implemented channel and networking protocols to run over the same physical interface and media. Consequently, there is a rich set of standards and terminology associated with its architecture. FICON uses the Single-Byte Command Code Sets -3 (SB-3) and Single-Byte Command Code Sets -4 (SB-4) implementation within the Fibre Channel standard. The FC architecture, as illustrated in Figure 1-1, is an integrated set of rules consisting of five layers (FC-0 through FC-4) for serial data transfer between computers, devices, and peripherals developed by the InterNational Committee for Information Technology Standards (INCITS) and American National Standard Institute (ANSI).

Audio / Video

FC-FCP FICON IPI / SCSI / HIPPI / SB / IP / SB-3/4 Channels Upper Level Protocol (ULP)

P / 802.2 Networks

FC-4
Mapping

Multimedia

FC-3 FC-2
Signaling Protocol

Common Services

Framing Protocol/Flow Control

FC-1
Transmission Protocol

Encode/Decode (8/10)

FC-0
Interface/Media

Physical Characteristics Copper/Multi-Mode Fiber/Single-mode Fiber

Figure 1-1 Fibre Channel architectural levels

The FICON channel architecture consists of the following Fibre Channel (FC) protocols: FC-0 level - Interface and Media The Fibre Channel physical interface (FC-0), specified in FC-PI, consists of the transmission media, transmitters, receivers, and their interfaces. The physical interface specifies a variety of media and associated drivers and receivers capable of operating at various speeds. FC-1 level - Transmission Protocol This is a link control protocol that performs a conversion from the 8-bit EBCDIC code into a 10-bit transmission code; a unique bit-pattern is assigned to each known hexadecimal character. Encoding is done by the N_Port when sending the character stream over the fiber, and the decoding back to 8-bit code is performed by the receiving N_Port.

FICON Planning and Implementation Guide

FC-2 level - Signaling Protocol Fibre Channel physical framing and signaling interface (FC-PH) describes the point-to-point physical interface, transmission protocol, and signaling protocol of high-performance serial links for support of higher-level protocols associated with HIPPI, IPI, SCSI, FC-SB-2/3/4 (FICON), and others. FC-3 level - Common Services This layer is intended to provide the common services required for advanced features. FC-4 level - Mapping The Upper Level Protocol (ULP) is part of FC-4 and describes IPI/FC-FCP (SCSI)/HIPPI/SB/IP and FICON. FICON was introduced with Single Byte-2 Command Code Sets mapping protocol, and then later revised with FC-SB-3. In 2008 another revision was drafted, identified as FC-SB-4; this revision will be used to support additional FICON functions. FC-SB-2, FC-SB-3, and FC-SB-4 architecture information and other FC documentation can be obtained from the following Web site: http://www.t11.org Fibre Channel provides the capability to build a configuration, as shown in Figure 1-2, that can operate in a point-to-point, arbitrated loop, or switched configuration.

Switched Fabric
FC Link
F_Port E_Port E_Port N_Port F_Port

N_Port

ISL
F_Port

FC Link
N_Port

FL_Port NL_Port NL_Port NL_Port

NL_Port

Figure 1-2 FC configuration with port types

Note: System z FICON channels do not support the arbitrated loop topology.

Chapter 1. Introduction to FICON

1.1.1 Node
A node is an endpoint that contains information. It can be a computer (host), a device controller, or a peripheral device (such as a disk or tape drive). A node has a unique 64-bit identifier known as the Node_Name. The Node_Name is typically used for management purposes.

1.1.2 Port
Each node must have at least one port (hardware interface) to connect the node to the FC topology. This node port is referred to as an N_Port. Each N_Port has a Port_Name, which is a unique 64-bit identifier that is assigned at the time it is manufactured. The N_Port is used to associate an access point to a nodes resources. Other port types include: E_Port F_Port FL_Port G_Port L_Port NL_Port U_Port An expansion port is used to interconnect switches and build a switched fabric. A fabric port is used to connect an N_Port to a switch that is not loop-capable. A fabric loop port is used to connect NL_Ports to a switch in a loop configuration. A generic port is a port that has not yet participated in a fabric. A loop port is a port in a Fibre Channel Arbitrated Loop (FC-AL) topology. A node loop port is an N_Port with loop capabilities. A universal port is a port that has not yet assumed a specific function in the fabric. It is a generic switch port that can operate as an E_Port, an F_Port, or an FL_Port.

The port type is determined by the nodes role in the topology, as shown in Figure 1-2 on page 5.

1.1.3 Switched fabric

One or more switches are interconnected to create a fabric, to which the N_Ports are connected. Switches may be interconnected with a Inter-Switch Link (ISL) to which the E_Ports are connected. A switched fabric takes advantage of aggregated bandwidth via switched connections between N_Ports.

1.1.4 FC link
The port connects to the topology through an FC link. The FC link is a fiber optic cable that has two strands. One strand is used to transmit a signal and the other strand is used to receive a signal (see Figure 1-3 on page 7). An FC link is used to interconnect nodes and switches.

FICON Planning and Implementation Guide

FC port

FC port
Outbound Outbound

Tx Rx

Tx Rx

Inbound

Inbound

Fibre Channel link

Figure 1-3 Fibre Channel link

For example, an FC link (port-to-port connection) can be: Node-to-node link (N_Port-to-N_Port) Node-to-switch link (N_Port-to-F_Port) Switch-to-switch link (E_Port-to-E_Port)

1.1.5 World Wide Name

Nodes and ports have unique 64-bit addresses that are used to identify them in an FC topology. These addresses are assigned by the manufacturer, with a vendor-specific portion defined by the IEEE standards committee. These addresses (in the FC standard) are called Node_Names and Port_Names, and when they are worldwide unique, they are referred to as: World Wide Node_Name (WWNN) World Wide Port_Name (WWPN) A WWN (any WWNN or WWPN) is usually written in sets of two hex digits, separated by colons (for example, 08:45:12:56:43:00:D5:A0). Figure 1-4 illustrates where the WWNs are used.
Server (Node) WWNN1
WWPN11 WWPN31

Controller (Node) WWNN3

N_Port
WWPN21

Switch (Node) WWNN2

WWPN22

N_Port

F_Port

F_Port

F_Port
WWPN23

F_Port
WWPN24

N_Port
WWPN12

N_Port
WWPN32

Figure 1-4 World Wide Names

Chapter 1. Introduction to FICON

On a System z server, the WWNN is constant for all FICON channels (ports); however, the WWPN is unique to each FICON channel on the server.

1.2 System z FICON

Because FICON is an IBM I/O technology built on the Fibre Channel (FC) architecture, many of the FC terms and definitions are used in the various FICON configurations. FICON channel architecture is compatible with: Fibre Channel Physical and Signaling standard (FC-FS) Fibre Channel Switch Fabric and Switch Control Requirements (FC-SW) Fibre Channel Single-Byte-3 (FC-SB-3) and Fibre Channel Single-Byte-4 (FC-SB-4) Even though FICON uses the FC architecture to transport commands and data over the fiber link, FICON is implemented differently from other FC implementations. For example, in a switched point-to-point topology, a server using FC may use a discovery or definition process to obtain a port address. System z FICON only uses the definition process. Discovery Uses the WWN and discovers the associated N_Port address by querying the Fabric Name Server. This approach is used by all Fibre Channel protocol Host Bus Adapters (HBAs) users connected to a switched point-to-point topology Definition Uses the Hardware Configuration Definition (HCD) tool or Input Output Configuration Program (IOCP) to define the associated N_Ports to which communication is allowed. The definitions in HCD/IOCP must match the actual port connections of the attached control units. FICON is widely used in the System z environment, and provides additional strengths and capabilities compared to the ESCON technology. Many additional capabilities have been included in support of FICON since it was originally introduced. Some control units and control unit functions may require FICON use exclusively. For example, Hyper Parallel Access Volume requires the use of FICON and will not work with ESCON.

Hyper Parallel Access Volume (HyperPAV)

Hyper Parallel Access Volume (HyperPAV) is an enhancement to the original PAV feature supported on DASD subsystems. HyperPAVs can be used to improve I/O efficiency and performance, while simultaneously reducing the number of I/O resources used in System z environments. HyperPAVs require FICON connections. Additional support or licenses may be required on the control unit. Refer to IBM System Storage DS8000: Architecture and Implementation, SG24-6786 for more information about HyperPAV. Other FICON enhancements may require the use of certain FICON features used on a specific System z server, such as the System z10 or System z9 servers. I/O adapters or ports supporting FICON reside on control units, switches, and servers. The I/O adapter may support various FICON operating modes. In this book we focus on the FICON native operating mode and the latest FICON capabilities available on the System z10 and System z9 servers.

FICON Planning and Implementation Guide

1.2.1 FICON advantages over ESCON

FICON offers many advantages over ESCON. It enables multiple concurrent I/O operations to occur simultaneously to multiple control units. This is one of the fundamental differences between FICON and ESCON, with its sequential I/O operations. FICON uses fiber cabling more efficiently, thereby enabling improved performance, longer distances, and more addresses as shown in Table 1-1. FICON channels also permit intermixing of large and small I/O operations on the same link, which is a significant enhancement compared with ESCON channels. The data center I/O configuration now has increased connectivity flexibility because of the increased I/O rate, increased bandwidth, and multiplexing of mixed workloads.
Table 1-1 FICON-to-ESCON comparison ESCON Channel concurrent I/O operations Command execution Fiber use Link data rate Switching Unrepeated distance from channel Max distance without performance degradation Unit addresses per channel One Synchronous Channel/CU handshake Half duplex 200 Mbps Circuit switching 3 km 9 km 1024 FICON Up to 32 or more depending on FICON feature and servera Asynchronous Full duplex Up to 8 Gbps b Packet switching 10 km c 100 km or more d 16384

a. Maximum number of concurrent I/O operations can vary depending on the use of high performance FICON and FICON feature. b. The maximum link rate depends on the specific FICON feature. FICON Express8 provides up to 8 Gbps link rate. c. Standard unrepeated distance for most FICON LX features is 10 km; additional distance may be achieved with an RPQ on some FICON features. d. To avoid performance degradation at extended distances, FICON switches, directors, or channel extension devices may be required.

FICON-to-ESCON solutions
The FICON Express feature was the last FICON channel feature to support CHPID type FCV. The FICON Express channel in FCV mode allows the accessing of ESCON interface control units with ESCON interfaces via a 9032-5 ESCON Director FICON Bridge adapter. Many System z environments only use FICON Express4 features and above. The FICON Express features are supported on System z10 and System z9 servers only if carried forward on a server upgrade. Therefore, another solution is needed for FICON channel-to-ESCON control unit connectivity, such as the PRIZM FICON to ESCON Converter from Optica Technologies Inc. The PRIZM FICON to ESCON Converter is a channel-based appliance that converts native FICON protocol to native ESCON protocol, thereby allowing attachment of existing ESCON devices directly to FICON channels.

Chapter 1. Introduction to FICON

Unlike the Model 5 ESCON Director FICON Bridge solution, PRIZM supports native FICON channel protocol (CHPID type FC), and supports attachment to existing FICON fabrics. PRIZM provides strategic flexibility to those migrating from ESCON to FICON on their System z servers. For additional details, refer to the following Web site: http://www.opticatech.com/?page_id=19 The PRIZM solution is also offered through IBM GTS Site and Facilities Services: http://www.ibm.com/services/us/index.wss/itservice/igs/a1026000

1.2.2 FICON operating modes and topologies

System z FICON features may be defined in HCD or IOCP to operate in one of up to three operating modes. The defined operating mode supported depends on the FICON feature installed in the System z server. The three operating modes that can be defined are: FICON conversion mode (FCV) This operating mode is only supported on the FICON Express feature FICON native mode (FC) Fibre Channel Protocol (FCP) As shown in Figure 1-5, a FICON channel in native mode can access FICON control units in any of the following three topologies: Point-to-point (direct connection) Switched point-to-point (via a FICON Director) Cascaded FICON Directors (through two FICON Directors)

Point-to-point

FC

FC Link

FICON CU

Switched point-to-point

FC

FC Link

FC Link

FICON CU

FICON CU

FC L

ink

FICON Director

FC Lin

FC

Cascaded FICON Directors

Site A
Site B
ISL
F

FC
FC Lin k

k C Lin

FICON CU

FICON Director

FICON Director
FC L in k

nk

FICON CU

FC

Figure 1-5 System z supported FICON native topologies

10

FICON Planning and Implementation Guide

FC

Li

A FICON channel also supports channel-to-channel (CTC) communications. The FICON channel at each end of the FICON CTC connection, supporting the FCTC control units, can also communicate with other FICON control units, such as disk and tape devices.

1.2.3 Terms used with FICON Directors

It is important to point out that some definition terms used in IOCP and HCD were carried over from the ESCON environment. For clarity, we use the following terms throughout this book to describe the definitions needed in IOCP and HCD for cascaded FICON Directors and switched point-to-point configurations: An entry switch is the FICON Director that is directly connected to the processors FICON channel and to the CU (destination) or another FICON Director. A cascaded switch is the FICON Director that connects to the CU (destination) and to the entry switch. The entry switch and cascaded switch are interconnected via an Inter-Switch Link (ISL). Switch ID and switch address (1-byte value) are terms used to address a FICON Director. A port address (1-byte value) is used to address the physical port on the FICON Director. The switch control unit port (CUP) function allows z/OS to manage a FICON Director with the same level of control and security as for an ESCON switch. Host communication includes control functions such as blocking and unblocking ports, as well as monitoring and error reporting functions.

1.2.4 Terms used with the Input/Output architecture

The System z channel environment uses many terms and acronyms. We define some of those terms here. The terms listed here will be used throughout this book as we discuss channel and I/O operations in the various chapters.

Input/output channels
Input/output (I/O) channels are components of the System z Channel Subsystem (CSS) and z/Architecture. They provide a pipeline through which data is exchanged between servers, or between a server and external devices. z/Architecture channel connections are referred to as channel paths.

Control unit
The most common attachment to a System z channel is a control unit (CU) accessed via an Enterprise Systems CONnection (ESCON) or a FIbre CONnection (FICON) channel. The CU controls I/O devices such as disk and tape devices.

I/O devices
An input/output (I/O) device provides external storage, a means of communication between data processing systems, or a means of communication between a system and its environment. In the simplest case, an I/O device is attached to one control unit and is accessible through one channel path.

Channel-to-Channel
The Channel-to-Channel (CTC) function simulates an I/O device that can be used by one system control program to communicate with another system control program. It provides the data path and synchronization for data transfer between two channels. When the CTC option is used to connect two channels that are associated with different systems, a loosely coupled

Chapter 1. Introduction to FICON

11

multiprocessing system is established. The CTC connection, as viewed by either of the channels it connects, has the appearance of an unshared input/output device.

Channel Subsystem
The Channel Subsystem provides the functionality for System z servers to communicate with input/output (I/O) devices and the network. The CSS has evolved with the increased scalability of IBM System z servers. The CSS architecture provides functionality in the form of multiple Channel Subsystems (CSSs). Multiple CSSs can be configured within the same System z (up to four on z10 EC servers and z9 EC servers, and up to two on z10 BC servers and z9 BC servers).

Subchannels
A subchannel provides the logical representation of a device to the program. It contains the information required for sustaining a single I/O operation. A subchannel is assigned for each device defined to the logical partition. Note that Multiple Subchannel Sets (MSS) are available on System z10 and System z9 servers to increase addressability. Two subchannel sets are provided: subchannel set-0 can have up to 63.75 K subchannels, and subchannel set-1 can have up to 64 K subchannels.

Channel Path Identifier

A Channel Path Identifier (CHPID) is a value assigned to each channel path of the system that uniquely identifies that path. A total of 256 CHPIDs are supported by the CSS.

Multiple Image Facility

Multiple Image Facility (MIF) enables resource sharing across logical partitions within a single CSS or across CSSs.

Channel spanning
Channel spanning extends the MIF concept of sharing channels across logical partitions to sharing channels across Channel Subsystems and logical partitions.

12

FICON Planning and Implementation Guide

Chapter 2.

System z FICON technical description

FICON uses the Fibre Channel Single Byte Command Sets-3 (FC-SB-3) standard as defined by the InterNational Committee of Information Technology Standards (INCITS), and published as ANSI standards. Late in 2008 IBM introduced System z High Performance FICON (zHPF) on the System z10. zHPF uses the architecture as described in the draft proposal for Fibre Channel Single Byte Command Sets-4 (FC-SB-4 ). FC-SB-4 is intended to be a complete replacement of the FC-SB-3 standard. This chapter provides a technical description of how System z servers use the FICON architecture to process I/O operations, as well as the FICON features and functions available on System z10 and System z9 servers.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

13

2.1 Using the FICON architecture for I/O operations

Before a System z FICON channel can send any I/O over the link, the link must be initialized. There are specific steps used in this initialization process, and they are described in detail in the next section. After the initialization process is complete, I/O operations can be started on the System z FICON channel. A FICON channel operates in what is known as command mode, exploiting the FC-SB-3 protocols. A System z10 FICON channel can operate in command mode and in transport mode (also known as zHPF) simultaneously, exploiting the FC-SB-4 protocols. Note the following points: When both the System z10 FICON channel and the control unit indicate support for zHPF, the channel will perform the I/O operation in transport mode. If the control unit does not support zHPF, the System z10 FICON channel will operate in command mode. zHPF must be enabled in z/OS to perform transport mode operations. Both modes (command and transport) are discussed in greater detail later in this chapter.

2.1.1 FICON initialization

The FICON channel initialization process establishes the necessary conditions for elements of a channel path to be able to sustain both link-level and device-level communication. Under normal conditions, the initialization process occurs infrequently. For example, it is usually required only after a power-on procedure or during a system initialization procedure, and is not part of the execution of an I/O operation. There is a set of initialization steps that is used with every connection. There is also a series of steps used for every control unit image stored in the HSA. For example, the initialization process for a FICON channel consists of the following procedures: Link initialization Channel login and security attribute determination Channel node-identifier acquisition Channel state-change registration Channel link-incident-record registration Process login, when supported by the channel and control unit Channel logical-path establishment For a control unit, the initialization process consists of the following procedures: Link initialization Control unit login Control unit node-identifier acquisition Control unit state-change registration Process login, if supported Control unit logical-path establishment The Fabric login (FLOGI) and Port login (PLOGI) process will determine the presence of a switch fabric. If a switch fabric is present, the channel performs an F_Port Login. If a switch fabric is not present, the channel performs an N_Port Login.

14

FICON Planning and Implementation Guide

Point-to-point
This is the initialization process for a point-to-point connection: 1. 2. 3. 4. 5. Channel N_Port Login process. Request Node Identifier (RNID) function to provide specific neighbor node information. Channel State-Change Registration (SCR) is sent to the control unit. Channel Link-Incident-Record Registration (LIRR). Process Login (PRLI) is used when channel and control unit support System z High Performance FICON (zHPF). 6. Establish Logical Path (ELP) to the control unit images that are stored in the Hardware Storage Area (HSA1) when each channel image is initialized.

Switched point-to-point
This is the initialization process for a switched point-to-point connection: 1. Channel F_Port Login process. 2. Query Security Attributes (QSA) determines if a cascaded switch configuration is supported and if two-byte destination addresses can be used. 3. RNID function provides specific neighbor node information. 4. Channel SCR is sent to the fabric controller. 5. LIRR channel to the management server. 6. Channel N_Port Login. The channel will also login to the defined control unit N_Port. The channel performs this for each control unit N_Port link address defined on this channel path. RNID to control unit. LIRR to control unit. PRLI used when channel and control unit support zHPF. ELP to the control unit images that are stored in the HSA.

Cascaded FICON Directors

The cascaded FICON Directors login process is similar to that of the switched point-to point process. The only difference is that the FICON channel and control unit log into (F_Port Login) two different FICON Directors. Figure 2-1 shows the key steps used in the FICON initialization process for a switched-point-to-point connection with a control unit that supports zHPF.

Channel initialization PLOGI/FLOGI QSA RNID PLOGI / LIRR / RNID ELP / LPE

CU initialization PLOGI/FLOGI RNID

FICON Director

PLOGI / LIRR / RNID

Control Unit

System z10

PRLI (zHPF support) ELP / LPE

Figure 2-1 FICON initialization - channel to control unit initialization

Also called the Hardware System Area, it is a protected area of storage where control blocks used by the channel subsystem are located.

Chapter 2. System z FICON technical description

15

Link initialization is described in FC-FS-2. When link initialization is complete, the N_Port or F_Port is in the active state. After link initialization is complete for an N_Port or F_Port, the port is considered to be operational as long as it remains in the active state. The link speed is also determined during this process, through auto-negotiation. The FICON link will start at the highest speed and work lower until both sides choose the same speed. FICON features support the following speeds: FICON Express2 will auto-negotiate to 2 Gbps or 1 Gbps. FICON Express4 will auto-negotiate to 4 Gbps, 2 Gbps, or 1 Gbps. FICON Express8 will auto-negotiate to 8 Gbps, 4 Gbps, or 2 Gbps.

FICON registration
System z10 servers support platform and name server registration to the fabric. When registered, information about the channels connected to a fabric will be available to other nodes or SAN managers. The attributes that are registered for the System z10 servers include: Platform information World Wide Node Name (WWNN) - This is the node name of the platform, and it is the same for all channels belonging to the platform. Platform type (host computer). Platform name - The name includes Vendor ID, product ID, and vendor-specific data from the node descriptor. Channel information World Wide Port Name (WWPN). Port type (N_Port_ID). FC-4 types supported. Classes of service supported by the channel. Platform registration is a service defined in the Fibre Channel - Generic Services - 4 (FC-GS-4) standard (INCITS (ANSI) T11 group).

2.1.2 FICON I/O request

Before the actual I/O operation starts and FC frames are transmitted over the fiber cable, the channel program must be created and passed to the Channel Subsystem (CSS). The Input/Output Supervisor (IOS) is a z/OS component, between the CSS and the application program, that issues the start subchannel command. The application program uses an access method that calls the I/O driver to build the channel program. The channel program may use transport commands when operating in transport mode, or channel commands when operating in command mode. Transport mode is used by the System z High Performance FICON (zHPF) architecture. A transport mode I/O operation uses a Transport Control Word (TCW) to send commands and data over the FC link and does not use CCWs. The channel and control unit must support zHPF. Command mode is used by the existing FICON architecture where an I/O operation consists of CCWs and data. When the SSCH is passed to the CSS (see Figure 2-2 on page 17) and the FICON channel is selected, the FC architecture and FICON protocols are used to build the frames required to send over the FC link. 16
FICON Planning and Implementation Guide

Application

z/OS I/O Requests

SSCH (ORB) I/O Interrupt

IOS UCB

System z Channel Subsystem FICON Channel FC4 (Protocol) FC3 (Services) FC2 (Framing) FC1
Encode/ Decode

FC0 (Optics)

FC-FS FC-2 frames, protocol type = FC-SB-3 or SB-4 (FICON)

SOF FC-2 Header

FICON payload

FC-2 CRC

EOF

FC Port addressing
Fibre Channel Fabric FICON CU
FC-2 Hdr - S_ID D_ID

S_ID

D_ID

D_ID

S_ID

Figure 2-2 FICON channel I/O request

A FICON I/O request flow is as follows: 1. An application or system component invokes an I/O request. The application or access method provides CCWs and additional parameters in the Operation Request Block (ORB). 2. The request is queued on the Unit Control Block (UCB)2. The Input Output Supervisor (IOS) will service the request from the UCB on a priority basis. 3. IOS issues a Start Subchannel (SSCH) instruction with the Subsystem Identification word (SSID) representing the device and ORB as operands. 4. The ORB contains start-specific control information. It indicates whether the channel is operating in transport mode (zHPF support) or command mode, and also indicates the starting address of the Channel Programs (CPA) Channel Command. 5. The CSS selects the most appropriate channel and passes the I/O request to it. 6. The channel fetches from storage the Channel Command Words and associated data (for Write Operations). 7. The channel assembles the required parameters and fields of the FC-2 and FC-SB-3 or FC-SB-4 for the I/O request and passes them to the Fibre Channel adapter (which is part of the FICON channel). Device-level information is transferred between a channel and a control unit in SB-3 or SB-4 Information Units (IUs). Information units are transferred using both link-level and device-level functions and protocols. For example, when the channel receives an initiative to start an I/O operation, the device-level functions and protocols obtain the command and other parameters from the current CCW and insert them into the appropriate fields within a command IU. When the
2

The UCB is a control block in memory that describes an I/O device to the operating system.

Chapter 2. System z FICON technical description

17

command IU is ready for transmission, link-level functions and protocols provide additional information (for example, address identifiers and exchange ID in the frame header) and then coordinate the actual transmission of the frame on the channel path. 8. The Fibre Channel adapter builds the complete FC-FS FC-2 serial frame and transmits it into the Fibre Channel link. As part of building the FC-FS FC-2 frame for the I/O request, the FICON channel in FICON native (FC) mode constructs the 24-bit FC port address of the destination N_Port of the control unit, and the control unit image and device address within the physical CU.

FICON frame layout

The FC-4 defines the mapping rules used by FICON. FC-SB-3 and FC-SB-4 are based on the FC-4 Information Unit constructs. Information associated with the execution of an I/O operation and the operation of a device is transferred between the channel and control unit as Information Units. IUs contain device-level commands, status, data, or control information or link-level control information. Figure 2-3 illustrates the FICON frame layout. The Fibre Channel (FC-2) sends the SB-3 and SB-4 IUs in FC-2 frames. The maximum amount of requester information that can fit into one FC frame is 2112 bytes. As many frames will be sent to transfer an IU as it takes to hold the IU data. The receiving end node (FC adapter at the CU or another channel) rebuilds the incoming FC frames for a given IU into an inbound IU. The FC header and its content determines how the frame is to be handled within the fabric The frame contains two parts: the FC-FS FC-2 header part, and the payload part (the FC-SB-4 or FC-SB-3 parts). The FC-2 header part of the frame holds the source FC port address (S_ID) and the destination FC port address (D_ID). The FC-SB-3 or FC-SB-4 header part of the frame holds the channel image address, the Control Unit Image address, and the Device Unit Address (UA).
When FC Header indicates SB-3 or SB-4 Payload: Contains channel command and/or data

FICON

Fibre Channel standard frame format

Word 0 1 2 3 4 5 Bits 31-24 R_CTL

4 Bytes Start of Frame (SOF)

2112 Bytes Payload

24 Bytes FC Header 64 Bytes Optional Header 4 Bytes CRC Error Check 4 Bytes End of Frame (EOF)

2048 Byte Payload

Bits 23-16

bits 15-8 Destination ID (Address) Source ID (Address) Sequence Count

bits 7-0

Sequence ID Originator Exchange ID

Responder Exchange ID

R_CTL indicates if it is a
Link control frame Data frame (FICON )

Exchange ID
Command mode (exchange pair is is used in an active I/O operation)

Figure 2-3 FICON frame layout

18

FICON Planning and Implementation Guide

The FC header and its content determine how the frame is to be handled. The R_CTL field indicates the type of frame it is, as follows: Link Control Acknowledge Link Response Link Command Data Frame Video_Data Link Data Basic Link Service (BLS) Extended Link Service (ELS) Device_Data (Type - IP, IPI-3, SCSI, FCP, SB-3 or SB-4) SB-3 and SB-4 is used for FICON The FC header also contains the Exchange ID. In FICON command mode, two exchange IDs are used and are called an exchange pair. Command mode uses one exchange pair for one I/O operation. An open exchange represents an I/O operation in progress over the channel. A detailed description of the FC frame and FICON FC-SB-3 or SB-4 usage documentation can be obtained from the following Web site: http://www.t11.org

2.1.3 Command mode

The System z FICON architecture defines the protocol for CCW and data pipelining, and how FICON channels operate (known as command mode). Command mode uses the CCWs and data as shown in Figure 2-4 to perform I/O operations.

SSCH I/O Request

FICON Channel
zArch CCW1 FC-SB-3 Cmd/Data IU Cmd/Data IU Cmd IU Cmd IU Cmd IU Cmd/Data IU Cmd/Data IU Data IU FC-FS FC2 FC2 FC2 FC2 FC2 FC2 FC2

Control Unit
Frame(s) Frame(s) Frame(s) Frame(s) Frame(s) Frame(s)
CCW1 CCW2 CCW3 CCW4 CCW5 CCW6 CCWn CMD2 CMR CMD1

Device

Data from CCW2 Memory CCW3

CCW4 CCW5 CCW6

CMD1 End CMD2 End

Data to Memory

CCWn

Frame(s) Frame
IU CE/DE (CCW-y)

CMDy

CMD-y End

I/O Status IU Request Control (StaAcc) Complete I/O Interrupt

Frame

Figure 2-4 FICON command mode operation

A fundamental difference with ESCON is the CCW chaining capability of the FICON architecture. ESCON channel program operation requires a Channel End/Device End (CE/DE) after executing each CCW. FICON supports CCW chaining without requiring a CE/DE at the completion of each CCW operation.

Chapter 2. System z FICON technical description

19

The ESCON channel transfers the CCW to the control unit and waits for a CE/DE presented by the control unit after execution of the CCW by the device (CCW interlock). After receiving CE/DE for the previous CCW, the next CCW is transferred to the control unit for execution. With a FICON channel, CCWs are transferred to the control unit without waiting for the first command response (CMR) from the control unit or for a CE/DE after each CCW execution. The device presents a logical end to the control unit after each CCW execution. After the last CCW of the CCW chain has been executed by the CU/device, the control unit presents a CE/DE to the channel. In addition, FICON channels can multiplex data transfer for several devices at the same time. This also allows workloads with low to moderate control unit cache hit ratios to achieve higher levels of activity rates per channel.

2.1.4 Transport mode

The way transport mode handles an I/O operation is significantly different from the CCW operation for command mode. While in command mode, each single CCW is sent to the control unit for execution. In transport mode, all commands are sent in a single frame to the control unit. The FICON channel uses transport mode if the control unit supports zHPF. When operating in transport mode, an I/O operation will use a transport control word (TCW) to send commands and data, as illustrated by Figure 2-5.

FICON Channel
zHPF Arch FC-SB-4 FC-FS

Control unit

SSCH I/O Request Data from Memory

4 KB read cmds
Prefix + 64 byes of data + Read cmd (4k) + Read cmd (4k) + Read cmd (4k) + Read cmd (4k) Cmd/Data IU FC2

Frame(s)

Data to Memory I/O Request Complete

TCW is used to obtain cmds and data for I/O operation

Frame(s)

Data (16 KB) + CRC

Frame(s)

Status
Total frames 11 20

I/O operation with four 4 KB reads Transport mode Command mode

Total commands 5 5

Total exchanges 1 2 (exchange pair)

Figure 2-5 FICON transport mode operation

The FICON Express8, FICON Express4, and FICON Express2 features support transport mode. A parameter in the Operation Request Block (ORB) is used to determine how the FICON channel will operate (command mode or transport mode). The mode used for an I/O operation also depends on the settings in the z/OS operating system.

20

FICON Planning and Implementation Guide

The application or access method provides the channel program commands and parameters in the ORB. Bit 13 in word 1 of the ORB specifies how to handle the channel program in either command mode or transport mode.

zHPF support determination

zHPF support is determined at link initialization. Figure 2-6 shows the initialization steps used to determine if transport mode will be used. During link initialization, both the channel and the control unit indicate whether they support zHPF. In the response to the Request Node ID (RNID) Extended Link Sequence (ELS), the Process Login (PRLI) support indicator is presented. In the RNID response, Process Login (PRLI) support is indicated in the node parameters, bit three. When bit three is one (1): The node supports the FC-SB-4 process login using the PRLI extended link service. FC-SB-4 protocol is used by zHPF When bit three is zero (0): The node does not support the FC-SB-4 process login. If PRLI is supported, the channel sends a PRLI ELS. The PRLI response then indicates that zHPF is supported by the control unit.

Channel-to-CU initialization PLOGI/ LIRR Send RNID Accept RNID (Indicate support for FC-SB-4 process login) Perform PRLI and determine transport mode support)

Control Unit

Figure 2-6 FICON sequence to determine zHPF support

A channel that supports the process login (PRLI) extended link service and transport mode operations will send a request to each control unit in its configuration that also supports the process login ELS. The PRLI is used to determine whether the control unit supports transport mode operations. zHPF provides additional capabilities such as being able to interrogate the CU before an actual missing interrupt occurs. Because zHPF uses TCWs (which do not time the channel operation in the same way as command mode), a mechanism must be used to prevent unnecessarily invoking the Missing Interrupt Handler (MIH) recovery actions. Transport mode provides an in-band method for z/OS to query the state of the I/O at the control unit without invoking error recovery or retry.

Missing Interrupt Handler

In channel operations, there is a Missing Interrupt Handler (MIH) time value set for various device classes. The MIH is constantly polling and is used to determine if there are any I/O operations that are not completing. For each of these I/O requests that are not completing, a recovery action will be performed. The recovery action could be one or more of the following: Issue a message Generate a logrec record for diagnostic purposes Terminate the I/O request Requeue the I/O request
Chapter 2. System z FICON technical description

21

Normally when a channel does not get a CMR response from a started I/O operation within the MIH value, it invokes MIH recovery procedures, which can then terminate the I/O in progress. The transport mode operation called interrogate is used to query the state of the CU. (Note that the interrogate process does not affect the state of the primary operation, and is sent before the MIH value is exceeded.) The CU provides an interrogate response IU that contains extended status that describes the state of the primary operation. z/OS can then decide whether recovery is needed, or to reset MIH value for this I/O operation.

2.1.5 Modified Indirect Data Address Word

On System z, the Modified Indirect Data Address Word (MIDAW) provides alternatives to using CCW data chaining in channel programs. The MIDAW facility has been added to z/Architecture and can coexist with the current CCW IDAW facility. MIDAW allows scattering of data in memory for non-contiguous real pages. (This is sometimes known as scatter-read or scatter-write.) Although the CCW IDAW function requires all but the first and last IDAW in a list to deal with complete 2 KB or 4 KB units of data, the MIDAW facility allows page boundary crossing on either 2 KB or 4 KB boundaries. This allows access to data buffers anywhere in a 64-bit buffer space. Figure 2-7 illustrates an example of MIDAW usage.

CCW with IDAW flag set (with ORB specifying 4K blocks) CCWs
cmd 04

IDAW address

Real address IDAWs Real address Real address

IDAWs usage limited because of addressing requirements

CCW with MIDAW flag set CCW

cmd 01

MIDAW address

MIDAWs

reserved reserved reserved L

2k 32 1k

Real address Real address Real address

IDAWs cannot be used for scattered reads or writes

Figure 2-7 Command mode MIDAWs

2.1.6 Transport Indirect Data Address Word (TIDAW)

When the System z channel is operating in transport mode, Transport Indirect Data Address Work (IDAW) is used. TIDAWs and MIDAWs are similar in concept and provide the capability of scattered reads and writes.

22

FICON Planning and Implementation Guide

TIDAWs are used when certain flag bits are set in the transport control word. Figure 2-8 illustrates an example of TIDAW usage.

TCW flag set for TIDAWS

TIDAWS

flags reserved count real address flags reserved count real address flags reserved count real address

Figure 2-8 Transport mode TIDAWs

TIDAWs and MIDAWs are used with z/OS extended format data sets that use internal structures (usually not visible to the application program) that require scatter-read or scatter-write operation.

2.1.7 Open exchange

An open exchange is part of FICON (and FC) terminology. Many I/O operations can be in progress over FICON channels at any one time. For example, a disk I/O operation might temporarily disconnect from the channel while performing a seek operation or while waiting for a disk rotation. During this disconnect time, other I/O operations can be managed, as explained here: Command mode open exchanges In command mode, the number of open exchanges is limited by the FICON Express feature. FICON Express8, FICON Express4, FICON Express2 allow up to 64 open exchanges. One open exchange (actually, this is an exchange pair) in command mode is the same as one I/O operation in progress. Transport mode open exchanges In transport mode, one exchange is sent from the channel to the control unit. Then the same exchange ID is sent back from the control unit to the channel to complete the I/O operation. The maximum number of simultaneous exchanges the channel can have open with the CU is up to 750 exchanges. The CU sets the maximum number of exchanges in the status area of the transport mode response IU (the default number is 64 and can be increased or decreased).

2.1.8 Buffer-to-buffer credit usage in FICON

Normally buffer credits are only an issue for extended distances. The buffer-to-buffer value determines the distance that the two nodes can be apart and still maintain the supported link data rate. All System z FICON channel features have enough buffer credits to support a distance of 10 km. When FICON frames travel over the link, a flow control mechanism must be in place to ensure the transmitter does not overrun the receiver. FICON mainly delivers class 3 type frames that use buffer-to-buffer credits to manage the flow control Ports have buffers for the temporary storage of frames. Each buffer can contain up to 2112 bytes of data. Buffers are referred to as credits. A credit represents a receiving port's ability to

Chapter 2. System z FICON technical description

23

accept one frame. At port initialization (login), buffer credit values are exchanged between two ports based on the number of buffers available for the ports. For example, the FICON Express8 feature has 40 buffer credits. This means that the FICON Express8 receiver is capable of storing up to 40 frames of data at any one time. The other FICON features support the following: FICON Express4 contains 212 buffer credits. FICON Express2 contains 107 buffer credits. FICON Express contains 64 buffer credits. For a more complete explanation about how buffer credits work, refer to Buffer credits on page 39.

2.1.9 Extended distance FICON

Degradation of performance at extended distances can be avoided by implementing an enhancement to the industry standard FICON architecture. The enhancement is a protocol for persistent IU pacing, and it can be used when the FC-SB-4 process login is supported. Control units that exploit the architecture can increase the pace count (the number of IUs allowed to be underway between the channel and the control unit). The IU pacing protocol controls the number of IUs that can be in flight from a channel to a control unit. The control unit may increase the pacing count (the number of IUs allowed to be in flight from a channel to a control unit) in the first Command Response IU sent to the channel. When the enhancement for persistent IU pacing is not used, the increased pacing count is valid only for the remainder of the current outbound exchange; see Figure 2-9.

FICON channel FC4 IU (command / data) FC 2 frames

frame frame frame frame frame

CU
CU response - I got frames send more IUs (CU may modify IU count at this time)

Figure 2-9 IU pacing for extended distance

At the start of every I/O, the IU count is reset to the default number. The IU pacing protocol, as defined, has the limitation that the first burst of IUs from the channel to the control unit may be no larger than a default value of 16. This causes a delay in the execution of channel programs with more than 16 IUs at long distances because a round trip to the control unit is required before the remainder of the IUs can be sent by the channel, upon the receipt of the first command response, as allowed by the increased pacing count. A channel may operate in the default IU pacing mode or in the persistent IU pacing mode. During initialization, the specific node descriptor information is exchanged between the channel and control unit. This information includes SB-4 support and will indicate whether the node supports concurrent enablement of the persistent IU pacing function. When a channel that supports concurrent enablement of the persistent IU pacing function receives a node descriptor from a control unit with bit 7 of byte 1 equal to one (1), the channel enables persistent IU pacing for all currently established logical paths with the control unit. When a control unit that supports concurrent enablement of the persistent IU pacing function receives a node descriptor from a channel with bit 7 of byte 1 equal to one (1), the 24
FICON Planning and Implementation Guide

control unit enables persistent IU pacing for all currently established logical paths with the channel. For logical paths that are established subsequent to the processing of the node descriptor, the persistent IU pacing bit in the ELP/LPE IU optional features field is used to enable or disable persistent IU pacing. Persistent IU pacing is a method for allowing a channel and control unit supporting the FC-SB-4 process login to retain a pacing count that can be used at the start of execution of a channel program. This may improve performance of long I/O programs at higher link speeds and long distances by allowing the channel to send more IUs to the control unit and eliminating the delay of waiting for the first Command Response. The channel retains the pacing count value, presented by the control unit in accordance with the standard, and uses that pacing count value as its new default pacing count for any new channel programs issued on the same logical path. Figure 2-10 illustrates an example of using persistent IU pacing for extended distance.

Channel

Initialization Establish logical path (ELP) LP established and increase IU count (Persistent IU pacing)

CU

Send FF IUs
FICON channel FC4 IU (command / data) FC 2 frames

frame frame frame frame frame frame frame frame frame CU response - I got frames (CU may modify IU count at this time)

CU DS8000

Figure 2-10 Persistent IU pacing for extended distance

Support is provided on the IBM System Storage DS8000 with the appropriate licensed machine code level, and is exclusive to System z10 servers.

2.2 System z FICON feature support

Table 2-1 on page 26 lists the FICON features available on the System z10 and System z9 servers, and the number of FICON channels support for each. Some features are only available when carried forward on a server upgrade and FICON Express8 features are only available on the System z10 server. The channels residing on a single FICON Express8, FICON Express4, or FICON Express2 feature can be configured individually, and can be defined in different channel modes (FC or FCP). The channels residing on a FICON Express feature may be defined for FC, FCP, or FCV channel mode; however, this feature is only supported on System z10 and System z9 severs if the feature was carried forward on an upgrade.

Chapter 2. System z FICON technical description

25

Table 2-1 System z server FICON feature support Maximum number of channels Channel feature Feature codes 2319 2320 3319 3320 3321 3322 3318 3323 3324 3325 3326 z9 EC 120 120 336 336 336 336 n/a n/a 336 n/a n/a z9 BC R07/S07 32/40 32/40 64/80a 64/80a 64/112 64/112 32/56 32/56 64/112 n/a n/a z10 EC 120 120 336a 336a 336 336 n/a n/a 336 336 336 z10 BC 40 40 112a 112a 128 128 64 64 128 128 128 Channels per feature 2 2 4 4 4 4 2 2 4 4 4 Channel increments (orderable) 2 2 4 4 4 4 2 2 4 4 4

FICON Express LXa FICON Express SXa

FICON Express2 LX FICON Express2 SX FICON Express4 10KM LXb FICON Express4 SX
b

FICON Express4-2C SX FICON Express4-2C 4KM LX FICON Express4 4KM LXb FICON Express8 10KM LX FICON Express8 SX

a. Carry forward on an upgrade b. Effective October 27, 2009, withdrawn from marketing for System z10 servers

FICON Express8
The FICON Express8 features are exclusive to System z10 and are designed to deliver increased performance compared to the FICON Express4 features. The FICON Express8 features have four independent channels, and each feature occupies a single I/O slot, utilizing one CHPID per channel. Each channel supports 2 Gbps, 4 Gbps, and 8 Gbps link data rates with auto-negotiation. A link data rate of 1 Gbps is not supported. All FICON Express8 features use Small Form Factor Pluggable (SFP) optics to permit each channel to be individually serviced in the event of a fiber optic module failure. The traffic on the other channels on the same feature can continue to flow if a channel requires servicing. The FICON Express8 features are ordered in 4-channel increments and are designed to be added concurrently. This concurrent update capability allows you to continue to run workloads through other channels while the FICON Express8 features are being added. FICON Express8 CHPIDs can be defined as a spanned channel and can be shared among logical partitions within and across CSSs. The FICON Express8 features are designed for connectivity to servers, switches, Directors, disks, tapes, and printers, and they can be defined as: CHPID type FC Native FICON, High Performance FICON for System z (zHPF), and FICON Channel-to-Channel (CTC) traffic Supported in the z/OS, z/VM, z/VSE, z/TPF, and Linux on System z environments CHPID type FCP Fibre Channel Protocol traffic for communication with SCSI devices Supported in z/VM, z/VSE, and Linux on System z environments

26

FICON Planning and Implementation Guide

FICON Express8 10KM LX

Feature code 3325 is designed to support unrepeated distances up to 10 km (6.2 miles) at 8 Gbps. Each channel supports 9/125 micrometer single mode fiber optic cable terminated with an LC Duplex connector.

FICON Express8 SX
Feature code 3326 is designed to support unrepeated distances up to 150 meters (492 feet) at 8 Gbps. Each channel supports 50/125 micrometer multimode fiber optic cable or a 62.5/125 micrometer multimode fiber optic cable terminated with an LC Duplex connector.

FICON Express4
The FICON Express4 features have four (or two, for the 2-port features) independent channels, and each feature occupies a single I/O slot, utilizing one CHPID per channel. Each channel supports 1 Gbps, 2 Gbps, and 4 Gbps link data rates with auto-negotiation. All FICON Express4 features use Small Form Factor Pluggable (SFP) optics to permit each channel to be individually serviced in the event of a fiber optic module failure. The traffic on the other channels on the same feature can continue to flow if a channel requires servicing. The FICON Express4 features are ordered in 4-channel (or 2-channel) increments and are designed to be added concurrently. This concurrent update capability allows you to continue to run workloads through other channels while the FICON Express4 features are being added. FICON Express4 CHPIDs can be defined as a spanned channel and can be shared among logical partitions within and across CSSs. The FICON Express4 features are designed for connectivity to servers, switches, Directors, disks, tapes, and printers, and they can be defined as: CHPID type FC Native FICON, High Performance FICON for System z (zHPF), and FICON Channel-to-Channel (CTC) traffic Supported in the z/OS, z/VM, z/VSE, z/TPF, and Linux on System z environments CHPID type FCP Fibre Channel Protocol traffic for communication with SCSI devices Supported in z/VM, z/VSE, and Linux on System z environments

FICON Express4 10KM LX

Feature code 3321is designed to support unrepeated distances up to 10 km (6.2 miles) at 4 Gbps. Each channel supports 9 micron single mode fiber optic cable terminated with an LC Duplex connector. Multimode (62.5 or 50 micron) fiber cable can be used with the FICON Express4 10KM LX feature. The use of multimode cable types requires a mode conditioning patch (MCP) cable. This feature should be used when the unrepeated distance between devices is greater than 4 km (2.5 miles) or the link loss budget between devices exceeds 2 dB. A 10KM LX transceiver is designed to interoperate with a 10KM LX transceiver. Interoperability of 10 km

Chapter 2. System z FICON technical description

27

transceivers with 4 km transceivers is supported if the unrepeated distance does not exceed 4 km.

FICON Express4 4KM LX

Feature code 3324 is designed to support unrepeated distances up to 4 km (2.5 miles) at 4Gbps. Interoperability of 4 km transceivers with 10 km(6.2 miles) transceivers is supported if the unrepeated distance does not exceed 4 km. Each channel supports 9 micron single mode fiber optic cable terminated with an LC Duplex connector. Multimode (62.5 or 50 micron) fiber cable can be used with the FICON Express4 4 KM LX feature. The use of multimode cable types requires a mode conditioning patch (MCP) cable.

FICON Express4-2C 4KM LX

Feature code 3323, with two channels per feature, is designed to support unrepeated distances up to 4 km (2.5 miles) at 4 Gbps. Interoperability of 4 km transceivers with 10 km (6.2 miles) transceivers is supported if the unrepeated distance does not exceed 4 km. Note: FICON Express4-2C 4KM LX is only available on System z10 BC and System z9 BC servers. Each channel supports 9 micron single mode fiber optic cable terminated with an LC Duplex connector. Multimode(62.5 or 50 micron) fiber cable can be used with the FICON Express4-2C 4 KM LX feature. The use of multimode cable types requires a mode conditioning patch (MCP) cable.

FICON Express4 SX
Feature code 3322 is designed to support unrepeated distances up to 270 meters (886 feet) at 4 Gbps. Each channel supports 62.5 micron or 50 micron multimode fiber optic cable terminated with an LC Duplex connector.

FICON Express4-2C SX
Feature code 3318, with two channels per feature, is designed to support unrepeated distances up to 270 meters (886 feet) at 4 Gbps. Note: FICON Express4-2C SX is only available on System z10 BC and System z9 BC servers. Each channel supports 62.5 micron or 50 micron multimode fiber optic cable terminated with an LC Duplex connector.

FICON Express2
The FICON Express2 SX and FICON Express2 LX features have four independent channels, with each feature occupying a single I/O slot, utilizing one CHPID per channel and four

28

FICON Planning and Implementation Guide

CHPIDs per feature, while continuing to support 1 Gbps and 2 Gbps link data rates. The link speed is auto-negotiated. The FICON Express2 SX and LX features are ordered in 4-channel increments and designed to be added concurrently. This concurrent update capability allows you to continue to run workloads through other channels while the FICON Express2 features are being added. FICON Express2 CHPIDs can be defined as a spanned channel and can be shared among logical partitions within and across CSSs. The FICON Express2 features are designed for connectivity to servers, switches, Directors, disks, tapes, and printers, and they can be defined as: CHPID type FC Native FICON, High Performance FICON for System z (zHPF), and FICON Channel-to-Channel (CTC) traffic Supported in the z/OS, z/VM, z/VSE, z/TPF, and Linux on System z environments CHPID type FCP Fibre Channel Protocol traffic for communication with SCSI devices Supported in z/VM, z/VSE, and Linux on System z environments Note: FICON Express2 is supported on System z10 and System z9 servers only if carried forward on an upgrade.

FICON Express2 LX
Feature code 3319 is designed to support unrepeated distances up to 10 km (6.2 miles) at 2 Gbps. Each channel supports 9 micron single mode fiber optic cable terminated with an LC Duplex connector. Multimode (62.5 or 50 micron) fiber cable can be used with the FICON Express2 LX feature. The use of multimode cable types requires a mode conditioning patch (MCP) cable.

FICON Express2 SX
Feature code 3320 is designed to support unrepeated distances up to 500 meters (1640 feet) at 2 Gbps. Each channel supports 62.5 micron or 50 micron multimode fiber optic cable terminated with an LC Duplex connector.

FICON Express
The two channels residing on a single FICON Express feature occupy one I/O slot in the System z I/O cage. Each channel can be configured individually and support 1 Gbps link speed. The FICON Express features are designed for connectivity to servers, switches, Directors, disks, tapes, and printers, and they can be defined as: CHPID type FC Native FICON and FICON Channel-to-Channel (CTC) traffic Supported in the z/OS, z/VM, z/VSE, z/TPF, and Linux on System z environments CHPID type FCP Fibre Channel Protocol traffic for communication with SCSI devices Supported in z/VM, z/VSE, and Linux on System z environments CHPID type FCV

Chapter 2. System z FICON technical description

29

 Supported in the z/OS, z/VM, z/VSE, z/TPF, and Linux on System z environments Note: FICON Express is supported on System z10 and System z9 servers only if carried forward on an upgrade.

FICON Express LX
Feature code 2319 is designed to support unrepeated distances up to 10 km (6.2 miles) at 1 Gbps. Each channel supports 9 micron single mode fiber optic cable terminated with an LC Duplex connector. Multimode (62.5 or 50 micron) fiber cable can be used with the FICON Express LX feature. The use of multimode cable types requires a mode conditioning patch (MCP) cable.

FICON Express SX
Feature code 2320 is designed to support unrepeated distances up to 860 meters (2822 feet) at 1 Gbps. Each channel supports 62.5 micron or 50 micron multimode fiber optic cable terminated with an LC Duplex connector. Table 2-2 summarizes the available FICON feature codes and their respective specifications. Notes: Mode Conditioning Patch (MCP) cables can be used with FICON features that can operate at a link data rate of 1 Gbps (100 MBps) only. FICON Express8 features do not support the attachment of MCP cables.
Table 2-2 FICON channel specifications Feature code Feature name Connector type Cable type Link data rate 1 or 2 Gbpsa Server z10 ECb z10 BCb z9 ECb z9 BCb z10 ECb z10 BCb z9 ECb z9 BCb z10 ECb z10 BCb z9 ECb z9 BCb

SM 9 m 2319 FICON Express LX LC Duplex with MCPc: MM 50 m or MM 62. 5m MM 62.5 m 2320 FICON Express SX LC Duplex MM 50 m

1 Gbps

1 or 2 Gbpsa

30

FICON Planning and Implementation Guide

Feature code

Feature name

Connector type

Cable type

Link data rate 1 or 2 Gbpsa

Server z10 ECb z10 BCb z9 ECb z9 BCb z10 ECb z10 BCb z9 ECb z9 BCb z10 ECb z10 BCb z9 ECb z9 BCb z10 BC z9 BC

SM 9 m 3319 FICON Express2 LX LC Duplex with MCPc : MM 50 m or MM 62.5 m MM 62.5 m 3320 FICON Express2 SX LC Duplex MM 50 m

1 Gbps

1 or 2 Gbpsa

MM 62.5 m 3318 FICON Express4-2C SX LC Duplex MM 50 m SM 9 m 3321 FICON Express4 10KM LX LC Duplex with MCPc : MM 50 m or MM 62.5 m MM 62.5 m 3322 FICON Express4 SX LC Duplex MM 50 m SM 9 m 3323 FICON Express4-2C 4KM LX LC Duplex with MCPc : MM 50 m or MM 62.5 m SM 9 m 3324 FICON Express4 4KM LX LC Duplex with MCPc : MM 50 m or MM 62.5 m SM 9 m MM 62.5 m 3326 FICON Express8 SX LC Duplex MM 50 m 1, 2, or 4 Gbpsa z10 BC z9 BC 1, 2, or 4 Gbpsa 1, 2, or 4 Gbpsa 1, 2, or 4 Gbpsa

z10 ECb z10 BCb z9 EC z9 BC z10 ECb z10 BCb z9 EC z9 BC

1, 2, or 4 Gbpsa

z10 ECb z10 BCb z9 EC z9 BC z10 EC z10 BC z10 EC z10 BC

3325

FICON Express8

LC Duplex

10KM LX

2, 4, or 8 Gbpsd 2, 4, or 8 Gbpsd

a. Supports auto-negotiate with neighbor node b. Only supported when carried forward on an upgrade c. Mode conditioning patch cables may be used d. Supports auto-negotiate with neighbor node for link data rates 8, 4, and 2 Gbps only

Note: IBM does not support a mix of 50 m and 62.5 m fiber optic cabling in the same physical link.

Chapter 2. System z FICON technical description

31

Refer to Table 4-1 on page 67 for the allowable maximum distances and link loss budgets based on the supported fiber optic cable types and link data rates.

32

FICON Planning and Implementation Guide

Chapter 3.

FICON Director technical description

This chapter provides a technical description of the common components and functions of a FICON Director that are needed to support FICON environments. It also offers information regarding FICON Directors that have been qualified by IBM for System z use. Terminology: Often the terms Director and switch are used interchangeably. However, compared to a switch, a Director typically means more fully redundant hardware resulting in higher availability, larger port capacity, and additional capabilities.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

33

3.1 The role of the FICON Director

The FICON Director provides a variety of features and functions. Its primary function is to allow the dynamic connections between System z servers, control units, or another FICON Director. The FICON Director provides the ability to logically connect multiple control units to a set of channels from one or more System z servers, which in turn minimizes the cabling infrastructure and maximizes the connectivity. One might say that the FICON Director is at the heart of I/O connectivity for System z servers. Figure 3-1 shows a conceptual view of frame processing in a FICON Director with multiple FICON channels and control units.

FICON Channel

FIC ON

Fra m

es

O FIC

es r am NF

FICON Control Unit

es

FIC ON F

ram es

FICON Director

ram NF ICO F

es ram NF O FIC s me Fra ON FIC

FIC O

FIC ON Fra me s NF ram es

FICON Channel

FICON Control Unit

Figure 3-1 FICON Director - frame processing

The FICON channel supports multiple concurrent I/O connections. Each concurrent I/O operation can be to the same FICON control unit (but to different devices/control unit images), or to different FICON control units. A FICON channel uses the Fibre Channel communication infrastructure to transfer channel programs and data through its FICON features to another FICON-capable node, such as a storage device, printer, or System z server (channel-to-channel). See Chapter 2, System z FICON technical description on page 13 for more details. A FICON channel, in conjunction with the FICON Director, can operate in two topologies: 1. Switched point-to-point (through a single FICON Director to FICON-capable control units) 2. Cascaded FICON Directors (through two FICON Directors to FICON-capable control units)

3.1.1 Switched point-to-point configuration

In a switched point-to-point connection, at least two Fibre Channel (FC) links are needed in the channel-control unit path: 1. Between the System z (N_Port) and the FICON Director (F_Port) 2. Between the FICON Director (F_Port) and the control unit (N_Port)

34

FICON Planning and Implementation Guide

The FICON channel determines whether the associated link is in a point-to-point or switched topology. It does so by logging into the fabric using fabric login (FLOGI ELS), and checking the accept response to the fabric login (ACC ELS). The FLOGI-ACC (accept) response indicates if the channel (N_Port) is connected to another N_Port (point-to-point) or an F_Port (fabric port). An example of a switched point-to-point topology is shown in Figure 3-2.
System z Server Storage
Determines fabric connection (point-to-point or switched)

Fabric Login (FLOGI) FICON Channel FICON Channel

N_Ports

F_Ports

FICON Director

F_Ports

N_Ports

FICON Control Unit Port FICON Control Unit Port

Logical Path

Figure 3-2 Switched point-to-point configuration

Multiple channel images and multiple control unit images can share the resources of the FC link and the FICON Director, such that multiplexed I/O operations can be performed. Channels and control unit links can be attached to the FICON Director in any combination, depending on configuration requirements and available resources in the Director. Sharing a control unit through a FICON Director means that communication from a number of System z channels to the control unit can take place over one Director. A FICON channel can communicate with a number of FICON control units on different ports as well. The communication path between a channel and a control unit is composed of two different parts: the physical channel path and the logical path. In a FICON switched point-to-point topology (with a single Director), the physical paths are the FC links, or an interconnection of two FC links through a FICON Director, that provide the physical transmission path between a channel and a control unit. A FICON (FC-SB-3 or FC-SB-4) logical path is the relationship established between a channel image and a control unit image for communication during execution of an I/O operation and presentation of status.

3.1.2 Cascaded FICON Director configuration

In a cascaded FICON Director configuration, at least three Fibre Channel (FC) links are needed in the channel-control unit path: 1. Between the System z (N_Port) and the FICON Director (F_Port) 2. Between the two FICON Directors (E_Ports) 3. Between the FICON Director (F_Port) and the control unit (N_Port) With this configuration, the connection between sites can consist of multiple FC links, also known as Inter-Switch Links (ISLs). An example of cascaded FICON Director topology is shown in Figure 3-3 on page 36. Multiple channel images and multiple control unit images can share resources of the FC link and FICON Directors, such that multiplexed I/O operations can be performed.

Chapter 3. FICON Director technical description

35

Channels and control unit FC links can be attached to the FICON Directors in any combination, depending on configuration requirements and available Director ports. Sharing a control unit through a FICON Director means that communication from a number of channels to the control unit can take place either over one Director-to-CU link (in the case where a control unit has only one FC link to the FICON Director), or over multiple link interfaces (in the case where a control unit has more than one FC link to the FICON Director). A FICON channel can also communicate with a number of FICON control units on different ports in the second FICON Director. The communication path between a channel and a control unit is composed of two different parts: the physical channel path and the logical path. In a cascaded FICON Director topology, the physical paths are the FC links, interconnected by the Directors, that provide the physical transmission path between a FICON channel and a control unit. Figure 3-3 illustrates the configuration.
System z Server
Determines fabric connection (point-to-point or switched)

Storage

Fabric Login (FLOGI) FICON Channel FICON Channel ISLs FICON Control Unit Port FICON Control Unit Port

FICON Director
Logical Path

E_Ports

FICON Director

Figure 3-3 Cascaded FICON Director configuration

A FICON (FC-SB-3 or FC-SB-4) logical path is the relationship established between a FICON channel image and a control unit image for communication during execution of an I/O operation and presentation of status.

3.1.3 Basic components of a FICON Director

A FICON Director is an expandable and scalable Fibre Channel platform designed to support FICON attachment of System z servers and storage products. It typically features an N+1 redundancy scheme for all of its components. This redundancy results in high availability. Even in the case of failure of any single component, the FICON Director remains operational. Most activities can be done non-disruptively; such as loading and activating firmware, and replacement of failed parts. A FICON Director consists of a chassis that usually holds the following types of components: Backplane module Power supply assembly Fan module assembly I/O module Switch module

36

FICON Planning and Implementation Guide

Backplane module
The backplane provides the connectivity within the chassis and between all system (I/O, switch, and control) modules, including the power supply assemblies and fan assemblies. The backplane gives you the capability to increase the number of ports by adding I/O modules.

Oversubscription The term oversubscription implies that the port may not be able to drive at full speed due to
limited backplane capacity. Oversubscription may only occur when simultaneous activities are present and backplane capacity is depleted. This usually applies to I/O modules with high port counts (for example, 48 ports operating at 8 Gbps).

Power supply assembly

Usually two power assemblies are installed in a chassis, providing full redundancy (and in some cases, load sharing). Each power supply assembly is a removable, hot-swappable unit.

Fan module assembly

Fan module assemblies provide the cooling capability for the FICON Director. In the event of a single fan failure, the remaining fans may increase their speeds to a higher operating rpm to compensate for the single fan failure. The fan module assembly is hot-swappable.

I/O module
The main function of the I/O module is to provide the physical connectivity between FICON channels and control units that are being connected to the Director. Due to the usage of Small Form Factor pluggable (SFP) ports instead of fixed optical port cards, port upgrades are simple and nondisruptive to other ports on the I/O module. Typically, any mix of short wavelength and long wavelength ports is allowed with any I/O module.

Switch module
The switch module on the FICON Director contains the microprocessor and associated logic that provides overall switching in the Director. The base configuration always contains two switch modules for redundancy reasons. If the active switch module fails or is uninstalled, the standby switch module automatically becomes the new active switch module. Failover occurs as soon as the active switch module is detected to be faulty or uninstalled. The active switch module performs following control functions: Switch initialization High availability and switch drivers Name server SNMP Zoning Typically, firmware upgrades can be performed concurrent to FICON Director operation.

3.1.4 Basic functions of a FICON Director

In this section we discuss the basic functions of a FICON Director that are needed to support a FICON environment. Some common FICON Director terms are described, as well.

Domain ID or switch address

The Domain ID is used to define the FICON Director address (switch address). It is specified in the FICON Director as a decimal number that is converted to a hexadecimal value for the

Chapter 3. FICON Director technical description

37

use in the System z server. The switch address is a hexadecimal number used in defining the Director in the System z environment. The Domain ID and the switch address must be the same when referring to a Director in a System z environment. Each FICON Director in a fabric must have a unique switch address. The switch address is used in the HCD or IOCP definitions, and it can be any hex value between x00 and xFF. The valid Domain ID range for the FICON Director is vendor-dependent. When defining the switch addresses in the HCD or IOCP, ensure that you use values within the FICON Directors range. The switch ID must be assigned by the user, and it must be unique within the scope of the definitions (IOCP and HCD). The switch ID in the CHPID statement is basically used as an identifier or label. Although the switch ID can be different than the switch address or Domain ID, we recommend that you use the same value as the switch address and the Domain ID when referring to a FICON Director. The Domain ID is assigned by the manufacturer, and it can be customized to a different value. It must be unique within the fabric and insistent. Insistent means that the FICON Director uses the same Domain ID every time, even if it is rebooted. Insistent Domain IDs are required in a cascaded FICON Director environment. There is no need for a Domain ID to ever change in a FICON environment, and fabrics will come up faster after recovering from a failure if the Domain ID is insistent.

Port address or link address

The port address is a hexadecimal number between x00 and xFF that is used in the IOCP or HCD to define a port on the FICON Director to which the device is connected. This is known as a one-byte link address. Some FICON Directors can modify the port address to any hexadecimal number that is required. The port address is also needed in cascaded FICON Directors, where it is part of the two-byte link address defined in the IOCP or HCD. The two-byte link address consists of the Domain ID and the port address to which the device is connected.

Port number
The port number is typically used in open systems environments to refer to a port on a Director. In some cases, port numbers and port addresses are the same. However, in many other cases they are different, depending on the Director port characteristics. Port addressing is platform- and vendor-specific.

Inter-Switch Link
Inter-Switch Links (ISLs) are the fiber optic cables that connect two Directors using expansion ports (E_Ports). ISLs carry frames originating from N_Ports, as well as frames that are generated within the fabric for management purposes. Multiple ISLs are usually installed between Directors to increase bandwidth and provide higher availability.

Fabric Shortest Path First

Fabric Shortest Path First (FSPF) is a Fibre Channel routing protocol that applies a weighting factor to paths based on link speed and status. FSPF associates a cost for each active link and computes paths from a Director to all the other Directors in the fabric. Paths are computed by adding the cost of all the links traversed by the path. Then a path is chosen with the lowest cost.

38

FICON Planning and Implementation Guide

Note: FSPF only provides basic utilization of ISLs with the same cost. It does not ensure all ISLs are equally utilized, and it does not inhibit congestion even if the other ISLs between Directors are not fully utilized.

Control Unit Port

The Control Unit Port (CUP) protocol is used by System z to provide in-band management for FICON Directors. It allows z/OS to manage a FICON Director with the same level of control and security as for an ESCON Director. Host communication is via IBM Tivoli Systems Automation for z/OS (SA for z/OS), and includes control functions like blocking and unblocking of ports, as well as monitoring and error reporting functions. Although the FICON Director is transparent to the operating system in the path to a FICON control unit or device during the execution of an I/O operation, the CUP can provide the following: I/O management of the FICON Director FICON Director performance monitoring Error detection of the FICON Director

Port swapping
Port swapping refers to the capability of a FICON Director to redirect traffic on a failed F_Port to a working F_Port without requiring a change in System z I/O configuration. In a Fibre Channel fabric, a failing port typically results in a cable being reconnected to another available port, followed by automatic discovery of the device through the name server. However, in a FICON environment, the control unit link address is defined in the channel configuration file (IOCP) of the System z server. Therefore, the FICON Director must ensure that the N_Port address for the control unit remains the same even after the cable is reconnected to a different switch port. After port swapping, the address assigned to the N_Port performing fabric login (FLOGI) by the alternate port should be the same as the one that would have been assigned by the original port. The original port will assume the port address associated with the alternate port.

Buffer credits
Buffer credits are used by the Fibre Channel architecture as a flow control mechanism to represent the number of frames a port can store. Buffer credits are mainly a concern when dealing with extended distances and higher speeds. However, a poorly designed configuration may also consume available buffer credits and have a negative impact on performance. Buffer credit flow control is done at the link level (that is, between an N_Port and an F_Port), and is used by Class 2 and Class 3 traffic. It relies on the receiver-ready (R_RDY) signal to replenish credits. The total number of available buffer credits (BCs) that a receiver has is determined during the link initialization process (see Figure on page 40).

Chapter 3. FICON Director technical description

39

Channel N_Port
Xmit Rec 40 BCs
BCs can be different across ports Number of BCs on receiver port

F_Port
Xmit Rec 10 BCs

F_Port
Xmit Rec 10 BCs
Number of BCs on receiver port

CU N_Port
Xmit Rec 5 BCs

Figure 3-4 Buffer credit determination

The sender monitors the receiver's ability to accept frames by managing a buffer credit count value of the receiver port. The buffer credit count value is decremented by the transmitter when a frame is sent (see Figure 3-5).

Channel N_Port
Xmit Rec

BC counter

7 Frame 3 Frame 2 Frame 1

10 Buffers Xmit Rec

BC counter

CU N_Port
5 Buffers

5 Xmit Rec

Figure 3-5 Buffer credit counters

An R_RDY signal is returned for every received frame when a receive port buffer is made available. The buffer credit count value is incremented when an acknowledgment (R_RDY) is received (see Figure 3-6).
BC counter

Channel N_Port
Xmit Rec

8
Credit 1

Frame 3 Frame 2

10 Buffers Frame 1

BC counter

CU N_Port
5 Buffers

4 Xmit Frame 1

R_RDY

R_RDY is returned to indicate a buffer is available for additional frame

BC count decremented when frame is sent

Figure 3-6 Buffer credit counters - R_RDY

If the buffer credit count is zero (0), frame transmission to the associated port is temporarily suspended. This flow control mechanism prevents data overruns.

Buffer credit calculation based on full frame size

To calculate the number of buffer credits needed to fully utilize an FC link with an average FICON frame size of 2112 bytes, use the following criteria: 1 buffer credit per kilometer (km) at 2 Gbps 2 buffer credits per kilometer (km) at 4 Gbps 4 buffer credits per kilometer (km) at 8 Gbps In this case, a 10 km FC link running at 8 Gbps for would need 40 buffer credits.

4 (per km at 8 Gbps) x 10 km = 40 buffer credits

40

FICON Planning and Implementation Guide

It is important to note that not all applications utilize the same frame size. FICON frame size information can be tracked using the FICON Director Activity Report in RMF. For RMF to receive the information, the CUP must be configured and the FDR parameter in SYS1.PARMLIB must be set to Yes. The FICON frame size of 2112 bytes can be achieved through the use of zHPF with large sequential reads/writes, or with a very large block size such as 4 x 4 KB or 12 x 4 KB blocks per I/O.

High integrity fabric support

Another important benefit of FICON support of cascaded Directors is its ability to provide high integrity data paths. The high integrity function is an integral component of the FICON architecture when configuring FICON channel paths through a cascaded fabric. End-to-end data integrity is designed to be maintained through the cascaded FICON Directors. High integrity helps to ensure that any changes to the data streams are always detected, and that the data frames are delivered to the correct endpoint (FICON channel or control unit port). For FICON channels, Cyclical Redundancy Checking (CRC) and Longitudinal Redundancy Checking (LRC) are bit patterns added to the data streams to allow the detection of any bit changes. High integrity features are used within the FICON channel and the cascaded FICON Directors to ensure the detection and reporting of any miscabling actions occurring during operational use, which will cause data frames to be delivered to the wrong endpoint. A FICON channel requires that the cascaded FICON Directors support high integrity. During initialization, the FICON channel will query the fabric to determine that it supports high integrity. If it does, then the channel will complete the initialization process, allowing the channel to operate with the fabric. After a FICON switched fabric has been customized to support FICON cascaded Directors, and the required World Wide Node Name (WWNN) and Domain IDs have been added in the fabric membership list, the Director will check that its Inter-Switch Links (ISLs) are attached to the correct Director before they are made operational. If an accidental cable swap occurs, the Director invokes logical path testing, reporting, isolation, and recovery. The high integrity fabric feature for cascaded FICON Directors protects against miscabling and misdirecting of data streams, as shown in Figure 3-7 on page 42.

Chapter 3. FICON Director technical description

41

5
3
FICON Director

FICON Director

System z servers

oops

Storage Devices

2
FICON Director FICON Director

Figure 3-7 High integrity fabric feature

The checking process proceeds in this way: 1. Channel initialization completes. 2. At some later time, miscabling occurs (for example, cables are swapped at a patch panel). 3. The Director port enters invalid attachment state and notifies state change back to System z. 4. The System z server invokes the channel logical path testing, reporting, and isolation, and error recovery. 5. Any I/O requests to the invalid route are discarded until the error is corrected. 6. Data is protected.

Fabric binding
Fabric binding is a security feature that enables explicit control over which FICON Directors can be interconnected by preventing non-authorized Directors from merging, either accidentally or intentionally. This is done by manually defining the authorized Director in a fabric binding database. The fabric binding database contains the World Wide Node Name (WWNN) and Domain ID of the connecting FICON Director. The FICON Director that is allowed to connect must be added to the fabric binding database of the other FICON Director. Note: Activating fabric binding is a prerequisite for a cascaded FICON Director topology.

Insistent Domain IDs

Insistent Domain ID support restricts a switch address (Domain ID) from being automatically changed when a duplicate switch address is added to the fabric. Overt operator action is required to change a switch address. Insistent Domain ID support also prohibits the use of dynamic Domain IDs to ensure that predictable Domain IDs are enforced within the fabric. For example, if a Director has this

42

FICON Planning and Implementation Guide

feature enabled, and a new Director is connected to it (through an ISL) without the Domain ID feature enabled, the new Director is segmented into a separate fabric and user data will not flow.

Zoning
Zoning is a method used in the FICON Director to enable or disable communication between different attached devices. A zone consists of a group of ports or WWNs. Connectivity is permitted only between connections to the Director that are in the same zone. There are two types of zoning: Name zoning, which permits connectivity between attached nodes based on WWN Port zoning, which restricts port connectivity based on port number WWN zoning is typically used for open systems connections. Port zones should be used for FICON. Nodes that are not already defined to a zone are automatically put into the default zone. Conversely, default zone members are automatically removed whenever that member is added to an active zone.

Protocol intermix
An intermix of FICON and Fibre Channel Protocol (FCP) is supported by FICON Directors at the port level (each port can run either FICON or FCP). This means that the SAN infrastructure can be shared by both protocols, to create a commonly managed infrastructure.

Management capabilities
FICON Directors can be managed using different communication methods, such as: In-band management via a FICON Channel (using the CUP) Out-of-band management via the following: IP-based client/server management software Simple network management protocol (SNMP) Trivial file transfer protocol (TFTP) to load firmware Command-line Interface (CLI) via Telnet GUI provided by the Director vendor or other vendors CLI via the serial interface (RS232) Call home via a modem connection (for notification purposes) Figure 3-8 on page 44 shows the various interfaces that can be used to manage the FICON Director.

Chapter 3. FICON Director technical description

43

CLI (via Telnet), SNMP, and TFTP

Vendor GUI or Web-based Interface

SA for z/OS and RMF (using the CUP)

Out-of-band via Ethernet (IP) CLI

In-band via the CUP FICON Channel Call Home

Serial Interface

Figure 3-8 FICON management interfaces

3.2 Qualified FICON Directors

The purpose of IBM qualification testing for FICON Directors is to ensure that any issues related to interoperability with System z servers (such as functionality, availability, data integrity, and performance) are identified and resolved before the FICON Directors are qualified for System z use. Qualification testing is carried out in the IBM Vendor Solutions Connectivity (VSC) Lab in Poughkeepsie, New York, U. S. The test facilities provide the capability to configure a fully functional FICON environment to drive the various components and protocols of the FICON Directors, as well as simulate failure and recovery actions. The recovery tests check for the correct detection of a fault or error condition by the attached FICON Directors, and ensure that the recovery adheres to System z architecture rules. For a test to be classified successful, all errors must be detected and properly handled by the attached FICON Directors. In addition, IBM proprietary software and microcode utility test suites stress the FICON Directors to ensure they can perform at levels beyond typical customer usage. In this section, we focus on five of the qualified FICON Directors that support a link data rate of 8 Gbps: two from the IBM System Storage SAN b-type family and three from the Cisco MDS 9500 Series. The SAN b-type family consists of the IBM SAN768B and the IBM SAN384B,. The Cisco MDS 9500 Series encompasses the Cisco MDS 9513, Cisco MDS 9509, and the Cisco MDS 9506. Table 3-1 on page 45 lists the IBM and Cisco FICON Directors and some of their key capabilities.

44

FICON Planning and Implementation Guide

Table 3-1 Product

IBM and Cisco FICON Directors IBM Model 2499-384 2499-192 2054-E11 2054-E07 2054-E04 Supported protocols FICON, FCP FICON, FCP FICON, FCP FICON, FCP FICON, FCP Protocol intermix Yes Yes Yes Yes Yes Number of ports 16 to 384 16 to 192 12 to 528 12 to 336 12 to 192 Port speed (Gbps) 1, 2, 4, 8 1, 2, 4, 8 1, 2, 4, 8 1, 2, 4, 8 1, 2, 4, 8 ISL speed (Gbps) 2, 4, 8, 10 2, 4, 8, 10 2, 4, 8, 10 2, 4, 8, 10 2, 4, 8, 10

IBM System Storage SAN768B IBM System Storage SAN384B Cisco MDS 9513 Cisco MDS 9509 Cisco MDS 9506

To find other System z qualified FICON Directors, go to: http://www-03.ibm.com/systems/storage/san/index.html

3.2.1 IBM System Storage SAN b-type family components

The SAN768B and the SAN384B are built on a common platform. The SAN768B can switch up to 384 ports simultaneously. The SAN384B can switch up to 192 ports simultaneously. The SAN768B utilizes a vertical chassis and measures 14U tall. The chassis has 12 slots. Four slots are utilized for two control processor (CP) blades and two core switching blades. The eight remaining slots can be used for various blade types. The SAN384B utilizes a horizontal chassis and measures 8U plus a 1U tall exhaust shelf. The chassis has 8 slots. Four slots are utilized for two control processor (CP) blades and two core switching blades. The four remaining slots can be used for various blade types. The blade slots in both platforms can be filled with: 16-port, 32-port, or 48-port 8 Gbps blades Six-port 10 Gbps blade for dark fiber connections (such as ISLs) An encryption blade Routing/extension blade, if using FCIP A fabric application blade A future FCoE/CEE blade The SAN768B provides 3.072 Tbps chassis bandwidth based on 384 ports running at 8 Gbps (link data rate). If all 384 ports are running at 8 Gbps and that traffic travels over the backplane, the oversubscription ratio is 1.5 : 1. However, any ingress/egress traffic moving through port groups on the same blade does not cross the backplane (local switching), thereby eliminating oversubscription. The SAN384B provides 1.536 Tbps of chassis bandwidth based on 192 ports running at 8 Gbps (link data rate). If all 192 ports are running at 8 Gbps and that traffic travels over the backplane, the oversubscription ratio is 1.5 : 1. However, any ingress/egress traffic moving through port groups on the same blade does not cross the backplane (local switching), thereby eliminating oversubscription. Table 3-2 on page 46 through Table 3-6 on page 47 list the component options available with the SAN768B and SAN384B FICON Directors.
Chapter 3. FICON Director technical description

45

Table 3-2 lists the available blade options and the associated feature codes.
Table 3-2 SAN768B and SAN384B blade options Blade options FC 8 Gbps 16-portsa FC 8 Gbps 32-portsa FC 8 Gbps 48-portsa (does not support loop devices) FC 10 Gbps 6-ports FC Routing (16 + 2 ports) Feature code 3816 3832 3848 3870 3850

a. All ports must be populated with SFPs (each with a minimum of 1 Gbps and a maximum of 8 Gbps).

Table 3-3 lists the available transceiver options and the associated feature codes.
Table 3-3 SAN768B and SAN384B transceiver options Transceiver options a 4 Gbps Short Wave SFP Transceiver 4 Gbps WSW 8-Pack SFP Transceiver 4 Gbps 10 km Long Wave SFP Transceiver 4 Gbps 10 km LW 8-Pack SFP Transceiver 4 Gbps 4 km Long Wave SFP Transceiver 4 Gbps 4 km LW SFP Transceiver 8 Pack 4 Gbps 30 km ELW SFP Transceiver 10 Gbps FC SW XFP Transceiver 10 Gbps FC LW XFP Transceiver 1 GbE Copper SFP Transceiver 8 Gbps Short Wave SFP Transceiver 8 Gbps SW SFP Transceiver 8 Pack 8 Gbps 10 km Long Wave SFP Transceiver 8 Gbps 10 km LW SFP Transceiver 8-Pack 8 Gbps 25 km ELW SFP Transceiver a. Maximum of 48 eight-packs or 255 singles Feature code 2401 2408 2411 2418 2441 2448 2480 2510 2520 2550 2801 2808 2821 2828 2881

Table 3-4 lists the available licensed and miscellaneous options and associated feature codes.
Table 3-4 SAN768B and SAN384B licensed and miscellaneous options Licensed and miscellaneous options SAN 768B Inter-Chassis Cable Kit SAN768B Pair of Upgrade Power Supplies Feature code 7870 7880

46

FICON Planning and Implementation Guide

Licensed and miscellaneous options SAN768B Inter-Chassis License SAN768B FICON w/CUP Activation FCIP/FC High-Performance Extension FICON Accelerator Integrated Routing

Feature code 7885 7886 7887 7888 7889

Table 3-5 lists the available fiber optic options and associated feature codes.
Table 3-5 SAN768B and SAN384B fiber optic cable options Fiber optic cable options a Fiber Cable LC/LC 5 meter 50 m multimode Fiber Cable LC/LC 25 meter 50 m multimode Fiber Cable LC/LC 5 meter 50 m multmode 4-Pack Fiber Cable LC/LC 25 meter 50 m multimode 4-Pack Fiber Cable LC/LC 31 meter 9 m single mode 4-Pack Fiber Cable LC/LC 31 meter 9 m single mode a. Maximum of 96 four-packs or 255 singles Feature code 5305 5325 5405 5425 5444 5731

Table 3-6 lists the available rack mounting options and associated feature codes.
Table 3-6 SAN768B and SAN384B rack mounting options Rack mounting options First SAN768B installed in a 2109-C36 rack Second SAB768B installed in a 2109-C36 rack Standalone mode Feature code 9281 9282 9284

3.2.2 IBM System Storage SAN b-type family functions

This section discusses platform-specific functions that are available on System Storage SAN b-type family Directors for use in FICON environments. The platform-specific functions complement the functions that are described in 3.1.4, Basic functions of a FICON Director on page 37.

IBM System Storage Data Center Fabric Manager

IBM System Storage Data Center Fabric Manager (DCFM) is a comprehensive network management application that enables end-to-end management of data center fabrics. DCFM features easy-to-use administrative tools that streamline and automate repetitive tasks. DCFM is designed for unified management of data center fabrics, from ports on storage devices all the way to the FICON channel of the System z server. It can be used to configure System Storage SAN b-type family Directors.

Chapter 3. FICON Director technical description

47

DCFM provides multiprotocol networking support for: Fibre Channel Fiber Connectivity (FICON) Fibre Channel over IP (FCIP) Fibre Channel Routing (FCR) Internet SCSI (iSCSI)

Inter-Chassis Link
An Inter-Chassis Link (ICL) is a licensed feature used to interconnect two System Storage SAN b-type family Directors. ICL ports in the core blades are used to interconnect the two Directors, potentially increasing the number of usable ports. The ICL ports on are internally managed as E_Ports and use proprietary connectors instead of traditional SFPs. When two System Storage SAN b-type family Directors are interconnected by ICLs, each chassis still requires a unique Domain ID and is managed as a separate Director.

Traffic Isolation
Traffic Isolation (TI) zoning allows you to direct traffic to certain paths in cascaded FICON Director environments. For this reason, ISLs (E_Ports) must be included in the TI zone. TI zones are a logical AND with classic zoning (port or WWNN zoning). Note that TI zoning does not replace port or WWNN zoning. TI zoning has two modes of operation: Failover enabled With failover enabled, traffic will take the path specified by the TI zone regardless of other paths that are available. If all paths in the TI zone fail, then traffic will rerouted to another TI zone. Failover disabled With failover disabled, true traffic isolation is achieved. Traffic will only take the path specified by the TI zone. If all paths in the TI zone fail, traffic stops and will not be rerouted to another path even if other paths are available outside the TI zone.

Adaptive Networking
Adaptive Networking is a suite of tools and capabilities that enable you to ensure optimized behavior in the SAN. Even under the worst congestion conditions, the Adaptive Networking features can maximize fabric behavior and provide necessary bandwidth for high priority, mission-critical applications and connections. The following features are part of the Adaptive Networking suite: Traffic Isolation Routing QoS Ingress Rate Limiting QoS SID/DID Traffic Prioritization

Traffic Isolation Routing

The Traffic Isolation Routing feature allows you to control the flow of interswitch traffic by creating a dedicated path for traffic flowing from a specific set of source ports (N_Ports). It can be used, for example, to dedicate an ISL to high priority, host-to-target traffic. Traffic isolation is implemented using a Traffic Isolation zone (TI zone).

48

FICON Planning and Implementation Guide

Quality of Service feature

Quality of Service (QoS) is a feature in Adaptive Networking advanced fabric services that enables users to set the desired QoS for applications running on both physical and virtual servers. A High QoS setting gives critical applications the required bandwidth priority, maximum availability, and optimal performance in the event of network congestion. Fabrics can automatically and continuously adapt to changing network dynamics to ensure that critical applications always receive high priority.

Top Talkers in Advanced Performance Monitoring feature

The Top Talkers feature is part of the licensed Advanced Performance Monitoring feature. It measures and ranks top bandwidth-consuming traffic in real time based on SID/DID pairs. Top Talkers can identify the ports that consume the most bandwidth and that information can be used to assist you with configuring the SID/DID pairs with certain QoS attributes so they get proper priority.

Ingress rate limiting feature

Ingress rate limiting is a licensed feature that requires the Adaptive Networking license. It can restrict the speed of traffic from a particular device to the Director port. Ingress rate limiting can also be used to reduce existing congestion in the network or proactively avoid congestion.

Frame-level trunking
Frame-level trunking automatically distributes data flows over multiple physical Inter-Switch Link (ISL) connections and logically combines them into a trunk to provide full bandwidth utilization while reducing congestion. Frame-level trunking can: Optimize link usage by evenly distributing traffic across all ISLs at the frame level Maintain in-order delivery to ensure data reliability Help ensure reliability and availability even if a link in the trunk fails Optimize fabric-wide performance and load balancing with Dynamic Path Selection (DPS) Simplify management by reducing the number of ISLs required Provide a high-performance solution for network- and data-intensive applications Up to eight ISLs can be combined into a trunk, providing up to 68 Gbps data transfers (with 8 Gbps ISLs).

Virtual ISL connection

When a base switch is connected to another base switch, a Virtual ISL (XISL) connection is created. When logical switches with the same fabric ID (FID) are configured to use the XISL (such as the logical switches #1 and #2), the base switches automatically create a Logical ISL (LISL) within the XISL. The LISL isolates traffic from multiple fabrics, because each LISL is dedicated to traffic for a single fabric.

Lossless Dynamic Load Sharing

Lossless Dynamic Load Sharing (DLS) allows you to rebalance trunk port paths without causing input/output (I/O) failures in cases where the end devices require in-order delivery (IOD) of frames. You can use DLS on trunks connecting switches to perform the following functions: Eliminate dropped frames and I/O failures by rebalancing the paths going over the ISLs whenever there is a fabric event that might result in sub-optimal utilization of the ISLs Eliminate the delay caused by establishing a new path when a topology change occurs Pause ingress traffic (by not returning credits)

Chapter 3. FICON Director technical description

49

Change the existing path to a more optimal path Wait for sufficient time for frames already received to be transmitted Resume traffic after a failure

Virtual fabrics or logically partition a physical SAN

The Virtual Fabrics feature enables a physical storage area network (SAN) switch to be partitioned into multiple logical switches, each of which is assigned to a separate logical fabric. A logical switch and logical fabric are independent entities that are managed just like a physical switch or fabric. The Virtual Fabrics feature provides data, control, and management isolation for logical switches and logical fabrics using a standards-based implementation of the ANSI standard for virtual fabrics. Vitalization supports the following use cases: Isolating traffic types Pooling resources for independent fabrics Isolating different application tiers and device management Facilitating backbone edge routing Increasing FICON consolidation and scalability Connect logical and unaware switches to form logical fabrics Isolate and manage by application, business group, customer, or traffic type (FC, FICON) Table 3-7 lists the logical switch types supported by the Virtual Fabrics feature.
Table 3-7 Logical switch types Switch type Base switch Description Not supported for FICON. The base switch can only contain ISLs that are shared by other logical switches. Although not supported for FICON, the base switch can be used for other logical switches. The default switch is FID = 128 and it is where all ports are placed when virtual fabrics are enabled. Virtual fabrics are enabled by default for all switches shipped from the factory with FOS 6.2.0 and above. Only port-based numbering is permitted in the default switch. Zero-based port numbering is not supported. As a result, the default switch is usually not used for FICON. The logical switch is the most common switch type used for FICON. A logical switch can have any user-defined FID between 0-127. Logical switches support both port-based and zero-based port numbering.

Default switch

Logical switch

Port numbering
A port number is a number assigned to an external port to give it a unique identifier to the FICON Director. Ports are identified by both the slot number in which the blade is located in the chassis and the port number on the blade (slot number/port number). Table 3-8 shows the port numbering scheme for each blade type.
Table 3-8 Port numbering scheme Blade type 16-port 32-port Blade port numbering Ports are numbered from 0 through 15 from bottom to top. Ports are numbered from 0 through 15 from bottom to top on the left set of ports, and 16 through 31 from bottom to top on the right set of ports.

50

FICON Planning and Implementation Guide

Blade type 48-port 6-port

Blade port numbering Ports are numbered from 0 through 23 from bottom to top on the left set of ports, and 24 through 47 from bottom to top on the right set of ports. Ports are numbered from 0 through 5 from bottom to top.

Any port can be assigned to any logical or default switch. When zero-based addressing is used, ports can have any number the user defines. Some forward planning, however, can simplify the task of numbering the ports and minimize service impacts.

Migration
The combination of logical switches and zero-based addressing make for easy migration from IBM System Storage SAN b-type and m-type FICON Directors to a SAN768B or SAN384B platform. For each FICON Director to be migrated, a logical switch with its Domain ID and port numbers identical to those in the old FICON Director can be created. Therefore, you can perform a migration without making any changes to IOCP or HCD. Table 3-9 lists some of the functions and their associated hardware, software, and firmware prerequisites that are discussed in this section for the SAN768B and SAN384B.
Table 3-9 Hardware, software, and firmware prerequisites Function Data Center Fabric Manager Hardware prerequisites All supported Directors Software and firmware prerequisites Refer to: ftp://ftp.software.ibm.com/common/ ssi/pm/sp/n/tsd03064usen/TSD03064U SEN.PDF FOS 6.1 FOS 6.0.0c and above FOS 6.2 FOS 6.0.0c and above FOS 6.0.0c

Inter-Chassis Link (ICL) Traffic Isolation (TI) Virtual Fabric (logical partition) Quality of Service (QoS) Adaptive Network TopTalker, QoS, Ingress Rate Limiting

SAN768B or SAN384B SAN768B or SAN384B SAN768B or SAN384B SAN768B or SAN384B SAN768B or SAN384B

For more details regarding the features and functions provided by the System Storage SAN b-type family, refer to the following: IBM System Storage SAN768B Director Installation, Service and User's Guide, GA32-0574 IBM System Storage SAN384B Director Installation, Service and User's Guide, GC52-1333 Brocade Fabric OS Administrators Guide, 53-1001185 IBM System Storage SAN768B Web page: http://www-03.ibm.com/systems/storage/san/b-type/san768b/index.html IBM System Storage SAN384B Web page: http://www-03.ibm.com/systems/storage/san/b-type/san384b/index.html

Chapter 3. FICON Director technical description

51

3.2.3 Cisco MDS 9500 Series

The Cisco MDS 9500 Series is a modular multi-layer platform with common switching and service modules. The Cisco MDS 9513 has 11 slots for various modules and can switch up to 528 ports simultaneously. The Cisco MDS 9509 has 7 slots for various modules and can switch up to 336 ports simultaneously. The Cisco MDS 9506 has 4 slots for various modules and can simultaneously switch up to 192 ports. The Cisco MDS 9500 Series utilizes a vertical chassis. The Cisco MDS 9513 and MDS 9509 with their dual Supervisor-2 modules measure 14U tall. The Cisco MDS 9506 has dual Supervisor modules and measures 7U tall. The module slots in all three platforms can be filled with: IP services, storage services, and multi-service modules 12-port, 24-port, and 48-port 4 Gbps switching modules 24-port, 48-port, and 4/44-port 8 Gbps switching modules 4-port 10 Gbps switching module When all 48 ports on a 48-port switching module are used simultaneously at 8 Gbps, the oversubscription factor could be as much as 4 : 1. On a 24-port switching, when all 24-ports module are used simultaneously at 8 Gbps, the oversubscription factor could be as much as 2 : 1. Table 3-10 through Table 3-14 on page 53 list the component options available with the Cisco MDS 9500 Series FICON Directors. Table 3-10 lists the available line card options and associated feature codes.
Table 3-10 Cisco MDS 9500 Series line card options Line card options 4 port 10 Gbps FC Switch Module 12 port 4 Gbps FC Switch Module 24 port 4 Gbps FC Switch Module 48 port 4 Gbps FC Switch Module 18/4 Multservice Module (requires OS 3.2.2 or later) 24 8 Gbps FC Switching Module (requires OS 4.1.1 or later) 4/44 port 8 Gbps Host Switching Module (requires OS 4.1.1 or later) 48 port 8 Gbps FC Switching Module (requires OS 4.1.1 or later) Feature code 2404 2412 2424 2448 2450 2824 2844 2848

Table 3-11 lists the available transceiver options and associated feature codes.
Table 3-11 Cisco MDS 9500 Series transceiver Transceiver options FC 10 Gbps SRX2 Transceiver FC Ethernet 10 Gbps SRX2 Transceiver FC 10 Gbps 10 km LWX2 SC Transceiver FC 10Gbps 40 km ERX2 Transceiver Feature code 5030 5032 5040 5050

52

FICON Planning and Implementation Guide

Transceiver options FC 4 Gbps SW SFP Transceiver - 4 Pack FC 4 Gbps 4 km LW SFP Transceiver 4 Pack FC 4 Gbps 10 km LW SFP Transceiver - 4 Pack FC 8 Gbps SW SFP+ Transceiver (requires OS 4.1.1 or later) FC 8 Gbps SW SFP+ Transceiver - 4 Pack (requires OS 4.1.1 or later) FC 8 Gbps 10 km LW SFP+ Transceiver (requires OS 4.1.1 or later) FC 8 Gbps 10 km LW SFP+ Transceiver - 4 Pack (requires OS 4.1.1 or later) Tri-Rate SW SFP Transceiver Tri-Rate LW SFP Transceiver Gigabit Ethernet Copper SFP (feature code 2450 is required)

Feature code 5434 5444 5454 5830 5834 5850 5854 5210 5220 5250

Table 3-12 lists the available hardware and software packages and associated feature codes.
Table 3-12 Cisco MDS 9500 Series hardware and software packages Hardware and software packages MDS 9500 Enterprise Package MDS 9500 Fabric Manager Server Package MDS 9500 Mainframe Server Package Feature code 7021 7026 7036

Table 3-13 lists the available fiber optic cable options and associated feature codes.
Table 3-13 Cisco MDS 9500 Series fiber optic cable options Fiber optic cable options 5 meter 50 m LC/LC Fiber Cable (multimode) 25 meter 50 m LC/LC Fiber Cable (multimode) 5 meter 50 m LC/LC Fiber Cable - 4 Pack 25 meter 50 m LC/LC Fiber Cable - 4 Pack Feature code 5605 5625 5642 5643

Table 3-14 lists the available rack mounting options and associated feature codes.
Table 3-14 Cisco MDS 9500 Series rack mounting options Rack mounting options Plant Install 9513 in 2109-C36 rack Field Merge 9513 in 2109-C36 rack Feature code 9543 9544

3.2.4 Functions of the Cisco MDS 9500 Series

This section discusses platform-specific functions that are available on the Cisco MDS 9500 Series of Directors for use in FICON environments. The platform-specific functions complement the functions that are described in 3.1.4, Basic functions of a FICON Director on page 37.
Chapter 3. FICON Director technical description

53

Fabric Manager
The Fabric Manager is a set of network management tools that supports secure Simple Network Management Protocol version 3 (SNMPv3) and earlier versions. It provides a graphical user interface (GUI) that displays real-time views of your network fabric and lets you manage the configuration of Cisco MDS 9500 Series FICON Directors. Detailed traffic analysis is also provided by capturing data with SNMP. The captured data is compiled into various graphs and charts that can be viewed with any Web browser. The Cisco Fabric Manager applications are: Fabric Manager Server Device Manager

Fabric Manager Server

Fabric Manager Server is a separately licensed feature that extends the capabilities of the basic Fabric Manager. It adds the ability for Fabric Manager to manage multiple fabrics simultaneously (note that in Fibre Channel terms, each Director in switched point-to-point or cascaded FICON Director topology, is a separate fabric). It also enables accumulation and analysis of historical performance data, which can be a valuable complement to RMF.

Device Manager
The Device Manager presents two views of a Director: 1. Device View displays a graphic representation of the switch configuration, and provides access to statistics and configuration information for a single switch. 2. Summary View displays a summary of E_Ports (Inter-Switch Links), F_Ports (fabric ports), and N_Ports (attached hosts and storage) on the Director.

Comprehensive security framework

The Comprehensive security framework supports RADIUS authentication, Simple Network Management Protocol Version 3 (SNMPv3), role-based access control, Secure Shell Protocol (SSH), Secure File Transfer Protocol (SFTP), Fibre Channel Security Protocol (FC-SP), VSANs, hardware-enforced zoning, port security, fabric binding and VSAN-based access control.

PortChannel (ISL Trunking)

PortChannels refer to the aggregation of multiple physical interfaces into one logical interface to provide higher aggregated bandwidth, load balancing, and link redundancy. PortChannels can connect to interfaces across switching modules, so a failure of a switching module cannot bring down the PortChannel link.

Trunk E port (TE port)

Trunking, also known as VSAN trunking, is a feature specific to switches in the Cisco MDS 9500 Series. Trunking enables interconnect ports to transmit and receive frames in more than one VSAN, over the same physical link, using enhanced ISL (EISL) frame format. The trunking feature includes the following: Trunking configurations are applicable to E_Ports. If trunk mode is enabled on an E_Port and that port becomes operational, it is referred to as a TE port. The trunk-allowed VSANs configured for TE ports are used by the trunking protocol to determine the allowed-active VSANs in which frames can be received or transmitted. If a trunking enabled E_Port is connected to a third-party switch, the trunking protocol ensures seamless operation as an E_Port

54

FICON Planning and Implementation Guide

Quality of Service (QoS)

The QoS feature in Cisco SAN-OS allows for data traffic to be classified into distinct levels for service differentiation ( low, medium, or high priority). Quality of service (QoS) offers the following: Provides a relative bandwidth guarantee to application traffic Controls the latency experienced by application traffic Prioritizes one application over another (for example, prioritizing transactional traffic over bulk traffic) through bandwidth and latency differentiation You can apply QoS to ensure that Fibre Channel data traffic for your latency-sensitive applications receive higher priority over throughput-intensive applications such as data warehousing.

Virtual Storage Area Network

Virtual Storage Area Network (SAN) technology is used for hardware-enforced, isolated environments within a single physical fabric for secure sharing of physical infrastructure and enhanced FICON intermix support. The Cisco MDS 9500 Series can have up to 256 unique VSANs, and eight of those can be used for FICON. Each VSAN has its own set of fabric services (fabric server, name server, and so on), FICON Control Unit Port, Domain ID, Fabric Shortest Path First (FSPF) routing, operating mode, IP address, and security profile. VSANs can span line cards and are dynamic in size. For example, one VSAN with 10 ports can span 10 different line cards. Adding ports to a VSAN is a nondisruptive process. The maximum number of ports for a VSAN is 255 due to FICON addressing rules. VSANs enable complete traffic isolation and are maintained throughout the FICON fabric. You can also assign specific VSANs to certain cascaded links.

VSAN-Based Access Control

This feature enables customers to define roles where the scope of the roles is limited to certain VSANs. For example, a network administrator role can be set up to allow configuration of all platform-specific capabilities, and VSAN-administrator roles can be set up to only allow configuration and management of specific VSANs. VSAN-based access control reduces SAN disruptions by localizing the effects of user errors to the VSANs for which the user has administrative privileges. It adds a layer of security where only administrators can configure switches within specified VSANs.

Port numbering
A range of 250 port numbers is available for you to assign to all the ports on a Director. You can have more than 250 physical ports assigned and the excess ports do not have port numbers in the default numbering scheme. You can have ports without a port number assigned if they are not in a FICON VSAN, or you can assign duplicate port numbers if they are not used in the same FICON VSAN. By default, port numbers are the same as port addresses, and the port addresses can be swapped. The following rules apply to FICON port numbers: Supervisor modules do not have port number assignments. Port numbers do not change based on TE ports. Because TE ports appear in multiple VSANs, chassis-wide unique port numbers should be reserved for TE ports.
Chapter 3. FICON Director technical description

55

Each PortChannel must be explicitly associated with a FICON port number. When the port number for a physical PortChannel becomes uninstalled, the relevant PortChannel configuration is applied to the physical port. Each FCIP tunnel must be explicitly associated with a FICON port number. If the port numbers are not assigned for PortChannels or for FCIP tunnels, then the associated ports will not come up.

Migration
The combination of VSANs and FICON port numbers make for easy migration from previous-generation FICON Directors to a Cisco MDS 9500 infrastructure. For each FICON Director to be migrated, a FICON VSAN with its Domain ID and FICON port numbers identical to those in the old FICON Director can be created. Therefore, you can perform a migration without making any changes to IOCP or HCD. For more details regarding the features and functions provided by the Cisco MDS 9500 Series, refer to the following: Cisco MDS 9513 for IBM System Storage http://www-03.ibm.com/systems/storage/san/ctype/9513/index.html Cisco MDS 9509 for IBM System Storage http://www-03.ibm.com/systems/storage/san/ctype/9509/index.html Cisco MDS 9506 for IBM System Storage http://www-03.ibm.com/systems/storage/san/ctype/9506/index.html For additional Cisco MDS 9500 Series-related information, go to: http://www.cisco.com/en/US/partner/products/ps5990/tsd_products_support_series_ home.html

56

FICON Planning and Implementation Guide

Part 2

Part

Planning the FICON environment

This part provides an overview of the items that you need to consider when planning the FICON environment. It provides a structured approach for planning and designing an end-to-end FICON infrastructure solution. It also gives a solid base for configuring the components of the solution during the implementation phase.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

57

58

FICON Planning and Implementation Guide

Chapter 4.

Planning the FICON environment

When you are building a communication infrastructure with the highest quality of service for a FICON environment, proper planning is crucial. This chapter provides you with a structured approach for planning and designing an end-to-end FICON infrastructure solution. It gives a solid base for configuring the components of the solution during the implementation phase.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

59

4.1 Structured approach for planning

Having a structured approach for planning is key to reaching your desired goal when you are designing any end-to-end solution. The approach used in this chapter offers a framework that is based on best practices and industry standards. The topics we cover follow a logical sequence that aids in the decision-making process for planning a FICON environment. The decision points are listed here and described in greater detail in the following sections: Step 1 - Documentation Step 2 - Requirements Step 3 - Context Step 4 - Topologies and supported distances Step 5 - Convergence Step 6 - Management Step 7 - Virtualization and availability Step 8 - Performance Step 9 - Prerequisites and interoperability Step 10 - Physical connectivity If you are not familiar with Fibre Channel and FICON architectures and terminology, then before proceeding we strongly recommend that you read Chapter 1, Introduction to FICON on page 3; Chapter 2, System z FICON technical description on page 13; and Chapter 3, FICON Director technical description on page 33. In addition, to be successful in designing a FICON solution, you must be well versed in the technologies and products that support typical FICON solutions. Although this book provides some guidance, it is not meant to provide in-depth technology and product knowledge. Therefore, we recommend that you also attend the appropriate training and education offered by IBM as described at the following URL: http://www-304.ibm.com/jct03001c/services/learning/ites.wss/zz/en?pageType=page&c= a0001742

Step 1 - Documentation
Current and accurate documentation (planning, design, installation, and final) is essential for a successful implementation and later operations of the FICON environment. However, because most information is not always readily available during the various phases of the planning process, there will be situations where it will require iterations of the process. To avoid delays during the planning and implementation process, ensure that the various types of documentation needed are always well-defined, complete, and current.

Step 2 - Requirements
Planning for any end-to-end solution is a process of iterations. It starts with gathering requirements and defining the desired outcome of the solution. After all relevant data has been collected, you will need to review and evaluate it, as well as outline the objectives of the solution. If you are migrating from an existing environment, you must also consider the current configuration, including the functions and features that are implemented and how they are being used.

60

FICON Planning and Implementation Guide

Step 3 - Context
The output of the requirements-gathering phase should provide a solid base for mapping to one or more of the following scenarios: Migrating from ESCON to FICON Moving to an 8 Gbps environment (FICON Express8) Migrating from a single site to a multi-site environment Implementing a new FICON environment

Step 4 - Topologies and supported distances

To fulfill the requirements of the chosen solution (or scenario), you must select the most suitable topology. The total end-to-end distance (from a channel to a control unit) is a main factor in determining the correct topology. The following topologies are supported: Point-to-point Switched point-to-point Cascaded FICON Directors For extended distances, you must also evaluate the IBM qualified solutions depending on your requirements and the supported distances of the FICON connections. Note that IBM owns a strict qualification process for giving assurance of an extended distance solution based on WDM technology.

Step 5 - Convergence
The I/O communication infrastructure (FICON fabric) should provide the ability to attain high availability and continuous operations in a predictable manner, with centralized management and scalability for future changes. This means that the switching platform must be robust enough to deliver such capability, as well as be ready to deploy new enhancements that improve throughput, security, and virtualization. Therefore, the FICON Director is the reference platform for FICON fabrics. Factors that influence the selection of the FICON Director platform are: Intermix of the FICON and Fibre Channel Protocol (FCP), which enables you to consolidate the I/O infrastructure, maximize the utilization of the bandwidth, and consolidate the fabric management. Security, high integrity, and zoning are integral parts of the entire FICON infrastructure. Other relevant factors for the FICON fabric, such as power consumption, cooling, and space inside a data center, should also be considered.

Step 6 - Management
The selection of the systems management platforms and their integration with the operation support environment is very important to the continuous operations of the FICON infrastructure. The possibilities for FICON management include the following: Command-line Interface Element management Fabric management Storage management initiative specification System z management for FICON Directors
Chapter 4. Planning the FICON environment

61

Step 7 - Virtualization and availability

Because the efficiency and effectiveness of a data center is determined by the levels of virtualization and availability, the complete infrastructure must be considered. All aspects of virtualization and physical connectivity for the FICON channel, Director, and control unit should be evaluated.

Step 8 - Performance
The selection of the right technology is determined by satisfying traffic patterns, further segmentation, and performance of the end-to-end solution. It is very important to understand the fan-in, fan-out, and oversubscription ratios for the entire FICON environment. Traffic management gives you the possibility to proactively manage any congestion inside the fabric and minimize latency. There are a number of design evaluation tools that can aid in this process.

Step 9 - Prerequisites and interoperability

Checking the compliance of all FICON components in the infrastructure with the prerequisites of the planned solution is key to ensuring that your implementation will proceed without delays and avoid interoperability problems. The FICON components include the System z FICON features and control units, as well as the FICON Directors and Wavelength Division Multiplexor (WDMs) platforms, if applicable to the selected topology.

Step 10 - Physical connectivity

The physical transport media has to be selected according to the installed or planned features. Choosing a qualified cabling solution based on industry standards for Data Center Cabling (US TIA-942/Europe EN50173-5) and best practices in accordance with System z qualified IBM Facilities Cabling Services will provide a secure and highly scalable consolidated physical layer for the complete data center. In the subsequent sections we describe considerations and recommendations for each of the identified decision points in more detail.

4.2 Documentation
Planning, design, installation, and final documentation is a requirement for a successful implementation and later operations of the FICON environment. Based on best practices and requirements we recommend the following types of documentation:

Planning documentation
Current environment Requirements Available features and functions Decision points

Design documentation
High level Detailed level

Installation documentation
Detailed implementation plan

62

FICON Planning and Implementation Guide

Final documentation
Technical Operations Disaster recovery plan Change procedure The number of documents that you must create depends on the complexity of the solution and the required level of detail. Creating and maintaining documentation throughout the planning, design, and implementation phases of your FICON infrastructure is an important part of running the environment without unplanned outages. For the most part, the implementation, operations, and support (troubleshooting) responsibilities are owned by different groups. From a security viewpoint, there is usually a a clear boundary between the responsibilities. Changes have to be monitored and audited to ensure that due diligence is observed; for example, execution must be strictly separated from auditing. And finally, you have to consider the planned and unplanned turnover of professionals at the workplace. To reduce the risk of disruption to your FICON infrastructure, keep your documentation as current and accurate as possible and store a copy in a secure place, preferably in a different location in case of a disaster.

4.3 Requirements
It is important to clearly understand and accurately document all requirements. Such documentation will help you throughout the planning and design process. After the requirements are collected and documented, each one will have to be carefully evaluated. For existing environments (FICON or ESCON), it is also important to identify all equipment that is currently installed and how it is being used. This means physical and logical inventories should be carried out. The goal of the physical inventory is to identify and verify what is installed in your FICON or ESCON environment. Through onsite visual inspections and documentation, all ESCON or FICON channels, Directors and ports, control units, and cabling should be identified. Although the overall goal of the physical inventory is to identify and verify what you have physically installed, the goal of the logical inventory is to understand the functions that are being used and how they are defined. Determine whether the functions must be replaced by a FICON function. During the planning and design process, the requirements will be mapped to features, functions, topology, and technologies, which in turn will determine an approximate design of your FICON environment. This does not mean you will only have one solution, but rather multiple alternatives. The alternatives will have to be assessed using all the steps in this chapter.

4.4 Context
To define the context of your environment, you must take into consideration the existing and planned components (System z, FICON Director, and control units) that were determined and
Chapter 4. Planning the FICON environment

63

documented in the requirements gathering phase. You will then have to define your transport layer strategy dependent on the upper layer requirements for high availability, disaster recovery, and business continuity. This all has to be synchronized with the communication layer requirements of your applications. After all relevant data is reviewed and analyzed, the output (alternatives) can be used to map to one or more of the following scenarios: Migrating from ESCON to FICON (changing the transport layer protocol) Moving to a high bandwidth environment (implementing FICON Express8) Migrating from a single site to a multi-site cascaded topology (building a highly available multi-site solution) Implementing a new FICON environment (building a completely new FICON environment) The following sections provide addition information for each scenario.

4.4.1 Migrating from ESCON to FICON

This scenario discusses the considerations for migration from an ESCON environment to a FICON environment. We concentrate on DASD migration, because it is the most commonly connected device type. FICON is addressing technical limitations of ESCON in bandwidth, distances, topology, channel and device addressing and performance. For the right number of FICON channels you need: The I/O operations per second and their characteristics The required bandwidth ESCON channels are planned with 50% utilization or less and depend on block size. With higher utilization, the traffic will be slower. Another limit is the I/O rate, which is typically less than 1000 I/O per second. The ESCON data droop is caused by the single CCW transmit at a time, and then waiting for the notification. We do not recommend planning over the actual limits, because that can cause performance degradation. It is also important to have enough channels to keep the number of concurrent I/Os low. You also have to determine the peak I/O and MB/s values for each subsystem. This reflects current workload and planned growth. We recommend that you use analysis tools over a longer time period to obtain realistic trends. For native tape, throughput is very important. Aggregating multiple ESCON tape channels onto a single FICON channel reduces the System z infrastructure. This can be reached due to high head-to-tape transfer rates, capacity, and accessibility of the cartridge. Using FICON for virtual tapes also reduces the number of required CUs, back-end tape drives, and channel paths. When planning the migration from an existing ESCON configuration to a FICON configuration, consider the following rules: Configure at least two channel paths to a logical control unit for high availability. (More channel paths may be required, depending on throughput requirements.) A logical control unit or device cannot be defined to be accessed more than once from the same channel path.

64

FICON Planning and Implementation Guide

A physical control unit that has multiple logical control units (specified by the CUADD parameter in IOCP/HCD) may be accessed more than once from the same FICON (FC) channel path, but the access is to different CUADDs (different logical control units) within the physical control unit. Configure the channel paths according to the quantity of resources available in the FICON channel and control unit. It is not possible to aggregate two or more ESCON channel paths that access the same logical control unit into only one FICON channel path. Even if there are the same number of paths from the operating system image to the disk subsystem in both ESCON and FICON configurations, the advantages of using FICON include: More concurrent I/Os to the same control unit Concurrent I/Os with other control units Link data rate (200 Mbps for ESCON versus up to 8 Gbps for FICON) Unrepeated distance from channel (3 km for ESCON and 10 km for FICON)

4.4.2 Moving to a high bandwidth environment (FICON Express8)

Moving to 8 Gbps FICON improves the performance, scalability, and efficiency of your System z and storage environment. With the use of a FICON Director you can utilize the physical channels even more when shared with multiple connections to different CUs at the same time, while improving performance. With an 8 Gbps FICON environment, you need to address the following requirements: Storage growth - number of ports and bandwidth Fabric growth - number of ISLs and scalability Performance - bandwidth and cabling infrastructure Server virtualization - I/O throughput and scalability Services - end-to-end management Backup - bandwidth Operational flexibility - scalability and efficiency Compatibility - end-to-end management and lifetime There are two System z qualified FICON Director platforms that support 8 Gbps: the IBM System Storage b-type family and the Cisco MDS 9500 Series. Chapter 3, FICON Director technical description on page 33, explains the differences in proprietary implementations for each product line. Some benefits of using 8 Gbps FICON include: The ISL oversubscription ratio is halved by upgrading from 4 Gbps to 8 Gbps ISLs. At long distances, 8 Gbps compared to lower speeds optimizes the number of dark fiber or WDM links. Building a high-performance infrastructure that provides the flexibility to deploy 8 Gbps as needed simplifies the fabric management and scales with the growth of applications and data. We recommend that you do not mix low speed and high speed devices on the same 8 Gbps channel. The physical layer up to 4 Gbps link data rate had different limitations as compared to 8 Gbps and 10 Gbps. Also, the 8 Gbps features only auto-negotiate to speeds of either 8 Gbps, 4 Gbps, or 2 Gbps. We also recommend that you review the capabilities of the existing cabling infrastructure.
Chapter 4. Planning the FICON environment

65

For more detailed information, refer to: Chapter 2, System z FICON technical description on page 13 Chapter 3, FICON Director technical description on page 33 Implementing an IBM/Brocade SAN with 8 Gbps Directors and Switches, SG24-6116 IBM System Storage SAN768B, TSD03037USEN Cisco MDS 9506 for IBM System Storage, TSD00069USEN Cisco MDS 9509 for IBM System Storage, TSD00070USEN Cisco MDS 9513 for IBM System Storage, TSD01754USEN

4.4.3 Migrating from a single site to a multi-site environment

Migrating from a single site to a multi-site configuration usually means going from a point-to-point or switched point-to-point topology to a cascaded topology. Cascaded support is important for disaster recovery and business continuity solutions. It can provide high availability connectivity, as well as the potential for fiber infrastructure cost savings for extended storage networks. Solutions such as Geographically Dispersed Parallel Sysplex (GDPS) can benefit from the reduced intersite configuration complexity that FICON support of cascaded Directors provides. Cascaded FICON Directors allow for FICON channels to connect a System z server to another, similar server or to peripheral devices (such as disk, tape library, or printer) via two FICON Directors. In cascaded environments, the cabling for both intersite and intrasite can be simplified. This decreases the management complexity. It also reduces the number of FICON channels, Director ports, and storage devices. The sharing of links between the two sites reduces the number of physical channels between sites and the number of Director ports. The faster the channel speeds between sites, the better the intersite throughput. With higher speed this factor can even multiply. FICON Directors typically have more buffer credits per port than the server and the disk, or the tape subsystems. The more buffer credits you have, the greater distances you can go without impacting the response times. The following factors affect the performance of a cascaded FICON Director configuration: Number of ISLs and the routing policy Number of FICON channels routed across the ISLs ISL bandwidth ISL traffic management I/O workload characteristics Distances between the components Buffer-to-buffer credits FICON support of cascaded Directors (sometimes referred to as cascaded switching or two-switch cascaded fabric) is for single-vendor fabrics only.

4.4.4 Implementing a new FICON environment

When you are defining a new FICON environment, we recommend that you take advantage of the latest FICON functions and features that support high performance and high availability. For planning this scenario you must also consider criteria such as: Facility requirements - floor space, power, and cooling Scalability and flexibility - ability to add ports and bandwidth dynamically 66
FICON Planning and Implementation Guide

Performance - I/O throughput and traffic patterns Availability - redundant ports, bandwidth, and logical paths End-to-end management Cabling infrastructure Distances between the components Control unit and device capabilities Interoperability of all components The remaining sections in this chapter help you define a baseline for your FICON infrastructure, using best practices and industry standards that also align with vendor recommendations.

4.5 Topologies and supported distances

The unrepeated distance supported by System z FICON features is dependent upon: Transceiver type Fiber optic cabling type Port speed Mode conditioning patch (MCP) cables Cabling specifications are defined by the Fibre Channel - Physical Interface - 4 (FC-PI-4) standard and used by System z FICON features. Table 4-1 identifies cabling types and link data rates that are supported in the FICON environment, including their allowable maximum distances and link loss budget. The link loss budget is derived from the channel insertion loss budget defined by the FC-PI-4 standard (Revision 8.00).
Table 4-1 Fiber optic cabling for FICON - maximum distances and link loss budget FC-PI-4
Fiber core (light source)

1 Gbps
Distance in meters (in feet) Link loss budget in dB

2 Gbps
Distance in meters (in feet) Link loss budget in dB

4 Gbps
Distance in meters (in feet) Link loss budget in dB

8 Gbps
Distance in meters (in feet) Link loss budget in dB

10 Gbps ISLa
Distance in meters (in feet) Link loss budget in dB

9 m SM (10 km LX laser) 9 m SM ( 4 km LX laser) 50 m MMb (SX laser) 50 m MMc (SX laser) 62.5 m MMd (SX laser)

10000 (32736)

7.8

10000 (32736)

7.8

10000 (32736)

7.8

10000 (32736)

6.4

10000 (32736)

6.0

4000 (13200)

4.8

4000 (13200)

4.8

4000 (13200)

4.8

NA

NA

NA

NA

860 (2822) 500 (1640) 300 (984)

4.62

500 (1640) 300 (984) 150 (492)

3.31

380 (1247) 150 (492) 70 (230)

2.88

150 (492) 50 (164) 21 (69)

2.04

300 (984) 82 (269) 33 (108)

2.6

3.85

2.62

2.06

1.68

2.3

3.0

2.1

1.78

1.58

2.4

a. Inter-Switch Link (ISL) between two FICON Directors b. OM3: 50/125 m laser optimized multimode fiber with a minimum overfilled launch bandwidth of 1500 MHz-km at 850nm as well as an effective laser launch bandwidth of 2000 MHz-km at 850 nm in accordance with IEC 60793-2-10 Type A1a.2 fiber c. OM2: 50/125 m multimode fiber with a bandwidth of 500 MHzkm at 850 nm and 500 MHz-km at 1300 nm in accordance with IEC 60793-2-10 Type A1a.1 fiber d. OM1: 62.5/125 m multimode fiber with a minimum overfilled launch bandwidth of 200 MHzkm at 850 nm and 500 MHz-km at 1300 nm in accordance with IEC 60793-2-10 Type A1b fiber

Chapter 4. Planning the FICON environment

67

Note: IBM does not support a mix of 50 m and 62.5 m fiber optic cabling in the same physical link. Refer to 2.2, System z FICON feature support on page 25 for details about the FICON features available on the System z10 and System z9 servers.

4.5.1 Point-to-point
As illustrated in Figure 4-1, the maximum unrepeated distance of a FICON LX link using a single mode fiber optic cable is: 10 km (6.2 miles) for 1, 2, 4, 8 Gbps LX links 20 km (12.43 miles) for 1 Gbps LX links with RPQ 8P2263 12 km (7.46 miles) for 2 Gbps LX links with RPQ 8P2263

FICON link
9 m fiber

FICON CU

System z

10 km >10 km (RPQ 8P2263)

Figure 4-1 FICON LX point-to-point maximum unrepeated distance

4.5.2 Switched point-to-point

As illustrated in Figure 4-2, one FICON Director increases the maximum supported distance of a FICON LX channel path using single mode fiber optic cables to the following distances: 20 km (12.43 miles) for 1, 2, 4, 8 Gbps LX links 30 km (18.64 miles) for 1 Gbps LX links with RPQ 8P2263 22 km (13.67 miles) for 2 Gbps LX links with RPQ 8P2263
FICON Director
FICON CU

FICON link
9 m fiber

FICON links
9 m fiber

FICON CU FICON CU

System z

10 km >10 km (RPQ 8P2263)

10 km

Figure 4-2 FICON LX switched point-to-point maximum unrepeated distance

4.5.3 Cascaded FICON Directors

A FICON channel path may include a maximum of two cascaded FICON Directors. The use of extended distance longwave transceivers on the Inter-switch links (ISL) between the

68

FICON Planning and Implementation Guide

FICON Directors may be required. Each ISL requires one fiber trunk (two fibers) between the FICON Directors. As illustrated in Figure 4-3, and using the example of 10 km ISL links between FICON Directors, the maximum supported distance of a FICON LX channel path using single mode fiber optic cables is: 30 km (18.64 miles) for 1, 2, 4, 8 Gbps LX links 40 km (24.86 miles) for 1 Gbps LX links with RPQ 8P2263 32 km (19.88 miles) for 2 Gbps LX links with RPQ 8P2263
FICON Director FICON Director
FICON CU

FICON link
9 m fiber

ISL
9 m fiber

FICON links
9 m fiber

FICON CU FICON CU

System z

10 km >10 km (RPQ8P2263)

10 km

10 km

Figure 4-3 FICON LX cascaded FICON Directors maximum unrepeated distance

A FICON channel path through one or two FICON Directors consists of multiple optical fiber links. Each link in the channel path can be either long wavelength (LX) or short wavelength (SX), allowing the channel path to be made up of an intermix of link types as illustrated in Figure 4-4. This is possible because the FICON Director performs an optical-to-electrical-to-optical conversion (OEO) of the channel path as it passes through the Director. The transceiver types (LX or SX) at each end of a given link must match.
Site 2 Site 1
FICON Director
SM

LX FICON CU

LX SX LX LX LX
MM
SM

SX FICON CU LX FICON CU

10 km
SM M

LX SX

SM

System z

LX LX SX

LX LX SX LX

SM

LX FICON CU

Figure 4-4 FICON Director media intermix

MM
SM

SX FICON CU LX FICON CU

FICON Director

Chapter 4. Planning the FICON environment

69

4.5.4 Extended distance

The most important element for an extended distance topology is latency. Latency is usually caused by the following: The length of the fiber, because of the time it takes for the light to travel over distance The opto-electro-opto converter (or the WDM equipment), if used This has to be compensated for by the buffer credit technology used inside the interface logic of the FICON Director or the channels at the System z server and the control unit. The repeated distance for a FICON channel is System z-qualified to a maximum of 100 km (62 miles). For all FICON features using repeaters, the end-to-end distance between the FICON channel and the FICON Director port can be up to 100 km. The same end-to-end distance is also available between the FICON Director port and the control unit port. However, the overall end-to-end distance between the FICON channel and control unit is System z-qualified up to 100 km only. Note: The maximum distance of the fiber optic link for a FICON Express8 LX feature to the first hop is limited to 10 km. Using Dense Wavelength Division Multiplexing (DWDM) technology, the FICON SX and LX implementation supports a maximum distance of 100 km between two sites, as illustrated in Figure 4-5.
Site 2
LX FICON CU

DWDM

Site 1
LX

km 100
SM

SM

LX FICON CU

LX SX SX LX
FICON Director
MM

DWDM

SX FICON CU LX FICON CU

SX SX

SM

SX
System z

Figure 4-5 FICON point-to-point and switched point-to-point maximum extended distance

By combining FICON cascaded Director technology with DWDM technology, the FICON implementation supports a maximum distance of 100 km between two sites, as illustrated in Figure 4-6 on page 71. DWDM technology also provides increased flexibility because multiple links and protocol types can be transported over a single dark fiber trunk.

M M
SX FICON CU
SM

LX FICON CU

LX SX SX LX
FICON Director
MM
SM

SX FICON CU LX FICON CU

70

FICON Planning and Implementation Guide

Site 2 Site 1
FICON Director
SM

LX FICON CU

LX SX
LX LX LX LX SX SX SX SX

MM

SX FICON CU LX FICON CU

LX

SM

10 0

DW DM
SM

SX SX

km
DW DM
M M
SX SX

System z

LX LX

LX LX

SX SX

SM

LX FICON CU

SX SX

LX SX LX
MM
SM

SX FICON CU LX FICON CU

FICON Director

Figure 4-6 Cascaded FICON Directors maximum extended distance

For more detailed information, refer to: Planning for Fiber Optic Links (ESCON, FICON, Coupling Links, and Open Systems Adapters), GA23-0367 For the list of IBM qualified extended distance FICON Director solutions, see: http://www-03.ibm.com/systems/storage/san/index.html For the list of IBM qualified extended distance WDM solutions, see: http://resourcelink.ibm.com

4.6 Convergence
Some System z data center fabrics may require the use of a fully integrated multiprotocol infrastructure that supports mixed I/O and traffic types for simultaneous storage connectivity. A converged infrastructure enables you to consolidate multiple transport layers in a single physical interconnect. This consolidation provides the flexibility to build virtual server and storage environments supporting low-latency, high-bandwidth applications, and simplify hardware and cabling. This approach also simplifies management and reduces power consumption, cooling, and space requirements inside a data center. Other relevant factors to consider are: High density port count (small form factor pluggable ports) Best energy effectiveness devices (per transferred byte) High availability (a redundant architecture of all components) High speed (switching capability - backplane throughput) If you selected a point-to-point topology in the previous step, you can skip this section and go to 4.7, Management on page 75.

Chapter 4. Planning the FICON environment

71

4.6.1 Intermix fabric

Following are reasons why you might move to an intermix FICON and FCP environment: Because of an open systems SAN that is delivering reliable, high-performance connectivity; these fabrics often have available ports on their SAN Directors. To consolidate switching and cabling infrastructure, which enables more effective maintenance, power consumption, cooling, and other operational tasks. Possible scenarios include the following: FICON and open systems environments with a common storage network. System z servers accessing remote storage. Linux on System z to access local storage or remote storage. Hardware-based remote disk mirroring. Different System z server LPARs are accessing local or remote storage via FICON or via FCP. Important technical considerations that you must take into account include fabric management techniques such as zoning, partitioning, binding, and other design considerations. First you must consider the management techniques used in each protocol, and the management requirements of the target environment. The management differences between the two protocols are not relevant unless you want to control the scope of the switching through zoning or connectivity control. For example, FCP devices use name server zoning as a way to provide fabric-wide connection control. FICON devices can use the Prohibit Dynamic Connectivity Mask (PDCM) to provide Director-wide connection control. The PDCM is a vector-defined addressing system that establishes which addresses are allowed to communicate with each other. The FICON Director is required to support hardware enforcement of the connectivity control. In intermix environments, this is more restrictive than the zoning information used by open systems devices. It is also important to understand the implications of FICON port addressing versus port numbering in FCP. FICON abstracts the concept of the port by creating an object known as the port address. An association is then made between the port address and the port number. This concept facilitates the ability to perform FICON port swaps for maintenance operations without the need to regenerate the host configuration. The port address abstraction is not in the Fibre Channel architecture and is foreign to FCP. FCP communications are name-centric, discovery-oriented, and fabric-assigned, and they use the Fibre Channel Name Server to determine device communication. When an FCP device attaches to the fabric, it queries the Name Server for the zoning information. FICON devices do not query because the allowable port and device relationships have been previously defined. FCP configurations support multiple fabrics and allow seven hops between source and target. FCP port-to-port connectivity has traditionally been enforced via zoning, although other techniques complementary to zoning (such as port, Director, and fabric binding) are used. Binding helps alleviate the security concerns that are experienced in intermix installations, because with FCP, any device in the fabric can access the ANSI standard FC management server by logging into the fabric. If you are implementing intermix, you must block the transfer of any and all frames from a FICON Director port to all SAN connected ports and vice versa.

72

FICON Planning and Implementation Guide

For more detailed information about these topics, refer to: Implementing an IBM/Brocade SAN with 8 Gbps Directors and Switches, SG24-6116 IBM/Cisco Multiprotocol Routing: An Introduction and Implementation, SG24-7543 IBM System Storage/Brocade Multiprotocol Routing: An Introduction and Implementation, SG24-7544 Implementing an IBM/Cisco SAN, SG24-7545

4.6.2 Fabric security

Best security practices typically address three main objectives: availability and destruction; confidentiality and disclosure; and integrity and alteration of data. However, it is important to remember that the majority of security threats are from insiders. Therefore, any SAN security strategy should also include the following: Restrict administrator privileges Isolate sensitive environments Audit SAN activities All-level documentation You need to consider the following topics when planning a secure fabric. Securing access to the management tools, by using: HTTPS Secure Shell (SSH) Secure Socket Layer (SSL) Secure Copy (SCP) LDAP-MS Active Directory Role-Based Access Control (RBAC) Password policies IP Filters (firewall) Passive FTP RADIUS server Zoning Fabric Configuration Server (FCS) Fabric Configuration Server (FCS) policy Switch Connection Control (SCC) policy Device Connection Control (DCC) policy CHAP and DH-CHAP IPsec P-EAP and MS-CHAP

Securing the fabric, by using:

Securing Long-Distance SAN, by using: Physical separation Electronic separation Logical separation The use of Intrusion Detection and Incident Response Automatic event detection and management Event forensics Incident procedure to exercise due diligence

Chapter 4. Planning the FICON environment

73

The use of Fabric-Based Encryption This technology enables you to encrypt specific or all disk data to prevent accidental data leaks when disk drives are replaced or disk arrays refreshed, and to protect enterprises from data theft. Although this is not supported with FICON, you can use the encryption capabilities of System z.

4.6.3 High integrity

Data integrity ensures that any changes to the data streams are always detected, and that the data frames are delivered to the right endpoint. A FICON channel requires the support of high-integrity architecture when configured to operate in a cascaded Director fabric. During initialization, the FICON channel queries the Director to determine that it supports high-integrity architecture. If it does, the channel completes the initialization process, allowing the channel to operate with the fabric. High-integrity fabric architecture support includes fabric binding and insistent Domain ID support. A FICON Director provides the following fabric, Director, and port binding features: Fabric binding is a security method for restricting Directors within a multiple-Director fabric. This policy prevents unauthorized Directors from joining a fabric. Director binding is a security method for restricting devices that connect to a particular Director. A strict fabric-wide consistency is necessary for FICON Director binding. Port binding is a security method for restricting server or storage devices that connect to particular Director ports. This policy binds device ports to Director ports. We recommend that you set the switch ID (in IOCP or HCD) to the same value as the domain ID of the FICON Director, which is defined to the FICON Director at installation time. This simplifies the configuration and reduces confusion by having a common designation across all definitions. Domain ID and switch id have to be unique. When defining the switch IDs in the IOCP definitions, ensure that you use values within the FICON Directors range. If you use 2-byte addressing on the System z, then strict fabric binding is mandatory (even with one FICON Director in the fabric).

4.6.4 Zoning
Zoning is a method used in the Director to restrict communication. There are two types of zoning: WWN zoning and port zoning. Zoning is used, for example, to separate FICON devices from each other, to restrict confidential data to specific servers, and to control traffic paths. The default zone automatically puts nodes attached to a fabric that are not already in a zone into a single zone and allow any-to-any connectivity. Conversely, default zone members are automatically removed whenever that member is added to an active zone. The best practice recommendation is to disable the default zone and create port zones, even if the port zone includes all the ports in the fabric. By disabling the default zone and defining port zones, you ensure that the maximum number of zone members is not exceeded and does not inadvertently allow unintended connectivity. 74
FICON Planning and Implementation Guide

WWN zoning permits connectivity between attached nodes based on WWN. That node can be moved anywhere in the fabric and it will remain in the same zone. WWN zoning is used in open systems environments and is not used for FICON. The IOCP effectively does what WWN zoning does in open systems environments. Adding WWN zoning on top of that adds an unnecessary layer of complexity, because every time a new channel card or control unit interface is added to the fabric, the WWN zoning would have to be modified. Port zoning limits port connectivity based on port number. Port zoning is what should be used for FICON. Ports can easily be added to port zones even if there is nothing attached to the port. In purely FICON environments, you put all ports in a single port zone. Define port zones for FICON and World Wide Name (WWN) zones for FCP. Disk and tape mirroring protocols are FCP even when the front end is FICON. We recommend that you separate this traffic from the FICON traffic either with WWN zoning or virtual fabrics. In more complex fabrics, more zones can provide additional security. In intermixed environments, only the FICON ports are put in the port zone. FICON prohibits dynamic connectivity mask (PDCM) controls whether or not communication between a pair of ports in the Director is prohibited or allowed. If there are any differences in the restrictions set up and PDCM, the most restrictive rules are automatically applied. Link incidents for FICON ports are reported only to registered FICON listener ports. The only exception to this is the loss of synchronization link incident. To create a solid base for a SAN, use standard open systems best practices for creating WWN zones for the FCP traffic to ensure that the SAN is secure, stable, and easy to manage: Implement zoning, even if LUN masking is used. Persistently disable all unused ports to increase security. Use port WWN identification for all zoning configuration. Limit the aliases and names to allow maximum scaling. Use frame-based hardware enforcement. Use single initiator zoning with separate zones for tape and disk traffic if an HBA is used for both traffic types. Disable access between devices inside the default zone.

4.7 Management
We recommend that you manage the FICON fabric at various levels. Each level has a specific mission to fulfill. The most simple is offered by the CLI. Unfortunately, this does not provide enough control over the entire environment. To operate the FICON infrastructure you need to automate the whole process. All management levels create together robust systems management and interconnections with data center management.

4.7.1 Command-line interface

The command-line interface (CLI) uses role-based access control (RBAC) to manage access for operations of the FICON Director. Changing the configuration of a FICON Director without a Telnet connection requires a serial connection. A serial connection may not be possible if the FICON Director is physically located in a restricted or remote area.

Chapter 4. Planning the FICON environment

75

Because all FICON Directors ship with the same default IP address, only one can be connected to the management network at a time. You will also need a sufficient number of IP addresses for each FICON Director dependent on the platform. If all FICON Directors will be connected to the same management network, then they must have the same subnet mask and have a common IP addressing range. For recommended tasks executed from the CLI, refer to platform-specific documentation.

4.7.2 Element management

Web Tools enables an intuitive and easy-to-use interface which helps to monitor and manage FICON Directors. You can perform tasks by using a Java-capable Web browser from your laptop, desktop PC, or workstation from any location with IP connectivity (Analogy to Cisco MDS Device Manager). Web Tools provides an extensive set of features that enable you to quickly and easily perform key administrative tasks: Setting up IP addresses, Director names, and Simple Network Management Protocol Performing firmware upgrades and controlling boot parameters Maintaining RBAC, license keys, and RADIUS Enabling Ports on Demand Routing setup Configuring links and ISL trunks

4.7.3 Fabric management

A client-server based application for managing the whole multi-fabric environment is called the Data Center Fabric Manager (DCFM). (Analogy to Cisco MDS Fabric Manager Server or Cisco Data Center Network Manager.) DCFM Enterprise includes following: High scalability of managed Director ports Multiprotocol support Status and connectivity visualization End-to-end performance monitoring Fault management Troubleshooting and diagnostics Call home Offline support analysis Configuration management Security management FICON environment configuration and management with analysis and diagnostic capabilities Virtual fabric management

4.7.4 Storage management initiative specification

The storage management initiative specification (SMI-S) defines the interface that allows storage management systems to manage and monitor storage area network (SAN) resources. SMI-S is the only standard that addresses manageability from the perspective of a device in a storage array, all the way through the data path, to an application running on a host.

76

FICON Planning and Implementation Guide

The SMI-S does not require any modification or upgrade to deployed fabrics when it is deployed. The implementation is vendor-specific. It can be a demon running directly on the Director platform or a separate host application. Each fabric device needs its own specific agent. The advantage of using an end-to-end storage management platform is the possibility of management automation of the whole storage network including the FICON fabric. For more details about IBM Storage Management Platform, refer to the following URL: http://www.storage.ibm.com

4.7.5 System z management for FICON Directors

Combined, IBM Tivoli System Automation for z/OS and the Control unit port (CUP) of the FICON Director have the capability to control and automate the entire I/O communication. Resource Monitoring Facility (RMF) and the CUP can provide activity statistics and resource utilization.

Control unit port

The control unit port (CUP) used by System z provides in-band management of the FICON Director. FICON management server mode must be enabled on the Director to enable CUP management features. This prevents local commands from interfering with host-based commands by starting serial access to Director parameters. You should use the CUP for the following reasons: It enables you to perform in-band management from the Hardware Management Console (HMC). It enables you to use the management tools used with ESCON, such as System Automation z/OS for I/O-Ops. It enables you to get Service Information Messages (SIMs) to the z/OS console for FICON device hardware failures. It enables you to use the Dynamic Channel Path Management function of Workload Manager and IRD (DCM) with FICON. It enables you to receive RMF Monitor I reports for reporting on FICON Director activity statistics in the SMF Record Type 74 Subtype 7. RMF Monitor I performs long-term data collection for resource utilization. A FICON Director that supports CUP can be controlled by one or more host-based management tools. A mode register controls the behavior of the FICON Director according to CUP and other management interfaces. We make the following recommendations: Only one RMF should attempt to access the CUP at any given time. Too much activity to the CUP can cause missing interrupts. Best practice is to have two or more CHPIDs with access to the CUP for backup. You also have to configure the Missing Interrupt Handler process timeout, which determines when the channel will time out waiting for an expected response from the CUP. If processing times are increased, it is possible to inadvertently time out the channel because the CUP is busier than normal. FICON Directors have an embedded port FE in the Control Processor for the CUP function. On a 256-port Director, this logical FE address overlaps the physical port FE address. When
Chapter 4. Planning the FICON environment

77

using the CUP, on 256 port boxes, physical port addresses FE and FF cannot be used. In this case, use physical port addresses FE and FF for port swapping and for intermix ports. The following applies to FICON Directors with at least 256 ports in use: The FICON Director is seen by the System z server as a 256-port Director. The Prohibit dynamic connectivity mask and port names are for ports 0 through 253. Ports 254 and 255 are reserved. For information about the CUP license, behavior, and addressing rules, refer to FICON Director-specific documentation.

IBM Tivoli System Automation for z/OS

With this management platform you can monitor, control, and automate a wide range of system elements spanning both the hardware and software resources of your infrastructure. Tivoli System Automation for z/OS performs the following major operations: System operations It monitors and controls system operations, applications, and subsystems. Processor operations It monitors and controls processor hardware operations, and provides a connection from a focal point processor to a target processor. I/O operations It provides a single point of control for managing connectivity in your active I/O configurations. It takes an active role in detecting unusual I/O conditions and allows you to view and change paths between a processor and an input/output device, which can involve using dynamic switching, such as the FICON Director. I/O operations changes paths by allowing you to control channels, ports, Directors, control units, and input/output devices. You can do this through an operator console or API. For more details, you can refer to: IBM Tivoli System Automation for z/OS Enterprise Automation, SG24-7308

4.8 Virtualization and availability

To achieve the most effective FICON infrastructure, consider virtualizing your FICON components. Virtualizing the FICON components enables a flexible and dynamic resource sharing solution. In addition, we always recommend that you have at least two physical connections to multiple FICON components to avoid single points of failures and ensure high availability. In the following sections we discuss the virtualization capabilities of the FICON components and what you need to take into account when considering high availability for your FICON infrastructure.

4.8.1 System z
The use of FICON channels gives you the ability to define shared channels that can request logical paths. However, control units can only allocate a limited number of logical paths in relation to the number of logical paths that FICON channels can request. In configurations

78

FICON Planning and Implementation Guide

where channels request more logical paths than a control unit can allocate, you must manage logical paths to help ensure that the I/O operations take place. With proper planning, you can create I/O configuration definitions that allow control units in the configuration to allocate logical paths for every possible request made by channels in either of the following ways: Create a one-to-one correspondence between the logical path capacity of all control units in the physical configuration and the channels attempting to request them. Create I/O configurations that can exceed the logical path capacity of all or some of the control units in the physical configuration, but at the same time provide the capability to selectively establish logical connectivity between control units and channels as needed. This capability can be useful or even necessary in several configuration scenarios. Several components of System z provide the capability for this virtual environment, as explained in the following sections.

Channel Subsystem
Each server has its own Channel Subsystem (CSS). The CSS enables communication from server memory to peripherals via channel connections. The channels in the CSS permit transfer of data between main storage and I/O devices or other servers under the control of a channel program. The CSS allows channel I/O operations to continue independently of other operations within the server. This allows other functions to resume after an I/O operation has been initiated. The CSS also provides communication between logical partitions within a physical server using internal channels.

Multiple Channel Subsystem

The design of System z servers offers a considerable increase in processing power, memory sizes, and I/O connectivity. In support of the larger I/O capability, the CSS has been scaled up correspondingly, and the Logical Channel Subsystem (LCSS) concept is designed to do that. Concepts were introduced to facilitate this architectural change and provide relief for the number of supported LPARs, channels, and devices available to the server.

Multiple Subchannel Sets

Multiple Subchannel Sets (MSS) provides relief for I/O device configurations in large System z environments. It also increases Parallel Access Volume (PAV) connectivity. Subchannel numbers (including their implied path information to a device) are limited to four hexadecimal digits by hardware and software architectures. PAV has made this limitation of subchannels a challenge for larger installations. A single disk drive (with PAV) often consumes at least four subchannels. The solution gives you the ability to have sets of subchannels (addresses), with a current implementation of two sets. For more information about this topic, refer to: Multiple Subchannel Sets: An Implementation View, REDP-4387

Multiple Image Facility

Multiple Image Facility (MIF) allows channels to be shared among multiple LPARs. Shared channels are configured to an LPAR, thus giving the LPAR a channel image of the shared channel that it can use. Each channel image allows an LPAR to independently access and control the shared channel as though it were a physical channel assigned to the LPAR. By providing the logical equivalent of multiple physical channels dedicated to multiple LPARs, a shared channel can reduce hardware requirements without a corresponding reduction in

Chapter 4. Planning the FICON environment

79

I/O connectivity. This reduction in hardware requirements can apply to physical channels, Director ports, and control unit ports, depending on the configuration. MIF further improves control unit connection topologies for System z servers with multiple LPARs. MIF enables many LPARs to share a physical channel path, thereby reducing the number of channels and control unit interfaces required without a corresponding reduction in I/O connectivity. Installation can take advantage of MIF performance enhancements offered by: Understanding and utilizing I/O-busy management enhancements Planning for concurrent data transfer Understanding examples of MIF consolidation Understanding and utilizing I/O-busy management enhancements Before you can consolidate channels, you must be aware of the channel requirements of the particular control units you are configuring. The number of channels needed is independent of the number of LPARs on a system. The number of channels is based on the number of concurrent data transfers that the control unit is capable of. Although the recommended number of channels satisfies connectivity and performance requirements, additional channels can be added for availability. Note that not all ESCON or FICON configurations benefit from the use of shared channels. There are some configurations where using an unshared channel is more appropriate, as explained here: When there are logical path limitations of the control unit Although many ESCON control units can communicate with multiple LPARs at a time using multiple logical paths, there are some ESCON-capable control units that can only communicate with one LPAR at a time. When the channel utilization of shared channels will be greater than unshared channels If you use shared channels to consolidate channel resources, you must consider the channel utilization of all the channels you consolidate. The channel utilization of a shared channel will roughly equal the sum of the channel utilizations of each unshared channel that it consolidates. If this total channel utilization is capable of decreasing performance, then you must consider using unshared channels or a different configuration of shared and unshared channels to meet your connectivity needs. MIF allows you to use shared channels when defining shared devices. Using shared channels reduces the number of channels required, allows for increased channel utilization, and reduces the complexity of your IOCP input. You cannot mix shared and unshared channel paths to the same control unit or device. Channel spanning extends the MIF concept of sharing channels across logical partitions to sharing channels across logical partitions and channel subsystems. When defined that way, the channels can be transparently shared by any or all of the configured logical partitions, regardless of the Logical Channel Subsystem to which the logical partition is configured. For more details see: System z Enterprise Class System Overview, SA22-1084 z/Architecture, Principles of Operation, SA22-7832 System z Input/Output Configuration Program Users Guide for ICP IOCP, SB10-7037 System z Processor resource/Systems Manager Planning Guide, SB10-7153 System z Support Element Operations Guide, SC28-6879 IBM System z Connectivity Handbook, SG24-5444 80
FICON Planning and Implementation Guide

4.8.2 Control unit

Consider the following factors concerning the allocation of logical paths by control units: Control units allocate logical paths dynamically on a first come-first served basis. Control units do not manage the allocation of logical paths but instead allow channels to compete for logical paths until all of the control units logical paths are used. Control units vary in the number of logical paths they support. The FICON infrastructure significantly affects the volume of logical path requests to a control unit as follows: Control units can attach to one or more ports on a Director or to additional ports on other Directors. Each Director port can dynamically connect to many other ports to which channels requesting logical paths are attached. For System z servers, each logical partition attaching to the same control unit will compete for the control units logical paths. In a configuration where control units are shared by different servers, I/O configuration definitions for individual control units are not coordinated automatically among the IOCDSs of the different servers. Each System z server competes for a control units logical paths. Shared channels require the establishment of a logical path for each channel image corresponding to an active LPAR sharing the channel. This can increase the number of logical paths that a single channel requests. FICON support offers a high throughput performance for a single channel, which reduces the number of required channels. The CU channel adapters maintain compatibility with the existing infrastructure through auto-negotiation. A high performance FICON feature provides a significant reduction in channel utilization. This reduction can allow more I/O input on a single channel, and can also allow for a reduction in the number of FICON channels required to support a given workload. The number of offered CU channel adapters is further increasing the ability to share the storage. On a DS8000 storage unit, each storage unit Fibre Channel adapter has four channels. Each channel has a unique World Wide Port Name (WWPN). On FICON, the Fibre Channel CU channel adapter can operate with fabric or point-to-point topologies. The CU FICON channels limit the number of devices per channel to 16384. To fully access 65280 devices on a storage unit, you need to connect a minimum of four FICON channels to the storage unit. You can access the devices through a FICON Director to a single storage unit FICON channel. With this method, you can expose 64 control-unit images (16384 devices) to each System z channel. Parallel Access Volume (PAV) enables a single System z server to simultaneously process multiple I/O operations to the same logical volume, which can help to significantly reduce device queue delays (IOSQ time). This is achieved by defining multiple addresses per volume. With dynamic PAV, the assignment of addresses to volumes can be automatically managed to help the workload meet its performance objectives and reduce overall queuing. With PAV, reads are simultaneous. Writes to different domains (a set of tracks the disk controller is working on) are simultaneous as well. However, writes to the same domain are serialized. No double updates are possible to preserve integrity. You can further enhance PAV by adding the IBM HyperPAV feature. IBM HyperPAV associates the volumes with either an alias address or a specified base logical volume number. IBM HyperPAV is only supported on FICON channel paths.

Chapter 4. Planning the FICON environment

81

For more information about these topics, refer to: IBM System Storage DS8000 Architecture and Implementation, SG24-6786 IBM System Storage DS8000 Introduction and Planning Guide, GC35-0515

4.8.3 FICON Director

The Virtual Fabrics feature provides data, control, and management isolation for logical Directors and logical fabrics according the ANSI standard for Virtual Fabrics. Virtualization allows you to: Isolate traffic types, applications, and management Pool resources Increase FICON consolidation and scalability You have to consider how many virtual fabrics for FICON with FICON Management Server you can create based on vendor limitations. All standard Director features are supported per virtual fabric. For a complete list of features refer to Chapter 3, FICON Director technical description on page 33 and to platform-specific documentation. The Virtual Fabrics option can be added to the existing fabric capabilities, but it does not replace them. FICON cascading only supports an interconnection between FICON Directors of the same platform. The following topics influence the logical layout of virtual fabrics.

Port characteristics
The port addressing mode is platform-specific. For more details refer to Chapter 3, FICON Director technical description on page 33 and to platform-specific documentation. For port addressing, use the worksheet provided in Appendix B, Configuration worksheets on page 251, to help document the physical layout of your FICON Director. Ports will automatically determine what they are connected to and at what speed they will operate. Manual settings should only be used with equipment that cannot auto-negotiate or login properly. Typically, the only interfaces that require manual settings are older 1 Gbps interfaces and long distance ISLs. Before installing the FICON Director, consider where you will connect your FICON channels, based on your availability requirements. Distribute the channels among different port cards. If two channels are defined to access the same CU, plug both fiber optic cables into different port cards in the Director. Distribute the CU ports among different port cards. If two paths are defined to attach the CU to the server through the Director, connect both fiber optic cables to ports on different cards. Distribute the ISLs across different port cards. If two or more ISLs are to be attached between the Directors, connect the fiber optic cables to ports on different cards. For frame-based trunking, the ISLs needs to be on the same ASIC. Following these rules will ensure that there is always one path available between the server and the CU in case of a defective port card in the Director.

82

FICON Planning and Implementation Guide

If multiple FICON Directors are planned based on the number of required connections and paths, consider spreading those paths across different Directors.

Inter-switch link
The number of Inter-switch links (ISLs) must be carefully planned based on the existing and planned topology, workload characteristics, traffic patterns, and system performance. For availability, there should always be two ISLs between any two cascaded FICON Directors. A general guideline is to plan for one ISL to four FICON channels that are routed over it. The ISL speed must reflect the speed of the channel. Disk and tape should not share the same ISLs. FICON and FCP traffic should not share the same ISLs. Use TI zones to direct traffic to different ISLs. This is vendor-specific. Plan on having enough ISL bandwidth so that the ISLs do not become congested when the maximum number of streaming tape devices are running, because that could result in tape traffic termination. Tape performance is adversely affected when tapes stop. Tapes are also more likely to break when stopping and starting.

Virtual ISL
When two Directors are connected, an ISL connection for two logical switches or extended ISL (XISL) connection for two base switches is created, dependent on the configuration. When logical switches with the same Fabric ID are configured to use the XISL, the Director automatically creates a logical ISL (LISL) within the XISL. The LISL isolates traffic from multiple fabrics. XISL and LISL technology is not supported for FICON at this time. For Cisco, the equivalent of virtual ISL is Trunking E-ports, which can transport virtual SAN Fabric information between Directors. For information regarding the support of virtual ISLs in FICON environments, refer to platform-specific documentation.

4.9 Performance
A single FICON channel can replace multiple ESCON channels. This means fewer channels, Director ports, and control unit ports, resulting in a simpler configuration to manage. Additionally, FICON supports the combined CTC and channel function. ESCON-to-FICON channel aggregation can be anywhere between 8:1 and 2:1. An understanding of the factors that affect performance must be established before deciding on your target configuration. For example, the characteristics of your workload have a direct impact on the performance you will realize in your FICON environment. In addition, keep the utilization of your FICON channels at or below 50% utilization to maintain optimum response times. The following items will require analysis: I/O rates Block sizes Data chaining Read/write ratio

Chapter 4. Planning the FICON environment

83

For more information about these topics, read the following performance papers: Performance Considerations for a Cascaded FICON Director Environment Version 0.2x, by Richard Basener and Catherine Cronin FICON Express2 Channel Performance Version 1.0, GM13-0702 IBM System z9 I/O and FICON Express4 Channel Performance, ZSW03005USEN IBM System z10 I/O and High Performance FICON for System z Channel Performance, ZSW03059USEN

I/O operations
Deploying FICON channels demands a further understanding of the I/O operations in your environment. The following terms are used to describe the different phases with respect to the measurements available for determining the duration of an I/O operation. I/O supervisor queue time (IOSQ), measured by the operating system. The I/O request may be queued in the operating system if the I/O device, represented by the UCB, is already being used by another I/O request from the same operating system image (UCB busy). The I/O Supervisor (IOS) does not issue a start subchannel (SSCH) command to the Channel Subsystem until the current I/O operation to this device ends, thereby freeing the UCB for use by another I/O operation. Pending time (PEND), measured by the channel subsystem. After IOS issues the start subchannel command, the channel subsystem may not be able to initiate the I/O operation if any path or device busy condition is encountered. Channel busy Director port busy Control unit adapter busy Device busy

Connect time, measured by the channel subsystem. This is the time that the channel is connected to the control unit, transferring data for the I/O operation. Disconnect time The channel is not being used for the I/O operation, because the control unit is disconnected from the channel, waiting for access to the data or to reconnect. In the following sections we describe factors that will influence your FICON performance design.

4.9.1 Frame pacing

Frame pacing is an FC-4 application data exchange measurement and throttling mechanism. It uses buffer credits to provide a flow control mechanism for FICON to assure delivery of data across the FICON fabric. After all buffer credits for a port are exhausted, a frame pacing delay occurs. These delays generally result in longer FICON connect time or longer PEND times that show up on the volumes attached to these links. Frame latency is the average amount of time it takes to deliver a frame from the source port to the destination port. Each element that handles the frame contributes to this latency measurement (channel, Director, storage port adapter, link distance, and so on). Only when using FICON Directors and when the CUP is enabled can RMF provide frame pacing delay information.

84

FICON Planning and Implementation Guide

You must also keep in mind that tape workloads have larger payloads in a FICON frame, while disk workloads might have much smaller. The average payload size for disk is often about 800 to 1500 bytes. By using the FICON Director activity reports, you can gain an understanding of your average read and write frames sizes on a port basis. The buffer credit represents the number of receive buffers supported by a port for receiving frames. The minimum value of buffer credits is one (1). This value is used as a controlling parameter in the flow of frames over the link to avoid possible overrun at the receiver. The number of buffer credits is not associated with the performance until high data rates are attempted over long distances. If there are insufficient buffer credits, there may be a hard limit on the data rate that can be sustained. The number of buffer credits available for each port on the FICON Director is implementation-dependent. The optimal amount of buffer credits is determined by the distance (frame delivery time), the processing time at the receiving port, link data rate, and the size of the frames being transmitted. There are four implications to consider when planning buffer credit allocation: Ports do not negotiate buffer credits down to the lowest common value. A receiver simply advertises buffer credits to the linked transmitter. The exhaustion of buffer credits at any point between an initiator and a target will limit overall performance. For write-intensive applications across an ISL (tape and disk replication), the buffer credit value advertised by the E_Port on the target cascaded FICON Director is the major factor that limits performance. For read-intensive applications across an ISL (regular transactions), the buffer credit value advertised by the E_Port on the local FICON Director is the major factor that limits performance. Buffer credits are mainly a concern for extended distances; however, a poorly designed configuration may consume all available buffer credits and have an impact on performance. For example, assume you have a FICON 8 Gbps channel attached to two different control units running at lower link rates 4 Gbps and 2 Gbps. Depending on the traffic pattern, it can happen that the low speed device consumes all the available buffer credits of the 8 Gbps link and you are not able to send more packets between the server and FICON Director. In this case the average link utilization goes down. For this reason, we recommend that you upgrade all low speed channels where possible, or physically separate low and high speed environments. For more information about this topic, refer to Buffer credits on page 39.

4.9.2 Extended distance FICON

The control units that support the Extended distance FICON feature are able to increase the IU pacing count, which eliminates handshakes between channel and control unit. This support allows the channel to remember the last pacing information and use this about subsequent operations to avoid performance degradation at the start of a new I/O operation. Improved IU pacing (255 IU instead of 16) improves the utilization of the link. Extended distance FICON can eliminate the need for channel extenders in 2-site and 3-site z/OS Global Mirror configurations, because of the increased number of read commands simultaneously in flight. Extended distance FICON enables greater throughput over distance for IBM z/OS global mirror (XRC) using the same distance.
Chapter 4. Planning the FICON environment

85

Note, however, that Extended distance FICON does not extend the achievable physical FICON distances or offer any performance enhancements in a non-z/OS global mirror environment. For more information about this topic, refer to 2.1.9, Extended distance FICON on page 24.

4.9.3 Multiple allegiance

Multiple allegiance (MA) is the concurrent operations capability of the IBM disk storage systems that facilitates multiple accesses to or from the same volume with multiple channel path groups and system images. It allows multiple servers to establish concurrent implicit allegiances if there is no possibility that any of the channel programs can alter any data that another may read or write. This can reduce the PEND time, because device busy is not presented to the channel. Resources that will benefit from multiple allegiance are: Volumes that may have concurrent read operations or a high read-to-write ratio Data sets that have a high read-to-write ratio, multiple extents on one volume, or are concurrently shared by multiple users

4.9.4 Parallel Access Volume and HyperPAV

Parallel Access Volume (PAV) enables a single System z server to simultaneously process multiple I/O operations to the same logical volume, which can help to significantly reduce device queue delays (IOSQ time). This is achieved by defining multiple addresses per volume. With dynamic PAV, the assignment of addresses to volumes can be automatically managed to help the workload meet its performance objectives and reduce overall queuing. With PAV, reads are simultaneous. Writes to different domains (a set of tracks the disk controller is working on) are simultaneous as well. However, writes to the same domain are serialized. No double updates are possible to preserve integrity. You can further enhance PAV by adding the IBM HyperPAV feature. IBM HyperPAV associates the volumes with either an alias address or a specified base logical volume number. IBM HyperPAV is only supported on FICON channel paths. HyperPAV is designed to: Provide a more efficient PAV function. For large volumes, to scale I/O rates without the need for additional PAV alias definitions. Exploit FICON architecture to reduce overhead, improve addressing efficiencies, and provide storage capacity and performance improvements. Dynamic assignment of PAV aliases improves efficiency. The number of PAV aliases needed might be reduced, taking fewer from the 64 K device limitation and leaving more storage for capacity use. Enable a more dynamic response to changing workloads. Simplified management of aliases. Avoid migration to larger volume sizes. HyperPAV provides a far more agile alias management algorithm because aliases are dynamically bound to a base for the duration of the I/O for the z/OS image that issued the I/O.

86

FICON Planning and Implementation Guide

When an I/O completes, the alias is returned to the pool for the logical control unit. It then becomes available to subsequent I/Os. The number of aliases required can be approximated by the peak I/O rates multiplied by the average response time. For example, if the average response time is 4 ms and the peak I/O rate is 2000 per second, then the average number of I/O operations executing at one time for that LCU during the peak is eight. Therefore, eight PAV-aliases should be all that is needed to handle the peak I/O rate for the LCU, along with all the other PAV-base addresses in the LCU. Depending on the kind of workload, there is a huge reduction in PAV-alias UCBs with HyperPAV. The combination of HyperPAV and EAV allows you to significantly reduce the constraint on the 64 K device address limit and in turn increase the amount of addressable storage available on z/OS. In conjunction with Multiple Subchannel Sets (MSS) on System z, you have even more flexibility in device configuration. For more information about this topic, refer to: Disk storage access with DB2 for z/OS, REDP-4187 http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/TD100311 Use the PAV analysis tool to achieve a better understanding of your workload profile and achieve a more precise split: http://www.ibm.com/servers/eserver/zseries/zos/unix/bpxa1ty2.html

4.9.5 Modified Indirect Data Address Word

Using Modified Indirect Data Address Words (MIDAWs) will not cause the bits to move any faster across the FICON link, but MIDAWs do reduce the number of frames and sequences flowing across the link, which makes the channel more efficient. Channel efficiency is essentially the relationship of throughput to channel utilization. For a homogenous I/O stream of some uniform set of I/Os, the I/O rate, throughput, and channel utilization are all linearly related to each other. A channel becomes more efficient if the throughput at a particular level of utilization is increased. The most significant performance benefit of MIDAWs is achieved with extended format data sets. The MIDAW facility is a conceptually simple modification to IDAWs. MIDAWs remove the 4 KB boundary restrictions of IDAWs. Using MIDAWs, the z/OS media manager can transfer a whole track using a single CCW. To span multiple tracks in one channel program, command chaining was used, and still is used. However, with the advent of MIDAWs, the media manager no longer uses data chaining. If track-level CCWs were the foundation for improved FICON performance, then MIDAWs represent the building that rests upon the foundation. MIDAWs and IDAWs have basically the same performance characteristics, and both perform better than data chaining, for two reasons: One reason pertains to the number of CCWs used, which in turn affects the number of frames sent across the link. By reducing the number of frames, it takes less time for the FICON channel and the control unit host adapter port to process the channel program. The second reason pertains to FICON multiplexing and the fact that MIDAWs remove some of the data-chaining activity on the link. Multiplexing only affects performance when there are concurrent I/Os on the channel.

Chapter 4. Planning the FICON environment

87

In addition to reducing channel utilization, indirect addressing also has the effect of reducing the utilization of the control unit FICON channel. This effect is hidden, because the system does not have any way of measuring the utilization of the channel. The channel has to process every CCW, with or without data chaining, but indirect data addressing is transparent to the channel. MIDAWs help reduce channel utilization and affect the relative performance of different block sizes. Physical disk performance is irrelevant if the data permanently resides in cache, or if the channels are faster than the speed at which data can be sequentially read from the disks. However, block size can also affect FICON channel performance. In other words, in addition to eliminating the extended format performance penalty, using MIDAWs also eliminates the small block performance channel penalty, except to the extent that the record size affects the number of bytes that fit on a track. Small blocks are usually better for database transaction workloads, because the buffer hit ratio for random access tends to be higher with smaller blocks. MIDAWs affect the performance of update writes in the same manner as they affect reads. For more information about this topic, refer to: How Does the MIDAW Facility Improve the Performance of FICON Channels Using DB2 and Other Workloads?, REDP-4201

4.9.6 High Performance FICON

Substantial performance improvements can be realized in terms of throughput (MBps) and I/Os per second (number of starts per channel), depending on the application workload. The initial implementation of High Performance FICON (zHPF) by the IBM DS8000 is exclusively for I/Os that transfer less than a single track of data. Realistic production workloads with a mix of data transfer sizes can see between 30% and 70% of FICON I/Os utilizing zHPF. This can result in between a 10% to 30% saving in channel utilization. The zHPF architecture has built-in reliability, availability, and serviceability enhancements. Significantly more information regarding errors is now provided by the control unit, thereby aiding fault isolation. There is also dedicated buffer for concurrent sense data now, thus making it unnecessary to separate sense commands. Missing interrupt problem determination is provided through an interrogate mechanism, thereby providing the operating system with the capability to query the state of operation at the control unit when detecting missing interrupts. Previously no information was available on missing interrupts and debugging involved attempting to correlate software, channel, and control unit log and trace data. Furthermore, the reason for aborts is provided to control units using an enhanced purge path mechanism. There is no change required to the IOCP when using the zHPF architecture. FICON channels are still defined the same way, as CHPID type FC. There is also no change required to the configuration or design of the fabric. I/O will flow over the fabric as either FICON or zHPF IUs. zHPF is transparent to FICON Directors. The potential exists to achieve some channel aggregation with the adoption of zHPF, but remember to give careful consideration to meeting all your workload requirements. The benefits of zHPF particularly apply to OLTP I/O workloads such as DB2, VSAM, PDSE, and zFS which typically transfer small blocks of fixed (4 K) data. zHPF-capable channels and devices support both FICON and zHPF traffic simultaneously. FICON continues to use CCWs and zHPF uses TCWs, so you have compatibility and coexistence.

88

FICON Planning and Implementation Guide

For more information about this topic, refer to: High Performance FICON for System z: Technical Summary for Customer Planning, ZSW03058USEN

4.9.7 Bandwidth management

This section discusses the features and function that can influence the fabric bandwidth.

Local switching
There are two Director architectures: Multi-stage switching with interconnected ASICs The architecture approach uses ASICs to both process and switch data traffic. Processing can measure performance, enforcing zoning and other protocol-related tasks. Switching is simply directing traffic from one port to another. If more user-facing ports are required, two or more ASICs are internally interconnected. Single-stage switching with one or more crossbars ASICs process data on the port blades, but they do not switch the traffic. Switching is performed by one or more serial crossbars. This provides highly predictable performance across all ports and an easier design for higher port capacity. The bandwidth maximum for a Director chassis is the crossbars bandwidth. There are three connection allocation possibilities: You can gain a local switching benefit by distributing storage, server, and ISL ports evenly across Director port blades, to minimize the impact of losing a port blade to a hardware fault or other issue. You can perform a slight modification by distributing storage, server, and ISL ports evenly across port blade ASICs. By going across port blades and alternating between the high and low ASICs, local switching occurs much more frequently. You can group storage, server, and ISL ports that will need to switch with each other within Local Switching groups. Ensure that array devices are mapped to a local switching group that also has server ports that will require those devices. Combinations of local and non-local traffic flows can occur simultaneously. The oversubscription rates can be optimized, depending on the design.

Fabric Shortest Path First

Port speed-weighted Fabric Shortest Path First (FSPF) Fibre Channel routing protocol applies a weighting factor based on link speed to ensure that data traffic uses higher-speed paths between Directors. It gives a basic utilization of multiple ISLs between Directors with the same cost. However, it does not assure that all links are equally utilized, which can cause a congestion on the link even others are not utilized at all. Because of this, we recommend that you use more effective ISL connectivity called frame-based trunking.

Frame-based trunking
Using frame-based trunking, data flows are automatically distributed over multiple physical Inter-switch link connections. Frame-based trunking combines up to eight ISLs into a single logical trunk, optimizes link usage by evenly distributing traffic across all ISLs at the frame

Chapter 4. Planning the FICON environment

89

level, maintains in-order delivery to ensure data reliability, ensures reliability and availability even if a link in the trunk fails, and simplifies management by reducing the number of ISLs. Frame-based trunking is the most effective way to utilize ISLs for FICON traffic while maintaining in-order-delivery (IOD) of frames inside the fabric.

Dynamic Path Selection

With Dynamic Path Selection (DPS), the routing path is based on the Source ID (SID), Destination ID (DID), and Fibre Channel originator exchange ID (OXID). Every exchange can take a different path through the fabric. Exchange-based routing requires the use of the Dynamic Load Sharing (DLS) feature. When there are multiple paths to a destination, the input traffic will be distributed across the different paths according to the bandwidth available on each of the paths. Every time there is a change in the network, the input is redistributed across the available paths. DPS technology is currently not supported for FICON, but can be utilized for Intermix Fabrics for FCP in a virtually-separated fabric.

Port-based DLS
DLS monitors link or trunk utilization to ensure load balancing. The choice of port-routing path is based only on the incoming port and destination domain. This feature is not supported for FICON Fabric.

Lossless DLS
Lossless Dynamic Load Sharing (DLS) allows you to rebalance trunk port paths without causing input/output (I/O) failures in cases where the end devices require in-order-delivery (IOD) of frames.

Port-channel
The Port-channel feature is an equivalent for link trunking on Cisco FICON Directors. The balancing is based either on a source-destination or source-destination-exchange ID policy. For FICON, only the source-destination policy is supported with guaranteed in-order-delivery of frames.

Traffic isolation
The term preferred path for FICON traffic was replaced by the use of static routes and Traffic Isolation (TI) zones (not defined by the Fibre Channel standard). Traffic Isolation allows data paths to be specified. It is used to: Separate disk and tape traffic Select traffic for diverse ISL routes Guarantee bandwidth for mission-critical data Traffic isolation can operate with failover enabled or disabled. Any traffic on ports that are not included in a TI zone follow the fabric shortest path first (FSPF) algorithm. TI zone can be used for failover traffic from other TI zones when the ISLs fail, and from ports that are not in a TI zone. It is recommended that you either put all ports in TI zones or do not use TI zoning at all. General rules for traffic isolation zones are: An N_Port can be a member of only a single TI zone. If administrative domains (ADs) are configured, this checking is done only against the current ADs zone database. An E_Port can be a member of only a single TI zone.

90

FICON Planning and Implementation Guide

If multiple E_Ports are configured that are on the lowest-cost route to a domain, load balancing is applied. Only source ports included in the zone are routed to zone E_Ports as long as other paths exist. If no other paths exist, the dedicated E_Ports are used for other traffic also.

Port fencing
Bad optics can cause errors to occur at a high enough frequency that error processing and sending and processing RSCNs can cause fabric performance problems. Port fencing allows the user to limit the number of errors that a port can receive by forcing a port offline when certain error thresholds are met. For FICON environments, port fencing is only set for CRC errors, Invalid Transmission Words, and Loss of Sync. The recommended thresholds are: Five errors per minute for CRC errors 25 errors per minute for Invalid Transmission Words Two errors per minute for Loss of Sync These settings are high enough to ignore occasional errors and transient errors due to recabling, but low enough to stop problematic optics from causing fabric issues. By default, the alarms are set to fence the port, log an alert, send and e-mail, and set an SNMP trap. In most FICON environments, only fencing the port and logging the alert are desired.

4.9.8 Traffic management

Virtual channel technology is the fundamental building block used to construct Adaptive Networking services. Virtual channels create multiple logical data paths across a single physical link or connection. They allocate their own resources (queues, flow control mechanisms such as buffer credits). Adaptive Networking can assign a Quality of Service (QoS) priority to individual data flows in each virtual fabric. It anticipates congestion and dynamically makes adjustments in the fabric so that application traffic continues to flow. Adaptive Networking introduces four new networking services: Quality of Service, traffic management, dynamic profiling, and resource recovery: QoS high, medium, and low priority are assigned to applications using zoning (Source ID/Destination ID pairs on WWN base). Automatic rate limiting and advanced queuing priority algorithms are implemented to remove congestion and dedicate bandwidth or assign data flow to specific path. Dynamic profiling provides knowledge about end-to-end resource use. Resource recovery prevents resource exhaustion and application disruption.

4.9.9 Evaluation tools

To maintain business-critical mainframe infrastructures you need the right tools, experience, and expertise. Without sufficient information, it is difficult to conduct capacity planning or predict the impact of FICON migration and other changes. For identifying performance bottlenecks, tracking the utilization levels of disk subsystems, channels, host adapters, and RAID ranks, you need additional performance data.

Chapter 4. Planning the FICON environment

91

Using the suite of evaluation tools, you can provide the data and insight that are needed to effectively manage mainframe storage performance, as well as optimize disk subsystem investments and fine-tune z/OS tape workloads. The tools are listed and explained here: Processor evaluation - IBM zCP3000 study Channel evaluation - IBM FICON aggregation study When planning to perform a capacity planning study for FICON channels, use the IBM System z Capacity Planning tool, zCP3000. zCP3000 is designed to support FICON channel aggregation (select channel candidates for aggregation onto FICON channels). Disk controller evaluation - IBM Disk Magic study This is limited to the storage FICON channel port and port busy utilization numbers for workloads and configurations. You can model, if additional channels are needed. IBM zHPF evaluation study This study is designed to quantify the amount of I/Os in an environment that appears to be zHPF-eligible. With this information, the impact/benefit of zHPF can be assessed for a specific situation or workload. For System z, the z/OS Resource Measurement Facility (RMF) can be used to analyze your environment. This provides online, interactive performance monitoring or alternatively, long-term overview reporting with post-processor reports. Some reports to assist in your analysis are: Channel path activity report This report identifies each channel path by identifier and channel path type and reports on the channel utilization by server and individual LPAR. Device activity report This report provides activity and performance information for selected devices. I/O queuing activity report This report provides information about the I/O configuration and activity rate, queue lengths, and percentages when one or more I/O components were busy. FICON Director activity report This report provides information about Director latency caused by port contention. Shared device activity report This report gives you an overall performance picture of disk and tape devices that are shared between MVS systems in a sysplex. For further information, refer to z/OS RMF Report Analysis, SC33-7991.

4.10 Prerequisites and interoperability

Device interoperability within a SAN is very important, because no single vendor can provide the entire infrastructure. In fact, most SAN environments have more than one server vendor and multiple storage vendors. SAN vendors provide comprehensive interoperability matrixes and have extensive experience designing, implementing, managing, and troubleshooting heterogeneous SANs. 92
FICON Planning and Implementation Guide

Directors providing connectivity for storage and servers must also have common services and settings to operate with stability and full performance. When interconnecting Directors, the name server, zoning database, FSPF, RSCNs, and time-out value information must be shared. Even if vendors follow Fibre Channel standards, coordinating fabric information between heterogeneous platforms is still an issue for fabric interoperability. Vendor implementation of the same standard could be at different stages and include varying enhancements, which can lead to fabric connectivity issues. Native connectivity is required in the following situations: Migration between two platforms Port optimization within a fabric Temporary platform merge for infrastructure change Double vendor policies Issues introduced by the native connectivity include: Management interface inconsistencies in fabric topology, node port states, zoning, and performance measurement Slow fabric recovery during adds, moves, and changes Failing fabric merges due to slight differences in standard implementations Name server synchronization issues Security features implementation Special vendor features to enhance fabric performance E_Port connectivity in these situations is used to form a single fabric. Using Fibre Channel Routing is an option to allow secure communication between hosts and storage in two or more separate, unmerged fabrics and to be managed independently. FCR is not supported by FICON. For this reason, any cascaded FICON Directors must be from the same vendor. For System z details, refer to Chapter 2, System z FICON technical description on page 13. For FICON Director details, refer to Chapter 3, FICON Director technical description on page 33. IBM System Storage Web site: http://www-03.ibm.com/systems/storage/san/index.html IBM System Storage SAN768B Fiber Backbone Interoperability Matrix IBM System Storage SAN384B Fiber Backbone Interoperability Matrix Cisco MDS 9506, 9509, 9513 for IBM System Storage Directors Interoperability Matrix For more details regarding IBM System z qualification letters, see: http://resourcelink.ibm.com

4.11 Physical connectivity

The physical layer for FICON infrastructure is based on fiber optic cabling. When planning this infrastructure you need to consider the following items: Existing physical layer of the data center Existing FICON components Planned FICON components Non-FICON components that also require fiber cabling Capacity planning of the data centers physical layer
Chapter 4. Planning the FICON environment

93

The supported standard media for System z FICON features are single mode OS1 and OS2, as well as multimode OM1, OM2, and OM3. With the increasing data rate on the link you need to utilize the latest fiber media types when designing the physical layer. New fiber media types provide more flexibility and better performance. A unified media platform is also very important inside the data center for simplified operations and troubleshooting. The highest flexibility and investment protection is with the single mode fiber. We recommend using the OS2 fiber type for System z FICON LX features, because of the lower attenuation compared to the OS1 fiber type. The System z FICON SX features support multimode fiber. For the best performance and investment protection for multimode, we recommend using the OM3 fiber type. There are higher bandwidth fiber cables on the market today, but their use could introduce a risk for interoperability and may not be supported by the equipment vendor. Note: OM1 and OM2 fiber types are not recommended for laser-based systems. When choosing the fiber media, we recommend using only qualified vendor fiber optic cabling components. Also, avoid mixing cabling components from different vendors even if they are all qualified. Qualified cabling components (connectors and cables) follow a standard color code. They are delivered with a measurement protocol directly from the manufacturer, which guarantees a high quality cable. This is important because the higher the bit rate on the channel, the more stringent the requirements for the fiber and the connectors. That means low attenuation, high return loss, and high number of cycles for reuse.

Data center cabling infrastructure

To allow for adding, moving, and changing devices inside a data center, multiple networks typically share a structured cabling infrastructure in a hierarchical star topology. Therefore, it is important to view the FICON interconnections as an integral part of the overall fiber cabling infrastructure. This means the passive cross-connect points and traces are shared. This type of approach also provides a highly scalable and flexible cabling infrastructure. At the hub of the hierarchical star topology is the Main Distribution Area (MDA). In IBM terminology it is the Central Patching Location, and it contains a large quantity of patch panels, placed in cabinets. The MDA is the central distribution point in the data center for telecommunication networks, local area networks (LANs), and SANs. It is the central cross-connect of the data center for all fiber optic and copper cables. The MDA uses backbone or horizontal cabling to extend connectivity to the Zone Distribution Area (ZDAs) in the server, storage, and Director environments of the data center. The use of high-density cabling between the ZDA and the MDA reduces congestion and improves manageability, scalability, modularity, and security in the data center infrastructure. All passive cross-connects, local distribution points, and equipment outlets are based on high density patch panels with a unified connector type for better scalability. We recommend using device-specific components, for example MTP brackets designed for System z. This provides a better starting point for cable routing inside the device rack. When planning the routing of cables inside the cabinets, remember to plan for a sufficient number of patch cords and length, according to the vertical and horizontal cable management

94

FICON Planning and Implementation Guide

inside the cabinets. Failure to do so can cause multiple outages during the operations of the devices, when moving or laying cables. There are various types of high density connectors available that offer differing parameters such as robustness, ease of use, number of channels, and fiber link parameters. We recommend using a unified connector type for the entire infrastructure. The standard-based high density connectors are LC, MU, and MTP. The IBM high-density connector type is referred to as the small form factor duplex fiber optic connector for the data center (SCDC). The SCDC was especially developed for the mainframe fiber optic cabling environment, starting with ESCON. When designing a fiber infrastructure, also consider the polarity management. We recommend using the standard-based ANSI/TIA-568-B.1-7 Method C. This is the default for the IBM Fiber Transport System. Each cabling infrastructure needs to be carefully documented. You have to consider the naming conventions for each environment specifically. When creating a naming convention for System z, keep in mind the format used in the FICON environment, for ease of use and understanding by the operators. When planning the optical interfaces, pay extra attention to the power level alignment between the receiver and the transmitter on the channel between two components. Low power levels do not necessarily mean a bad cable, but rather the potential of port errors and resending of frames, which in turn will impact performance. The FICON Express8 features utilize the existing single-mode and multimode cable plants. However, the 8 Gbps channel is more sensitive to the condition of the cable plant. The cable plant must satisfy the industry standard specification to minimize connector loss and reflections. We strongly recommend you thoroughly analyze the fiber optic link to ensure that the cable plant specifications (total cable plant optical loss as well as connectors and splices return loss) are being met for that link length. The most common source of cable plant optical link problems is associated with the various connections in the optical link. Dust, dirt, oil, or defective connections may cause a problem with high speed channels such as 8 Gbps, although lower link data rates such as 1 Gbps, 2 Gbps, or 4 Gbps may not be affected. If you are experiencing excessive bit errors on a link (regardless of the data link rate), we recommend that you first clean and reassemble the connections. Refer to IBM Fiber Optic Cleaning Procedure, SY27-2604 for the procedure and materials required. The cleaning is best performed by skilled personnel. The cleaning procedure may need to be performed more than once to ensure that all dust, dirt, or oil is removed. For more detailed information about fiber optic link testing and measurement, refer to System z Maintenance Information for Fiber Optic Links (ESCON, FICON, Coupling Links, and Open System Adapters), SY27-2597. For additional information, refer to the following URLs: http://www.tiaonline.org http://www.cenelec.eu http://www.ibm.com/services/siteandfacilities For an example of how you can use this decision-making process when planning a FICON environment, we provide a walkthrough of our FICON environment with a cascaded FICON Director topology in Appendix A, Example: planning workflow on page 243.

Chapter 4. Planning the FICON environment

95

96

FICON Planning and Implementation Guide

Part 3

Part

Configuring the FICON environment

This part provides you with a basic guide for configuring a FICON environment. We give step-by-step guidance with descriptions of the software components and definitions that are needed to make the FICON environment operational.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

97

98

FICON Planning and Implementation Guide

Chapter 5.

Configuring a point-to-point topology

This chapter describes all of the tasks that are required to configure and define a FICON environment on a point-to-point topology. Establishing a point-to-point topology involves the following items: Defining FICON channel, control unit, and device definitions in IOCP Activating the IOCDS Configuring the storage control unit Varying the channel path and device online Verifying the configuration

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

99

5.1 Establishing a point-to-point topology

This section describes the tasks that are required to configure a FICON point-to-point topology on a System z server, including configuring the attached storage control unit. Our discussion is based on an infrastructure that is already built and installed, including the physical installation of the following items: System z server DS8000 storage system Fiber cabling infrastructure

5.1.1 Description of our environment

Based on the considerations explained in Chapter 4, Planning the FICON environment on page 59, we are going to build a FICON point-to-point configuration that consists of a z10 EC server and a DS8000 storage device. In our case, all the required infrastructure is already installed in our environment. The configuration is shown Figure 5-1.

z10 EC
SCZP201
LPAR SC33 A29 z/OS V1R10
FICON Express8 LX CSS2 PCHID 5A3 CHPID 7F

DS8000
CU# D000
LX 0132

D0xx

FICON Express8 LX PCHID 5D3 CHPID 83

LX 0331

D1xx

CU# D100
* All cable connectors are LC Duplex type

Figure 5-1 Our point-to-point configuration

The z10 EC server (SCZP201) has one LPAR defined and activated. The system name running in this partition is SC33, which corresponds to the LPAR name A29. The operating system running in this partition is z/OS V1R10. As a standard feature of the z10 servers, zHPF protocol will be used for data transfer to the DS8000 storage device. Two FICON channels are defined as shared and spanned across CSS2 and CSS0 (only CCS2 is shown). The channels are FICON Express8 LX features (FC 3325), using PCHID 5A3 (CHPID 7F) and PCHID 5D3 (CHPID 83). An IBM System Storage DS8000 will be connected to the z10 server via CHPID 7F and CHPID 83. The two host adapters installed in the DS8000 are longwave (LX) laser. Port numbers 0132 and 0331 in the DS8000 are used to connect to the server. Two logical control

100

FICON Planning and Implementation Guide

units (D000 and D100) are defined, which have devices D000-D0FF and D100-D1FF assigned. The DS8000 will have the zHPF feature enabled. A maximum unrepeated distance of 10km (6.2 miles) is supported by the longwave laser (LX) feature when using 9m single mode (SM) fiber optic cable. The fiber optic cables have an LC Duplex connector at both ends to connect to the z10 FICON Express8 channels and to the DS8000 host adapters.

5.1.2 Tasks and checklist

This section provides an overview of required tasks and the checklist that we used to make sure all hardware and software prerequisites were met before starting the configuration tasks.

Tasks
Figure 5-2 shows the main steps required to define and activate a FICON point-to-point configuration.
Follow the verification checklist to ensure that all HW and SW prerequisites are met. Go to Verification checklist on page 101.

Verification checklist

Define channel, CU, and devices on a System z server

Information about defining the channel paths, control units, and devices is given in Defining the channel, CU, and storage device in IOCDS on page 103.

Configure storage CU

The configuration tasks for DS8000 storage system are described in Configuring the IBM Storage System DS8000 on page 108.

Plug fiber cables

Information about fiber optic cables and plugging rules to achieve the desired configuration is given in Connecting the fiber optic cables on page 108.

Verify configuration

Information about how to verify that your actual configuration matches the desired configuration is given in Verifying the installation on page 110.

Figure 5-2 Main steps for configuring and verifying a FICON point-to-point configuration

Verification checklist
Before configuring the point-to-point topology shown in Figure 5-1 on page 100, the following list was checked. All steps in the checklist must be finished and corrected (if required) to ensure a smooth and successful configuration of the topology. Both hardware and software requirements must be checked. Check that the appropriate FICON features are available on the System z server. For details about each feature, see System z FICON feature support on page 25. FICON Express2 LX FC 3319
Chapter 5. Configuring a point-to-point topology

101

SX FC 3320 FICON Express4 LX FC 3321, FC 3323, FC 3324 SX FC 3318, FC 3322 FICON Express8 LX FC 3325 SX FC 3326 If using the FICON Express8 feature, check the System z operating system requirements. z/OS V1.7 with the IBM Lifecycle Extension for z/OS V1.7 (5637-A01) is required to support the FICON Express8 feature in the z10 server. Check the 2097DEVICE PSP bucket for the latest information about FICON Express8 support on the operating system. Check DS8000 storage hardware requirements to support FICON longwave (LX) or shortwave (SX) connectivity. Check that FC 0709 and FC 7092 are installed to support zHPF. Note: FC 0709 and FC 7092 must be ordered to obtain the license key to enable zHPF support in a DS8000 storage controller. Check that DS8000 firmware is at level 6.1.16.0 or higher to support the zHPF feature. Check that the zHPF feature is enabled in the DS8000. Check that the correct fiber optic cables are available to connect the DS8000 storage controller to the z10 server. A 9m single mode (SM) fiber optic cable is required to support longwave laser (LX) for a maximum distance of 10km (6.2 miles). A 50m or 62.5m multi-mode (MM) fiber optic cable is required to support shortwave laser (SX). See System z FICON feature support on page 25 for the maximum supported distance depending on cable type and speed. An LC duplex connector is required at both ends of the fiber optic cables to connect to the System z server FICON adapter and the DS8000 host adapter. Note: All fiber optic cables used in a link must be of the same type. For example, they must be either all single mode or all multi-mode fiber optic cables.

5.1.3 Getting started

This section describes all the tasks that are required to reach the designated configuration shown in Figure 5-1 on page 100. It explains what needs to be done to get the FICON channels, storage control units, and the storage device online and operating. The steps include: Defining the channel, CU, and storage device in IOCDS Activating the changes (IOCDS) and configuring online Configuring the IBM Storage System DS8000 Connecting the fiber optic cables Configuring the CHPIDs online Verifying the installation 102
FICON Planning and Implementation Guide

For our scenario, we had an active partition in a System z server running z/OS with HCD. We used HCD to create, save, and activate our I/O definitions.

Defining the channel, CU, and storage device in IOCDS

There are several tools available that you can use to define and configure a FICON environment on a System z server. The definition tools and an explanation of the statements and keywords used in the IOCP are described in Appendix C, Configuration and definition tools on page 255. Only statements and keywords relevant to our desired configuration are explained. Refer to IOCP Users Guide, SB10-7037, for details about all available IOCP statements and keywords. HCD was used to define the CHPIDs, CUs, and devices for the sample configuration we describe in this section. Refer to HCD Users Guide, SC33-7988, for detailed information regarding the use of HCD. After all definition tasks are completed, the definition data is stored in the I/O definition file (IODF) data set. Any time you make changes to the I/O configuration, the IOCDS has to be updated to ensure consistency between the software and the microcode in the hardware. The newly created IOCDS will be sent to the Support Element of the system server. The Support Element in the System z server stores the IOCDS on its built-in HDD. The IOCDS stored in the Support Element is loaded into the servers storage during a Power-on Reset (POR) of the System z server. Changes or additions of channel paths, CUs, or devices can be performed dynamically. Because no system outage is required for dynamic changes, this is the preferred method for changing I/O definitions. If changes to the I/O definitions are performed dynamically, those changes are directly updated into the Hardware System Area (HSA) of the server. The HSA is a separate storage area in the System z server which holds all I/O specifications and provides status information for all defined I/Os. The operating system uses the HSA to manage the I/O subsystem. For more details about System z channel subsystem characteristics, refer to 4.8.1, System z on page 78. The following examples show the CHPID, CNTLUNIT, and IODEVICE statements when all the definition tasks based on the configuration shown in Figure 5-1 on page 100 are finished. First we explain how the CHPIDs are defined. See Example 5-1 for the CHPID statement and its keywords.
Example 5-1 CHPID definition for point-to-point configuration

CHPID PATH=(CSS(2),7F),SHARED,PARTITION=((A29),(=)),PCHID=5A3,* TYPE=FC CHPID PATH=(CSS(2),83),SHARED,PARTITION=((A29),(=)),PCHID=5D3,* TYPE=FC The PATH keyword in the CHPID statement defines CHPID 7E in CSS 2 as SHARED. The LPAR name (A29) is specified in the PARTITION keyword to use CSS 2 to access CHPID 7F. LPAR names specified in ICODS correspond to system name SC33. The PCHID 5A3 has been merged to the IOCDS by the CMT, which assigned CHPID 7F to PCHID 5A3. The TYPE of the channel is specified as FC (native FICON). The same definition rules apply to the definition of CHPID 83.

Chapter 5. Configuring a point-to-point topology

103

With the CHPIDs defined, we next show how the CUs attached to the CHPIDs are defined; Example 5-2 displays the CNTLUNIT statement and keywords.
Example 5-2 CNTLUNIT definition for point-to-point configuration

CNTLUNIT CUNUMBR=D000, * PATH=((CSS(0),50,54,58,5C),(CSS(1),50,54,58,5C),(CSS(2),* 50,54,58,5C,7E,82,7F,83),(CSS(3),50,54,58,5C)), * UNITADD=((00,256)), * LINK=((CSS(0),2C,2C,0A,0A),(CSS(1),2C,2C,0A,0A),(CSS(2),* 2C,2C,0A,0A,6688,66CF,**,**),(CSS(3),2C,2C,0A,0A)), * CUADD=0,UNIT=2107 CNTLUNIT CUNUMBR=D100, * PATH=((CSS(0),51,55,59,5D),(CSS(1),51,55,59,5D),(CSS(2),* 51,55,59,5D,7E,82,7F,83),(CSS(3),51,55,59,5D)), * UNITADD=((00,256)), * LINK=((CSS(0),28,28,0E,0E),(CSS(1),28,28,0E,0E),(CSS(2),* 28,28,0E,0E,6688,66CF,**,**),(CSS(3),28,28,0E,0E)), * CUADD=1,UNIT=2107 There are two control units defined by the CUNUMBR keyword: D000 and D100. The logical CU image number 0 is specified for control unit D000. For control unit D100, the logical CU image number 1 is specified in the CUADD keyword. CHIPDs 7F and 83 have access to CUs D000 and D1000 because both CHPIDs (among others) are specified in CSS2 in the PATH keyword for both CUs. The LINK keyword is required because some other FICON channels have access to the CUs through a FICON Director. No link address is required for CHPID 7F and 83. Two asterisks (**) are specified in the LINK keyword for point-to-point connectivity. After the CHPIDs and the CUs are defined, the next step is to define the devices owned by the CUs; Example 5-3 displays the IODEVICE statement and keywords.
Example 5-3 IODEVICE definition for point-to-point configuration

IODEVICE IODEVICE IODEVICE IODEVICE

ADDRESS=(D000,113),CUNUMBR=(D000),STADET=Y,UNIT=3390B ADDRESS=(D071,143),CUNUMBR=(D000),STADET=Y,UNIT=3390A ADDRESS=(D100,113),CUNUMBR=(D100),STADET=Y,UNIT=3390B ADDRESS=(D171,143),CUNUMBR=(D100),STADET=Y,UNIT=3390A

The IODEVICE statement specifies the characteristics of devices D000 and D100. For control unit D000 there are 113 devices defined, starting at device address D000. The same definitions are done for devices associated to control unit D100. These are base devices in the storage unit. In addition, 143 Hyper PAV alias devices are defined starting at device address D071 and D171.

Defining a FICON Channel-to-Channel connection

In addition to storage control units and storage devices, we have defined FICON Channel-to-Channel (FCTC) CUs and devices to allow data exchange between LPARs. For more information about considerations regarding FCTCs, refer to Appendix G, Adding FICON CTC connections on page 305.

104

FICON Planning and Implementation Guide

Based on those considerations, we are going to configure FCTCs in a point-to-point FICON environment. Figure 5-3 illustrates the desired FCTC configuration. To simplify the illustration only four CUs and two LPARs are shown in this sample configuration, although more than four CUs could be defined.

z10 EC
SCZP201
LPAR SC30 A23 z/OS V1R10
FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E

CU# CU# CU# CU# 4044 5044 4034 5034

LPAR SC31 A24 z/OS V1R10

FICON Express8 LX CSS2 PCHID 5D2 CHPID 82

* All cable connectors are LC Duplex type

Figure 5-3 FCTC configuration point-to-point

Both LPARs have access to CHPID 7E and 82 because the CHPIDs are defined as shared. There is a direct connection between CHPID 7E and CHPID 82 by a fiber optic cable. Each LPAR will have access to a logical control unit and their logical devices defined in the corresponding LPAR. LPAR A23 will communicate via a FICON channel to CU 5034, which is assigned to LPAR A24. LPAR A24 will communicate via a FICON channel to CU 4044, which is assigned to LPAR A23. Example 5-4 shows the definition of FCTC control units and devices.
Example 5-4 FCTC definitions point-to-point

CNTLUNIT CUNUMBR=5034,PATH=((CSS(2),7E)),UNITADD=((04,004)), CUADD=23,UNIT=FCTC IODEVICE ADDRESS=(5034,004),UNITADD=04,CUNUMBR=(5034), STADET=Y,PARTITION=((CSS(2),A24)),UNIT=FCTC CNTLUNIT CUNUMBR=4044,PATH=((CSS(2),82)),UNITADD=((04,004)), CUADD=24,UNIT=FCTC IODEVICE ADDRESS=(4044,004),UNITADD=04,CUNUMBR=(4044), STADET=Y,PARTITION=((CSS(2),A23)),UNIT=FCTC

* *

* *

One control unit is defined for each LPAR: CU 5034 for LPAR A23, and CU 4044 for LPAR A24. The PARTITION keyword in the IODEVICE statement specifies the logical partition that can access the device using the FICON channel. There are four devices specified for each LPAR (LPAR A23 and LPAR A24). This allows data traffic to flow between both LPARs using the corresponding devices: SC30 SC31 4044 <---> 5034
Chapter 5. Configuring a point-to-point topology

105

4045 <---> 5035 4046 <---> 5036 4047 <---> 5037 The rules we followed for device numbering are described in FCTC device numbering scheme on page 306. Figure 5-4 illustrates the logical view of the FCTC configuration and the data path between LPARs. To simplify the illustration, only two logical CUs are shown.

LPAR A23
Data transfer

CU 4044

7E

82

CU 5034

Data transfer

LPAR A24
Figure 5-4 FCTC data transfer (point-to-point)

Data is transferred over the FICON link in both directions between the CUs and the logical partitions (LPARs). LPAR A23 sends data to A24 via CU 5034. LPAR A23 receives data from A24 via CU 4044. The reverse applies to LPAR A24 when data is sent to or received from A23.

Activating the changes (IOCDS) and configuring online

After the definition of the desired configuration is finished, the data must be saved into the IODF data set and the IOCDS file on the Support Element. Up to four IOCDS files can be stored in a System z server. This allows to you plan and define future I/O configurations and store them on the Support Element of the server. The tasks that are required to save definition data and activate changes dynamically are performed via HCD. Refer to HCD Users Guide, SC33-7988, for detailed descriptions of all activation procedures. On the HCD start menu, select option 2 (Activate or process configuration data). Figure 5-5 on page 107 shows the Activate or Process Configuration Data menu.

106

FICON Planning and Implementation Guide

Figure 5-5 HCD Activate or Process Configuration Data menu

The options used to activate the changes are: Option 1. Build production I/O definition file Option 2. Build ICODS Option 6. Activate or verify configuration dynamically The Build production I/O definition file (IODF) function (option 1) saves definition data into an IODFxx dataset, where xx in the data set name is a suffix used to identify the IODF among other IODFs which may already exist. The suffix of the IODF data set is used in the IPL parameter when the z/OS image is brought up. This ensures that the desired IODF used by the operating system image matches the IOCDS that is loaded into the HSA. Using Build IOCDS (option 2) allows you to build an IOCDS and save it to the CPC. The IOCDS is built using the definitions saved in the IODFxx data set. Using Activate or verify configuration dynamically (option 6) allows you to dynamically activate changes in the I/O definitions. There is no Power-on Reset (POR) required because changes are directly updated to the HSA in the CPC. This is the preferred method for activating changes. Another way to activate changes in I/O definitions is to perform a POR at the CPC with the newly-built IOCDS. Keep in mind, however, that a POR causes a system outage. For this reason, if a POR is required, then we suggest that you plan ahead to make any changes in I/O definitions so that the new IOCDS is ready for use at the next POR. Changes can be done at any time and saved in an IODF data set. The newly-built IOCDS based on the I/O definition in the IODF should be sent to the CPCs Support Element and marked for use at the next POR. This ensures that the newly-built IOCDS will be used the next time the CPC is activated. To verify which I/O definitions (IODFs) are currently active, enter the D IOS,CONFIG command at the z/OS operator console. The response to the command is shown in Example 5-5.
Chapter 5. Configuring a point-to-point topology

107

Example 5-5 D IOS,CONFIG command

IOS506I 15.20.05 I/O CONFIG DATA 345 ACTIVE IODF DATA SET = SC33.IODF01 CONFIGURATION ID = TEST2094 EDT ID = 01 TOKEN: PROCESSOR DATE TIME DESCRIPTION SOURCE: SCZP201 09-05-11 20:50:41 SC33 IODF01 ACTIVE CSS: 2 SUBCHANNEL SETS CONFIGURED: 0, 1 CHANNEL MEASUREMENT BLOCK FACILITY IS ACTIVE Example 5-5 shows that the IODF01 data set is currently active in z/OS image SC33. It also provides information about the CSSs used in this z/OS image. CSS2 is the active CSS in system SC33. Compare the information provided by the D IOS,CONFIG command with the data of the ID statement in the IOCDS, shown in Example 5-6.
Example 5-6 ID statement in IOCDS

ID

MSG1='IODF01',MSG2='SC33.IODF01 - 2009-05-11 20:50', * SYSTEM=(2097,1),LSYSTEM=SCZP201, * TOK=('SCZP201',00800006991E2094205041680109131F00000000,* 00000000,'09-05-11','20:50:41','SC33','IODF01')

The suffix of the IODF data set shown in the IOCDS ID statement and displayed by the D IOS command should be the same number (that is, IODF01). This will prove that the IOCDS was built from the current active IODF data set and that the latest changes and new IO definitions were successfully activated.

Configuring the IBM Storage System DS8000

The host adapter ports in the DS8000 storage control unit have to be configured for native FICON protocol. Follow the procedure described in Appendix D, Configuring the DS8000 for FICON on page 259 to configure the host adapter ports. The configuration steps are based on the configuration shown in Figure 5-1 on page 100. For planning and implementation information, refer to IBM System Storage DS8000: Architecture and Implementation, SG24-6786.

Connecting the fiber optic cables

After all definition and configuration tasks are completed, fiber optic cables have to be plugged into the FICON channels and CU host adapters. All fiber links between the FICON channel and the CU host adapters use the longwave (LX) type of laser. For the LX type of laser, using a 9m single mode (SM) fiber optic cable is recommended to achieve maximum performance and distance. All fiber optic cables must have LC duplex connectors at both ends. Plug the fiber optic cable connectors to the designated ports, shown in Figure 5-1 on page 100, to the FICON channel and the CU ports, as listed here: PCHID 5A3 to DS8000 host adapter port 0003 PCHID 5D3 to DS8000 host adapter port 0242 Make sure the fiber optic cable ends are cleaned before plugging them into the transceivers. Also ensure that all fiber optic cables are labeled and documented for later use (for example, when recabling or for problem determination).

108

FICON Planning and Implementation Guide

For further information and considerations regarding fiber cabling and documentation, refer to 4.11, Physical connectivity on page 93 and 4.2, Documentation on page 62.

Configuring the CHPIDs online

After you successfully activate the new I/O definitions, you can configure the channel path online and try to communicate to the devices. Note that even if all the definitions in the IOCDS are correct, the channel status might be offline; therefore, enter D M=CHP(7F) on the system console to display the current status of CHPID 7F. If the CHPID is offline, enter CF CHP(7F),ON to configure the channel online. The channel status should change to ONLINE, as shown in Example 5-7.
Example 5-7 CF CHP(7F)

CF CHP(7F) IEE502I CHP(7F),ONLINE IEE712I CONFIG PROCESSING COMPLETE To achieve the best performance on the FICON channel, make sure that zHPF is enabled. Refer to 4.9.6, High Performance FICON on page 88 for considerations regarding how to exploit zHPF on System z servers. Enter D IOS,ZHPF at the z/OS console to display zHPF settings for the z/OS image. If zHPF is disabled, enter SETIOS ZHPF,YES. This enables zHPF, as shown in Example 5-8.
Example 5-8 Enabling zHPF

D IOS,ZHPF RESPONSE=SC33 IOS630I 13.27.23 ZHPF FACILITY 021 HIGH PERFORMANCE FICON FACILITY IS ENABLED Using the SETIOS ZHPF=YES command enables zHPF temporarily. However, after the next system IPL, the zHPF facility will be reset to the default (disabled). To permanently enable zHPF for z/OS, change the FCX parameter in the SYS1.PARMLIB member IECIOSxx to FCX=YES. Now you can query the status and functional details of the channel by entering D M=CHP(7F) at the operator console. The command output is shown in Example 5-9. It provides information about the channel and the attached devices.
Example 5-9 D M=CHP(7F)

D M=CHP(7F) IEE174I 11.58.59 DISPLAY M 290 CHPID 7F: TYPE=1A, DESC=FICON POINT TO POINT, DEVICE STATUS FOR CHANNEL PATH 7F 0 1 2 3 4 5 6 7 8 9 A B C D 0D00 + + + + + + + + + + + +@ +@ + 0D01 + + + + + + + + + + + + + + 0D02 + + + + + + + + + + + + + + ... 0DFD HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFE HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFF HA HA HA HA HA HA HA HA HA HA HA HA HA HA SWITCH DEVICE NUMBER = NONE ATTACHED ND = 002107.922.IBM.75.0000000BALB1

ONLINE E + + + F + + +

HA HA HA HA HA HA

Chapter 5. Configuring a point-to-point topology

109

PHYSICAL CHANNEL ID = 05A3 FACILITIES SUPPORTED = ZHPF Example 5-9 on page 109 shows that CHPID 7F is online and operating in a point-to-point configuration. Information is also displayed about the attached devices and the facilities (for example, zHPF) supported by the channel. To verify that communication to the attached devices is working properly, enter D M=DEV(xxxx), where xxxx is any device number. For example, to check the status of device D000, which is a storage device, enter D M=DEV(D000) on a z/OS console (see Example 5-10).
Example 5-10 D M=DEV(D000)

D M=DEV(D000) IEE174I 12.08.55 DISPLAY M 296 DEVICE D000 STATUS=ONLINE CHP 50 54 58 5C 7E 82 7F ENTRY LINK ADDRESS 0F 0F 1B 1B 6503 651B .. DEST LINK ADDRESS 2C 2C 0A 0A 6688 66CF 0D PATH ONLINE N N N N N N Y CHP PHYSICALLY ONLINE N N N N N N Y PATH OPERATIONAL N N N N N N Y MANAGED N N N N N N N CU NUMBER D000 D000 D000 D000 D000 D000 D000 MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107.900.IBM.75.0000000BALB1.0132 SCP TOKEN NED = 002107.900.IBM.75.0000000BALB1.0000 SCP DEVICE NED = 002107.900.IBM.75.0000000BALB1.0000 HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW, ZHPF

83 .. 0D Y Y Y N D000

The response to D M=DEV(D000) displays all available paths to the devices and their status. Information about the device (for example, device type), the control unit number, and the functions supported by the device (for example, MIDAW and zHPF) is shown.

Verifying the installation

After all the configuration tasks in the z10 server and the DS8000 storage are complete, you can verify that the current configuration matches the desired configuration shown in Figure 5-1 on page 100. This verification should be done to ensure that the fiber optic cables are correctly plugged to the designated ports in the System z server and the DS8000, and to check the status of channels and devices. First, check the status of PCHID 5A3 in the z10 server. From any HMC where the desired CPC (server) is defined (for example, SCZP201), start a Single Object Operation (SOO) to that CPC. Display the list of channels installed on the CPC. Search the list for the desired PCHID (5A3) and double-click the PCHID icon. This will display the PCHID details panel shown in Figure 5-6.

110

FICON Planning and Implementation Guide

Figure 5-6 PCHID details

Essential information for PCHID 5A3 is shown on the PCHID Details panel: The PCHID status is Operating. The PCHID type is FICON Express8. CSS.CHPID 2.7F is assigned to PCHID 5A3. The owning image of PCHID 5A3 is A29. The CHPID assigned to PCHID 5A3 is shared. Similar information can be found on the CHPID details panel, shown in Figure 5-7.

Figure 5-7 CHPID details

The CHPID Details panel provides information that is similar to the PCHID details panel. To display the CHPID details, select the channel list for an operating system image from the Support Element workplace. Notice that the information for PCHID 5A3 (CHPID 2.7F) provided on the details window matches the designated configuration. This proves that PCHID 5A3 has CHPID 2.7F assigned and that image A29 can access the channel. Repeat these checks on other channels that were recently configured. Important: If any of the data displayed on the PCHID or CHPID detail panel does not match the desired configuration, you must correct the definitions in the IOCDS.

Chapter 5. Configuring a point-to-point topology

111

Next, check that the FICON channels are connected to the correct host adapter port in the DS8000 storage controller. On either the PCHID details or CHPID details panel, click the Channel Problem Determination button. This will display the Channel Problem Determination panel, where you can select which information you want to display (see Figure 5-8).

Figure 5-8 Channel Problem Determination panel

Select Analyze channel information and click OK. The Analyze Channel Information window is displayed, as shown in Figure 5-9 on page 113, which provides information about the node attached to the FICON channel.

112

FICON Planning and Implementation Guide

Figure 5-9 Analyze Channel Information

Information about the nodes is displayed at the bottom part of the window. The lower left side displays information about the node in the System z server. The lower right side displays information about the attached node. Important: Make sure that the Node status for both nodes is displayed as Valid. If any other status is shown, none of the displayed information is valid. Check that the Type/model information as well as the serial number (Seq. number) is as expected. Next, examine the Tag field for each node. The Tag provides information about the port number of the attached node. The two right-most digits of the Tag value represent the CHPID number for the channel node (7F). For the attached node, the four digits represent the port number (0132). Be aware, however, that the tag value is provided by the attached device during link initialization and may have different meanings, depending on the vendor. The World Wide Node Name (WWN) and the World Wide Port Name (WWP) are also shown for each port, and may be used to prove that the channel is connected to the correct FICON adapter port if the WWN or WWP of the attached device is known. If the node status is not Valid or the Tag value and WWPN value are not correct, check the fiber optic cable link between the z10 server and the FICON Director to ensure that it is plugged correctly.

Chapter 5. Configuring a point-to-point topology

113

In our scenario we are now sure that PCHID 5A3 has connectivity to port number 0132 at the DS8000 host adapter, which matches our desired configuration (see Figure 5-1 on page 100). If the displayed values are not as expected, the fiber optic cables may not be plugged correctly and must be checked. After completing the preceding steps and proving that the physical path to the DS8000 storage controller and the logical definitions of the link are correct, check that the path to the control unit image is initialized correctly and properly defined. On the Channel Problem Determination panel, shown in Figure 5-8 on page 112, select Analyze Serial Link Status and click OK. The Analyze Serial Link Status window is displayed, as shown in Figure 5-10.

Figure 5-10 Analyze Serial Link status

The Analyze Serial Link Status window provides status information about the link to the control unit images defined in the IOCDS. Scroll through the list of CU images and check that the status for all CUs displays as Initialization Complete. Figure 5-10 shows a link status of Initialization Complete for all defined CU images on CPID 7F. Although there is no FICON Director attached, a link address of 0D is displayed by default for a point-to-point FICON link. If the link status Initialization Complete is not shown, you must check that the ports in the CU are correctly configured, and that the fiber optic cable link to the CU has the correct cable type and plugging.

114

FICON Planning and Implementation Guide

Chapter 6.

Configuring a switched point-to-point topology

This chapter describes all the tasks that are required to configure and define a FICON environment on a switched point-to-point topology. Establishing a switched point-to-point topology involves the following: Defining FICON channel, control unit, and device definitions in IOCP Activating the IOCDS Configuring the storage control unit Varying the channel path and device online Verifying the configuration

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

115

6.1 Establishing a switched point-to-point topology

This section describes the tasks that are required for configuring a FICON switched point-to-point topology on a System z server, including setup of a FICON Director and configuring the attached storage controller. Our discussion is based on an infrastructure that is already built and installed, including the physical installation of the following: System z server DS8000 storage system FICON Director Fiber cabling infrastructure

6.1.1 Description of our environment

Based on considerations explained in Chapter 4, Planning the FICON environment on page 59, we are going to build a FICON switched point-to-point configuration, which consists of an IBM z10 EC server, an IBM System Storage SAN768B FICON Director, and an IBM DS8000 System Storage device. In our case, all the infrastructure is already installed in our environment. The configuration is shown Figure 6-1.

z10 EC
SCZP201
LPAR SC30 (A23) z/OS V1R10
FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E

System Storage SAN768B

Port 03 Port 28

DS8000
CU# D000
LX 0003

D0xx

Switch # 65 Switch @ 65

LPAR SC31 (A24) z/OS V1R10

FICON Express8 LX CSS2 PCHID 5D2 CHPID 82

Port 1B

Port 3F

LX 0242

D1xx

CU# D100
* All cable connectors are LC Duplex type

Figure 6-1 Switched point-to-point configuration

The z10 EC server (SCZP201) has two LPARS defined and activated. The system names running in these partitions are SC30 and SC31, which correspond to the LPAR names A23 and A24, respectively. The operating system running in both partitions is z/OS V1R10. As a standard feature of the z10 server, zHPF protocol will be used for data transfer to the DS8000 storage device. Two FICON channels are defined as shared and in CSS2. Both LPARs have access to the two FICON channels. The channels are FICON Express8 LX features (FC 3325) using PCHID 5A2 (CHPID 7E) and PCHID 5D2 (CHPID 82).

116

FICON Planning and Implementation Guide

The SAN768B FICON Director has connectivity to the z10 server on port 03 and port 1B, and connectivity to the DS8000 storage system on port 28 and 3F. All the ports in the FICON Director are longwave (LX) ports. The switch number (Switch #) and the switch address (Switch @) are both set to 65. Although the switch number is specified in the IOCP, the switch address is the Domain ID specified in the FICON Director. The two host adapters installed in the DS8000 are longwave (LX) lasers. Port number 0003 and port number 0242 are used to connect to the FICON Director. Two logical control units (D000 and D100) are defined and have devices D000-D0FF and D1000-D1FF assigned. The DS8000 will have the zHPF feature enabled. A maximum unrepeated distance of 10 km (6.2 miles) is supported by the longwave laser (LX) feature when using 9m single mode (SM). The fiber optic cables have an LC Duplex connecter at both ends to connect to the z10 FICON Express8 channels; to the ports in the FICON Director; and to the DS8000 host adapter.

6.1.2 Tasks and checklist

This section provides an overview of the required tasks and a checklist that we used to make sure all hardware and software prerequisites were met before starting the configuration tasks.

Tasks
Figure 6-2 shows the main steps required to define and activate a FICON switched point-to-point configuration.

Verification checklist

Follow the verification checklist to ensure that all hardware and software prerequisites are met. Go to Verification checklist on page 118. Information about defining the channel paths, control units, and devices is given in Defining the channel, CU, and device in IOCP on page 119.

Define channel, CU, and devices on a System z server

Configure storage CU

The configuration tasks for a DS8000 storage system are described in Configuring the IBM Storage System DS8000 on page 125. All required steps to set up the FICON Director are shown in 8.1.1 Configuration flowchart on page 152.

Configure FICON director

Plug fiber cables

Information about fiber optic cables and plugging rules is given in Connecting the fiber optic cables on page 126.

Verify configuration

Information about how to verify that your actual configuration matches the desired configuration is given in Verifying the installation on page 128.

Figure 6-2 Main steps for configuring and verifying a FICON switched point-to-point configuration

Chapter 6. Configuring a switched point-to-point topology

117

Verification checklist
Before configuring the switched point-to-point topology shown in Figure 6-1 on page 116, the following list was checked. All steps in the checklist must be finished and corrected (if required) to ensure a smooth and successful configuration of the topology. Hardware and software requirements must be checked. Check that the appropriate FICON features are available on the System z server. For details about each feature, refer to System z FICON feature support on page 25. FICON Express2 LX FC 3319 SX FC 3320 FICON Express4 LX FC 3321, FC 3323, FC 3324 SX FC 3318, FC 3322 FICON Express8 LX FC 3325 SX FC 3326 If using FICON Express8 features, check System z operating system requirements. z/OS V1.7 with the IBM Lifecycle Extension for z/OS V1.7 (5637-A01) is required to support the FICON Express8 feature in z10 server. Check the 2097DEVICE PSP bucket for the latest information about FICON Express8 support on the operating system. Check that the number and types of FICON Director ports match the configuration requirements. LX SFPs SX SFPs Check DS8000 storage hardware requirements to support FICON longwave (LX) connectivity to the z10 server. Check that FC 0709 and FC 7092 are installed to support zHPF. Note: FC 0709 and FC 7092 must be ordered to obtain the license key to enable zHPF support in a DS8000 storage controller. Check that the DS8000 firmware is at level 6.1.16.0 or higher to support the zHPF feature. Check that the zHPF feature is enabled in the DS8000. Check that the correct fiber optic cables are available to connect the FICON Director to the System z server, and the DS8000 storage to the FICON Director. A 9m single mode (SM) fiber optic cable is required to support longwave laser (LX) for a maximum distance of 10km (6.2 miles). A 50m or 62.5m multi mode (MM) fiber optic cable is required to support shortwave laser (SX). See System z FICON feature support on page 25 for the maximum supported distance depending on cable type and speed. An LC duplex connector is required at both ends of the fiber optic cables to connect to the z10 server FICON channel, the ports in the FICON Director, and the DS8000 host adapter ports.

118

FICON Planning and Implementation Guide

Note: All fiber optic cables used in a link must be of the same type. For example, they must be either all single mode or all multi-mode fiber optic cables.

6.1.3 Getting started

This section describes all the tasks that are required to achieve the designated configuration shown in Figure 6-1 on page 116. It explains what needs to be done to get the FICON channels, FICON Director, storage control units, and the storage device online and operating. The steps described are: Defining the channel, CU, and storage device in IOCP Activating the changes (IOCDS) and configuring online Configuring the IBM Storage System DS8000 Configuring the FICON Director Connecting the fiber optic cables Configuring the CHPIDs online Verifying the installation For our scenario, we had an active partition in a System z server running z/OS with HCD. We used HCD to create, save, and activate our I/O definitions.

Defining the channel, CU, and device in IOCP

There are several tools available that you can use to define and configure a FICON environment on a System z server. The definition tools and an explanation of the statements and keywords used in the IOCP are described in Appendix C, Configuration and definition tools on page 255. Only statements and keywords relevant to our configuration are explained. Refer to IOCP Users Guide, SB10-7037, for details about all available IOCP statements and keywords. HCD was used to define the CHPIDs, CUs, and devices for the sample configuration we describe in this section. Refer to HCD Users Guide, SC33-7988, for detailed information regarding the use of HCD. After all definition tasks are completed, the definition data is stored in the I/O definition file (IODF) data set. Any time you make changes to the I/O configuration, the IOCDS has to be updated to ensure consistency between the software and the microcode in the hardware. The newly created IOCDS will be sent to the Support Element of the System z server. The Support Element in the server stores the IOCDS on its built-in HDD. The IOCDS stored in the Support Element is loaded into server storage during a Power-on Reset (POR) of the System z server. Changes or additions of channel paths, CUs, or devices can be performed dynamically. Because no system outage is required for dynamic changes, this is the preferred method for changing I/O definitions. If changes to the I/O definitions are performed dynamically, the changes are directly updated into the Hardware System Area (HSA) of the System z server. The HSA is a separate storage area in the server which holds all I/O specifications and provides status information for all defined I/Os. The operating system uses the HSA to manage the I/O subsystem. For more details about System z channel subsystem characteristics, refer to 4.8.1 System z on page 78. The following examples show the CHPID, CNTLUNIT, and IODEVICE statements if all the definition tasks based on the configuration shown in Figure 6-1 on page 116 are finished.
Chapter 6. Configuring a switched point-to-point topology

119

First we explain how the CHPIDs are defined. See Example 6-1 for the CHPID statement and its keywords.
Example 6-1 CHPID definition for switched point-to-point configuration

CHPID PATH=(CSS(0,2),7E),SHARED, * PARTITION=((CSS(0),(A01),(=)),(CSS(2),(A23,A24,A25,A29),* (=))),SWITCH=65,PCHID=5A2,TYPE=FC CHPID PATH=(CSS(1,2),82),SHARED, * PARTITION=((CSS(1),(A11),(=)),(CSS(2),(A23,A24,A25,A29),* (=))),SWITCH=65,PCHID=5D2,TYPE=FC The PATH keyword in the CHPID statement defines CHPID 7E in CSS 2 as shared. The LPAR names A23 and A24 are specified in the PARTITION keyword to use CSS 2 to access CHPID 7E. LPAR names specified in IOCDS correspond to system names SC30 and SC31. The SWITCH keyword is used to provide the switch number (65) of the Director to which the channel is connected. The PCHID 5A2 has been merged to the IOCDS by the CMT which assigned CHPID 7E to PCHID 5A2. The TYPE of the channel is specified as FC (native FICON). The same definition rules apply to the definition of CHPID 82. With the CHPIDs defined, we next show how the CUs attached to the CHPIDs are defined. Example 6-2 displays the CNTLUNIT statement and keywords.
Example 6-2 CNTLUNIT definition for switched point-to-point configuration

CNTLUNIT CUNUMBR=0065,PATH=((CSS(2),7E,82)), * UNITADD=((00,001)),LINK=((CSS(2),FE,FE)),UNIT=2032 CNTLUNIT CUNUMBR=D000, * PATH=((CSS(0),50,54,58,5C),(CSS(1),50,54,58,5C),(CSS(2),* 50,54,58,5C,7E,82),(CSS(3),50,54,58,5C)), * UNITADD=((00,256)), * LINK=((CSS(0),2C,2C,0A,0A),(CSS(1),2C,2C,0A,0A),(CSS(2),* 2C,2C,0A,0A,28,3F),(CSS(3),2C,2C,0A,0A)),CUADD=0, * UNIT=2107 CNTLUNIT CUNUMBR=D100, * PATH=((CSS(0),51,55,59,5D),(CSS(1),51,55,59,5D),(CSS(2),* 51,55,59,5D,7E,82),(CSS(3),51,55,59,5D)), * UNITADD=((00,256)), * LINK=((CSS(0),28,28,0E,0E),(CSS(1),28,28,0E,0E),(CSS(2),* 28,28,0E,0E,28,3F),(CSS(3),28,28,0E,0E)),CUADD=1, * UNIT=2107 There are three control units defined by the CUNUMBER keyword: 0065, D000, and D100. The link address of control unit 0065 is specified as FE in the LINK keyword, which is the port number of CUP in the FICON Director. The destination ports assigned for control units D000 and D100 are ports 28 and 3F in the FICON Director.

120

FICON Planning and Implementation Guide

The logical CU image number 0 is specified for control unit D000. The logical CU image number 1 is specified for control unit D100 in the CUADD keyword. CHIPDs 7E and 82 have access to all CUs because both CHPIDs are defined in the PATH keyword on all three CUs. The CHPIDs and the CUs are now defined. The next step is to define the devices owned by the CU. See Example 6-3 for the IODEVICE statement and its keywords.
Example 6-3 IODEVICE definition for switched point-to-point configuration

IODEVICE ADDRESS=065,UNITADD=00,CUNUMBR=(0065),STADET=Y, * UNIT=2032 IODEVICE ADDRESS=(D000,113),CUNUMBR=(D000),STADET=Y,UNIT=3390B IODEVICE ADDRESS=(D071,143),CUNUMBR=(D000),STADET=Y,UNIT=3390A IODEVICE ADDRESS=(D100,113),CUNUMBR=(D100),STADET=Y,UNIT=3390B IODEVICE ADDRESS=(D171,143),CUNUMBR=(D100),STADET=Y,UNIT=3390A The IODEVICE statement specifies the characteristics of devices 0065, D000, and D100. Device 0065 belongs to control unit 0065, which is the CUP port in FICON Director. On control unit D000 there are 113 devices are defined, starting at device address D000. The same definitions are done for devices associated to control unit D100. These are base devices in the storage unit. In addition 143 alias Hyper PAV devices are defined starting at device address D071 and D171.

Defining FICON Channel-to-Channel (FCTC)

In addition to storage control units, storage devices and the CUPs in the FICON Directors, we have defined FICON Channel-to-Channel (FCTC) CUs and devices to allow data exchange between LPARs. For considerations regarding FCTCs, refer to Appendix G, Adding FICON CTC connections on page 305. Based on those considerations, we are going to configure FCTCs in a switched point-to-point environment. Figure 6-3 on page 122 illustrates our configuration.

Chapter 6. Configuring a switched point-to-point topology

121

z10 EC
SCZP201
LPAR SC30 A23 z/OS V1R10
FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E

SAN768B
Port 03

CU# CU# CU# CU# 4044 5044 4034 5034

LPAR SC31 A24 z/OS V1R10

FICON Express8 LX CSS2 PCHID 5D2 CHPID 82

Switch # 65 Switch @ 65
Port 1B

Figure 6-3 FICON CTC (FCTC) configuration

In this configuration there are two LPARs (A23 and A24) residing in the same z10 server. Both LPARs can access CHPIDs 7E and 82 due to shared definition of those channel paths. Channels are attached to FICON Director 65 on ports 03 and 1B. For FCTC communication, there four logical control units defined: two of them for each partition for direct access in the same server, and two of them for each partition to access over the FICON link. Control units 4044 and 5044 allow direct access to LPAR A23. Control units 4034 and 5034 allow access to LPAR A24. Example 6-4 shows the definition of FCTC CUs and devices.
Example 6-4 IOCDS definitions for FICON CTC

CNTLUNIT CUNUMBR=5034,PATH=((CSS(2),7E)),UNITADD=((04,004)), LINK=((CSS(2),1B)),CUADD=23,UNIT=FCTC IODEVICE ADDRESS=(5034,004),UNITADD=04,CUNUMBR=(5034), STADET=Y,PARTITION=((CSS(2),A24)),UNIT=FCTC CNTLUNIT CUNUMBR=5044,PATH=((CSS(2),7E)),UNITADD=((04,004)), LINK=((CSS(2),1B)),CUADD=24,UNIT=FCTC IODEVICE ADDRESS=(5044,004),UNITADD=04,CUNUMBR=(5044), STADET=Y,PARTITION=((CSS(2),A23)),UNIT=FCTC CNTLUNIT CUNUMBR=4034,PATH=((CSS(2),82)),UNITADD=((04,004)), LINK=((CSS(2),03)),CUADD=23,UNIT=FCTC IODEVICE ADDRESS=(4034,004),UNITADD=04,CUNUMBR=(4034), STADET=Y,PARTITION=((CSS(2),A24)),UNIT=FCTC CNTLUNIT CUNUMBR=4044,PATH=((CSS(2),82)),UNITADD=((04,004)), LINK=((CSS(2),03)),CUADD=24,UNIT=FCTC IODEVICE ADDRESS=(4044,004),UNITADD=04,CUNUMBR=(4044), STADET=Y,PARTITION=((CSS(2),A23)),UNIT=FCTC

* * * * * * * *

Four control units are defined by the CNTUNIT statement: 4044 and 5044 for direct access to LPAR A23, and 4034 and 5034 for LPAR A24. Because CHPIDs 7E and 82 are shared by logical partitions, both CHPIDs can be used by both partitions for FCTC communication.

122

FICON Planning and Implementation Guide

The PARTITON keyword specifies that devices in control unit 5034 and 4034 allow direct access only by LPAR A24. These devices can be accessed by LPAR A23 over the FICON link by CHPIDs 7E and 82. The same rule applies to LPAR A24. This allows data traffic to flow between both LPARs using the corresponding devices, as shown here: SC30 4044 4045 4046 4047 <----> <----> <----> <----> SC31 5034 5035 5036 5037 SC30 5044 5045 5046 5047 <----> <----> <----> <----> SC31 4034 4035 4036 4037

The rules we followed for device numbering are described in FCTC device numbering scheme on page 306. Figure 6-4 illustrates the logical view of the FCTC configuration and the data path between two corresponding FCTC devices. To simplify the illustration, only two logical CUs are shown.

LPAR A23
Data transfer

CU 4044

7E

03

@65

1B

82

CU 5034

Data transfer

LPAR A24
Figure 6-4 Corresponding FCTC devices

Data is transferred over the FICON link in both directions between the CUs and the logical partitions (LPARs). LPAR A23 sends data to A24 via CU 5034 and receives data from A24 via CU 4044. The reverse applies to LPAR A24 when data is sent to or received from A23.

Activating the changes (IOCDS) and configuring online

After the definition of the desired configuration is finished, the data must be saved into the IODF data set and the IOCDS file on the Support Element. Up to four IOCDS files can be stored in a System z server. This allows you to plan and define future I/O configurations and store them on the Support Element of the server. The tasks that are required to save definition data and activate the changes dynamically can be done via HCD. Refer to HCD Users Guide, SC33-7988, for detailed descriptions of all activation procedures. On the HCD start menu, select option 2 (Activate or process configuration data). Figure 6-5 on page 124 shows the Activate or Process Configuration Data menu.

Chapter 6. Configuring a switched point-to-point topology

123

Figure 6-5 HCD Activate or Process Configuration Data menu

The options on the menu to activate the changes are: Option 1. Build production I/O definition file Option 2. Build IOCDS Option 3. Activate or verify configuration dynamically The Build production I/O definition file (IODF) function (option 1) saves definition data into an IODFxx dataset, where xx in the data set name is a suffix to identify the IODF among other IODFs, which may already exist. The suffix of the IODF data set is used in the IPL parameter when the z/OS image is brought up. This ensures that the desired IODF used by the operating system image matches the IOCDS that is loaded into the HSA. Using Build IOCDS (option 2) allows you to build an IOCDS and save it to the CPC. The IOCDS is built using the definitions saved in the IODFxx data set. Using Activate or verify configuration dynamically (option 6) allows you to dynamically activate changes in the I/O definitions. There is no Power-on Reset (POR) required because changes are directly updated to the HSA in the CPC. This is the preferred method for activating changes. Another way to activate changes in I/O definitions is to perform a POR at the CPC with the newly built IOCDS. Keep in mind, however, that a POR causes a system outage. For this reason, if a POR is required, then we suggest that you plan ahead to make any changes in I/O definitions so that the new IOCDS is ready for use at the next POR. Changes can be made at any time and saved in an IODF data set. The newly built IOCDS based on the I/O definition in the IODF should be sent to the CPCs Support Element and marked for use at the next POR. This ensures that the newly built IOCDS will be used the next time the CPC is activated.

124

FICON Planning and Implementation Guide

To verify which I/O definitions (IODF) are currently active, enter D IOS,CONFIG at the z/OS operator console. The response to the command is shown in Example 6-5.
Example 6-5 D IOS,CONFIG command

IOS506I 09.08.24 I/O CONFIG DATA 030 ACTIVE IODF DATA SET = SC31.IODF73 CONFIGURATION ID = TEST2097 EDT ID = 01 TOKEN: PROCESSOR DATE TIME DESCRIPTION SOURCE: SCZP201 09-04-21 16:12:24 SC31 IODF73 ACTIVE CSS: 2 SUBCHANNEL SETS CONFIGURED: 0, 1 CHANNEL MEASUREMENT BLOCK FACILITY IS ACTIVE Example 6-5 shows that the IODF73 data set is currently active in z/OS image SC31. It also provides information about the CSSs used in this z/OS image. CSS2 is the active CSS in system SC31. Compare the information provided by the D IOS,CONFIG command with the data of the ID statement in the ICODS, shown in Example 6-6.
Example 6-6 ID statement in IOCDS

ID

MSG1='IODF73',MSG2='SC31.IODF73 - 2009-04-21 16:12', * SYSTEM=(2097,1),LSYSTEM=SCZP201, * TOK=('SCZP201',00800006991E2094161224920109111F00000000,* 00000000,'09-04-21','16:12:24','SC31','IODF73')

The suffix of the IODF data set shown in the IOCDS ID statement and displayed by the D IOS command should be the same number (that is, IODF73). This proves that the IOCDS was built from the current active IODF data set and that the latest changes and new IO definitions were successfully activated.

Configuring the IBM Storage System DS8000

The host adapter ports in the DS8000 storage control unit have to be configured for native FICON protocol. Follow the procedure described in Appendix D, Configuring the DS8000 for FICON on page 259 to configure the host adapter ports. The configuration steps are based on the configuration shown in Figure 6-1 on page 116. For planning and implementation information, refer to IBM System Storage DS8000: Architecture and Implementation, SG24-6786.

Configuring the FICON Director

The FICON Director must be configured to match the designated configuration (shown in Figure 6-1 on page 116). For the desired switched point-to-point topology, this includes: Specifying the Domain ID in the Director Enabling the ports Checking the port settings Other settings in the FICON Director are optional and should be specified after reading the considerations explained in Chapter 4, Planning the FICON environment on page 59. Make sure that the selected ports (LX or SX) in the Director match the FICON node (LX or SX) that will be attached.

Chapter 6. Configuring a switched point-to-point topology

125

All the required steps to configure the FICON Director are described in detail in Chapter 8, Configuring FICON Directors on page 151. Go to 8.1.1 Configuration flowchart on page 152 and follow the procedures described there. Return to this section after the FICON Director is configured and ready for use.

Connecting the fiber optic cables

After all definition and configuration tasks are completed, the fiber optic cables have to be plugged into the FICON channels, CU host adapters, and the FICON Director. All the ports in our configuration are long wavelength (LX) laser types.For LX type lasers, a 9m single mode (SM) fiber optic cable is recommended to achieve maximum performance and distance. All fiber optic cables must have an LC duplex connector at both ends. Plug the fiber optic cable connectors to the designated ports, shown in Figure 6-1 on page 116, to the FICON channel, the CU, and the FICON Director, as listed here: PCHID 5A2 to FICON Director port 03 PCHID 5D2 to FICON Director port 1B CU host adapter port 0003 to FICON Director port 28 CU host adapter port 0242 to FICON Director port 3F Make sure the fiber optic cable ends are cleaned before plugging them into the transceiver. Ensure all fiber optic cables are labeled and documented for later use (for example, when recabling or for problem determination). For further information and considerations regarding fiber optic cabling and documentation, refer to 4.11 Physical connectivity on page 93 and 4.2 Documentation on page 62.

Configuring the CHPIDs online

After you successfully activate the new I/O definitions and the FICON Director and the storage control unit are configured, you can configure the channel path online and try to communicate with the devices. Note that even if all the definitions in the IOCDS are correct, the channel status might be offline, so enter D M=CHP(7E) on the system console to display the current status of CHPID 7E. If the CHPID is offline, enter CF CHP(7E),ON to configure the channel online. The channel status should change to ONLINE, as shown in Example 6-7).
Example 6-7 CHPID 7E online state

CF CHP(7E) IEE502I CHP(7E),ONLINE IEE712I CONFIG PROCESSING COMPLETE To achieve the best performance on the FICON channel, make sure that zHPF is enabled. Refer to 4.9.6 High Performance FICON on page 88 for considerations regarding how to exploit zHPF on System z servers. Enter D IOS,ZHPF at the z/OS console to display zHPF settings for the system. If zHPF is disabled, enter SETIOS ZHPF,YES. This enables zHPF, as shown in Example 6-8.
Example 6-8 Enable zHPF

D IOS,ZHPF RESPONSE=SC30 IOS630I 13.27.23 ZHPF FACILITY 021

126

FICON Planning and Implementation Guide

HIGH PERFORMANCE FICON FACILITY IS ENABLED The SETIOS ZHPF=YES command enables zHPF temporarily. However, after the next system IPL, the zHPF facility will be reset to the default (disabled). To permanently enable zHPF for z/OS, change the FCX parameter in SYS1.PARMLIB member IECIOSxx to FCX=YES. Now you can query the status and functional details of the channel path by entering D M=CHP(7E) at the operator console. The command output is shown in Example 6-9. It provides information about the channel and the attached devices.
Example 6-9 D M=CHP(7E)

D M=CHP(7E) IEE174I 13.22.46 DISPLAY M 144 CHPID 7E: TYPE=1B, DESC=FICON SWITCHED, ONLINE DEVICE STATUS FOR CHANNEL PATH 7E 0 1 2 3 4 5 6 7 8 9 A B C D E 0006 . . . . . + . . . . . . . . . 0D00 + + + + + + + + + + + + + + + 0D01 + + + + + + + + + + + + + + + 0D02 + + + + + + + + + + + + + + + 0DFD HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFE HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFF HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA SWITCH DEVICE NUMBER = 0065 DEFINED ENTRY SWITCH - LOGICAL SWITCH ID = 65 ATTACHED ND = SLKWRM.DCX.BRD.CA.0AFX0642C00Z PHYSICAL CHANNEL ID = 05A2 FACILITIES SUPPORTED = ZHPF

F . + + + HA HA HA

Example 6-9 shows that CHPID 7E is online and working in a switched point-to-point configuration. Information is also displayed about the attached node and the supported facilities (for example, zHPF). To check that communication to the attached devices is working properly, enter D M=DEV(xxxx), where xxxx is any device number. In Example 6-10 we checked the status of device 0065, which is the CUP in FICON Director, and device D000 which is a storage device.
Example 6-10 D M=DEV(0065,D000)

IEE174I 14.11.57 DISPLAY M 161 DEVICE 0065 STATUS=ONLINE CHP 7E 82 ENTRY LINK ADDRESS 03 1B DEST LINK ADDRESS FE FE PATH ONLINE Y Y CHP PHYSICALLY ONLINE Y Y PATH OPERATIONAL Y Y MANAGED N N CU NUMBER 0065 0065 MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = SLKWRM.DCX.BRD.CA.1AFX0642C00Z.002F SCP TOKEN NED = SLKWRM.DCX.BRD.CA.1AFX0642C00Z.0000 SCP DEVICE NED = SLKWRM.DCX.BRD.CA.1AFX0642C00Z.0000

Chapter 6. Configuring a switched point-to-point topology

127

DEVICE D000 STATUS=ONLINE CHP 50 54 58 5C 7E 82 ENTRY LINK ADDRESS 0F 0F 1B 1B 03 1B DEST LINK ADDRESS 2C 2C 0A 0A 28 3F PATH ONLINE N N Y Y Y Y CHP PHYSICALLY ONLINE N N Y Y Y Y PATH OPERATIONAL N N Y Y Y Y MANAGED N N N N N N CU NUMBER D000 D000 D000 D000 D000 D000 MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107.900.IBM.75.0000000BALB1.0300 SCP TOKEN NED = 002107.900.IBM.75.0000000BALB1.0000 SCP DEVICE NED = 002107.900.IBM.75.0000000BALB1.0000 HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW, ZHPF The response to D M=DEV(0065,D000) displays all available paths to the devices and their status. Information about the device (for example, device type), the control unit number, the link address, and the functions supported by the device (MIDAW and zHPF) is shown.

Verifying the installation

After all the configuration tasks in the z10 server, the DS8000 storage, and the FICON Director are done, you can verify that the current configuration matches the desired configuration shown in Figure 6-1 on page 116. This verification should be done to ensure that the fiber optic cables are correctly plugged to the designated ports in the z10 server, the DS8000, and the FICON Director, and to check the status of channels and devices. First, check the status of PCHID 5A2 in the z10 server. From any HMC where the CPC (server) is defined (for example, SCZP201), start a Single Object Operation (SOO) to that CPC. Display the list of channels installed on the CPC. Search the list for the desired PCHID (5A2) and double-click the PCHID icon. This will display the PCHID details panel shown in Figure 6-6.

Figure 6-6 PCHID details

128

FICON Planning and Implementation Guide

Essential information for PCHID 5A2 is shown on the PCHID Details panel: PCHID status is Operating. PCHID type is FICON Express8. CSS.CHPID 2.7E (and 0.7E) is assigned to PCHID 5A2. The owning image of PCHID 5A2 is A23 and A24 (among other images). CHPIDs assigned to PCHID 5A2 are shared across images. Similar information can be found on the CHPID Details panel, shown in Figure 6-7.

Figure 6-7 CHPID details

The CHPID Details panel provides similar information to the PCHID details panel. To display the CHPID details, select the channel list for an operAting system image from the Support Element workplace. As you can see, information for PCHID 5A2 (CHPID 2.7E) provided on the details window matches the designated configuration. This proves that PCHID 5A2 has CHPID 2.7E assigned and images A23 and A24 can access the channel. Repeat this checking on other channels that were recently configured. Note: If any of the data displayed on the PCHID or CHPID detail panel does not match to the desired configuration, you must correct the definitions in the IOCDS. Next, check that the FICON channels are connected to the correct FICON switch and port. On either the PCHID details panel or the CHPID details panel, click the Channel Problem Determination button. This will display the Channel Problem Determination panel, where you can select which information you want to display; see Figure 6-8 on page 130.

Chapter 6. Configuring a switched point-to-point topology

129

Figure 6-8 Channel Problem Determination panel

Select Analyze channel information and click OK. The Analyze Channel Information window is displayed, as shown in Figure 6-9, which provides information about the node attached to the FICON channel.

Figure 6-9 Analyze Channel Information window

130

FICON Planning and Implementation Guide

Information about the nodes are shown at the bottom part of the window. The lower left side displays information about the node in the System z server. The lower right side displays information about the attached node. Important: Make sure that the Node status for both nodes is displayed as Valid. If any other status is shown, none of the displayed information is valid. Check that the Type/model information and the serial number (Seq. number) are as expected. Next, examine the Tag field for each node. The Tag provides information about the port number of the node. The two right-most digits of the Tag value represent the CHPID number for the channel node (7E). For the attached node, the Tag value represents the port number (03). The World Wide Node Name (WWN) and the World Wide Port Name (WWP) are also shown for each port, and may be used to prove that the channel is connected to the correct FICON Director if the WWN or WWP is known. If the node status is not Valid, or the Tag value and the WWNN or WWPN are not correct, then check the fiber optic cable link between the z10 server and the FICON Director to ensure that it is plugged correctly. Next, check the Channel link address field, which displays the switch address and the port number of the attached Director. The two left-most digits represent the switch address (65). The two digits in the middle represent the port number (03). In our scenario, we are now sure that PCHID 5A2 has connectivity to port number 03 on switch address 65, which matches our desired configuration (see Figure 6-1 on page 116). If the displayed values are not as expected, the fiber optic cables may not be plugged correctly and must be checked. After completing the preceding steps and proving that the physical path to the FICON Director and the logical definitions of the link are correct, check that the path to the control unit image is initialized correctly and properly defined. On the Channel Problem Determination panel (shown in Figure 6-8 on page 130), select Analyze Serial Link Status and click OK. The Analyze Serial Link Status window is displayed, as shown in Figure 6-10 on page 132.

Chapter 6. Configuring a switched point-to-point topology

131

Figure 6-10 Analyze Serial Link Status

The Analyze Serial Link Status window provides status information about the link to the control unit images defined in the IOCDS. Figure 6-10 shows a link status of Initialization Complete for all defined CU images on link address FE and 28. Information is also displayed showing that PCHID 05A2 (CHPID 2.7E) is connected to port 03 on switch address 65 (switch number and channel link address). Link address FE is the CUP port in the FICON Director, which is used for communication and managing the FICON Director. Link address 28 is the destination port where the CU is physically attached. The link to CU address 00 and 01 (among other CU images) is initialized. CU address 00 and 01 correspond to control unit number D000 (CUADD=0) and D1000 (CUADD=1), as defined in the IOCDS. If the link status Initialization Complete is not shown, you must check that the ports in the FICON Director and the CU are correctly configured, and that the fiber optic cable link between the Director and the CU has the correct cable type and plugging.

132

FICON Planning and Implementation Guide

Chapter 7.

Configuring a cascaded FICON Director topology

This chapter describes all the tasks that are required to configure and define a FICON environment for a cascaded FICON Director topology. Establishing a cascaded FICON Director topology involves the following: Defining FICON channel, control unit, and device definitions in IOCP Activating the IOCDS Configuring the storage control unit Varying the channel path and device online Verifying the configuration

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

133

7.1 Establishing a cascaded FICON Director topology

This section describes the tasks required to configure a cascaded FICON Director topology on a System z server, including setting up FICON Directors and configuring the attached storage controller. Our discussion is based on an infrastructure that is already built and installed, with the following items physically installed: System z server DS8000 storage system FICON Directors Fiber cabling infrastructure

7.1.1 Description of our environment

Based on the considerations explained in Chapter 4, Planning the FICON environment on page 59, a cascaded FICON configuration will be built, consisting of a z10 EC server, a SAN768B and SAN384B FICON Director, and a DS8000 storage device. Longwave laser optic and SM fiber optic cables are used to achieve maximum distance and performance. In our example case, the entire infrastructure is already installed in our environment. Our configuration is shown Figure 7-1.

z10 EC
SCZP201
LPAR SC30 A23 z/OS V1R10
FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E

System Storage SAN384B

Switch # 66 Switch @ 66
Port 8D Port C4 Port 88 Port CF

DS8000
CU# D000
LX 0003

D0xx

LPAR SC31 A24 z/OS V1R10

FICON Express8 LX CSS2 PCHID 5D2 CHPID 82

Port 03

Port 2D

Port 34

Switch # 65 Switch @ 65

LX 0242

D1xx

CU# D100
Port 1B * All cable connectors are LC Duplex type

SAN768B
Figure 7-1 Cascaded FICON configuration

The z10 EC server (SCZP201) has two LPARS defined and activated. The system names running in this partition are SC30 and SC31, which correspond to the LPAR names A23 and A24. The operating system that is running in both partitions is z/OS V1R10. As a standard feature, the z10 server supports zHPF protocol, which will be used for data transfer to the DS8000 storage device. Two FICON Express8 channels are defined as shared and assigned to CSS2. Both LPARs have access to the two FICON Express8 channels. The channel adapters are FICON 134
FICON Planning and Implementation Guide

Express8 LX features (FC 3325), using a longwave laser for data traffic through the fiber optic cable. PCHID 5A2 has CHPID 7E assigned. PCHID 5D2 has CHPID 82 assigned. The SAN768B FICON Director has connectivity to the z10 server on port 03 and port 1B. All the ports in the FICON Director are longwave (LX) ports. The switch number (Switch #) and the switch address (Switch @) are both set to 65. The switch number is specified in the IOCDS. The switch address is the Domain ID specified in the Director. The SAN384B FICON Director has connectivity to storage devices on port 88 and CF. All the ports in the FICON Director are longwave (LX) ports. The switch number (Switch #) and the switch address (Switch @) are both set to 66. The switch number is specified in the IOCDS. The switch address is the Domain ID specified in the Director. The FICON Directors are attached by an Inter-Switch Link (ISL), which consists of two fiber optic cables connected to ports 2D and 34 in switch 65 and ports 8D and 34 in switch 66. If the Directors are connected they build a High Integrity Fabric, which is required to support cascaded FICON Directors. The ISLs are transparent in the path from the System z server to the CU and do not require any definitions in IOCDS. An IBM System Storage DS8000 will be connected to the SAN384B FICON Director on port 88 and the CF. The two host adapters are installed in the DS8000 supporting longwave (LX) laser. Port numbers 0003 and 0242 in the DS8000 are used to connect to the FICON Director. Two logical control units (D000 and D100) are defined which have devices D000-D0FF or D1000-D1FF assigned. The DS8000 will have the zHPF feature enabled to communicate to the z10 server. A maximum unrepeated distance of 10km (6.2 miles) is supported by the longwave laser (LX) feature when using 9m single mode (SM) fiber optic cables. The fiber optic cables have an LC Duplex connector at both ends to connect to the z10 FICON Express8 channels, as well as to the ports in the FICON Directors and to the DS8000 host adapters.

7.1.2 Tasks and checklist

This section provides an overview of required tasks, as well as a checklist that we used to make sure that all hardware and software prerequisites were met before starting configuration tasks.

Tasks
Figure 7-2 on page 136 shows the main steps required to define and activate a cascaded FICON Director configuration.

Chapter 7. Configuring a cascaded FICON Director topology

135

Verification checklist

Follow the verification checklist to ensure that all hardware and software prerequisites are met. Go to Verification checklist on page 136. Information about defining the channel paths, control units, and devices is given in Defining channel, CU, and storage device in IOCDS on page 138.

Define channel, CU, and devices on a System z server

Configure storage CU

The configuration tasks for a DS8000 storage system are described in Configuring the IBM Storage System DS8000 on page 143. All required steps to set up the FICON Directors are shown in 8.1.1, Configuration flowchart on page 152.

Configure FICON Director

Plug fiber cables

Information about fiber optic cables and plugging rules to achieve the desired configuration is given in Connecting the fiber optic cables on page 144. Information about how to verify that your actual configuration matches the desired configuration is given in Verifying the installation on page 146.

Verify configuration

Figure 7-2 Main steps for configuring and verifying a cascaded FICON Director configuration

Verification checklist
Before configuring the cascaded FICON topology shown in Figure 7-1 on page 134, the following list was checked. All steps in the checklist must be finished and corrected (if required) to ensure a smooth and successful configuration of the topology. Both hardware and software requirements must be checked. Check that FICON features are available on the System z server to establish the desired configuration. For details about each feature code, see System z FICON feature support on page 25. FICON Express2 LX FC 3319 SX FC 3320 FICON Express4 LX FC 3321, FC 3323, FC 3324 SX FC 3318, FC 3322 FICON Express8 LX FC 3325 SX FC 3326 If using the FICON Express8 feature, check the System z operating system requirements. z/OS V1.7 with the IBM Lifecycle Extension for z/OS V1.7 (5637-A01) is required to support the FICON Express8 feature in a z10 server.

136

FICON Planning and Implementation Guide

Check the 2097DEVICE PSP bucket for the latest information about FICON Express support on the operating system. Check that the number and types of FICON Director ports match the configuration requirements. LX SFPs SX SFPs Check that high-integrity fabric support is enabled on the FICON Directors. Check the DS8000 storage hardware requirements to support FICON longwave (LX) or shortwave (SX) connectivity. Check that FC 0709 and FC 7092 are installed to support zHPF. Note: FC 0709 and FC 7092 must be ordered to obtain the license key to enable zHPF support in a DS8000 storage controller. Check that the DS8000 firmware is at level 6.1.16.0 or higher to support the zHPF feature. Check that the zHPF feature is enabled in the DS8000. Check that the correct fiber optic cables are available to connect the FICON Director to the z10 server and the DS8000 storage to the FICON Director. A 9m single mode (SM) fiber optic cable is required to support longwave laser (LX) for a maximum distance of 10km (6.2 miles) at maximum speed. A 50m or 62.5m multi mode (MM) fiber optic cable is required to support shortwave laser (SX). See System z FICON feature support on page 25 for the maximum supported distance, depending on cable type and speed. An LC duplex connector is required at both ends of the fiber optic cables to connect to the System z server FICON adapter, as well as the ports in the FICON Director and the DS8000 host adapter. Note: All fiber optic cables used in a link must be of the same type. For example, they must be either all single mode or all multi-mode fiber optic cables.

7.1.3 Getting started

This section describes all the tasks that are required to achieve the designated configuration shown in Figure 7-1 on page 134. It explains what needs to be done to get the FICON channels, FICON Director, storage control units, and the storage device online and operating. The steps include: Defining channel, CU, and storage device in IOCDS Activating changes (IOCDS) and configuring online Configuring the IBM Storage System DS8000 Configuring the FICON Directors Connecting the fiber optic cables Configuring the CHPIDs online Verifying the installation For our scenario, we had an active partition in a System z server running z/OS with HCD. We used HCD to create, save, and activate our I/O definitions.

Chapter 7. Configuring a cascaded FICON Director topology

137

Defining channel, CU, and storage device in IOCDS

There are several tools available that you can use to define and configure a FICON environment on a System z server. The definition tools and an explanation of the statements and keywords used in the IOCP are described in Appendix C, Configuration and definition tools on page 255. Only statements and keywords relevant to our desired configuration are explained. Refer to IOCP Users Guide, SB10-7037, for details about all IOCP statements and keywords. HCD was used to define the CHPIDs, CUs, and devices for the sample configuration we describe in this section. Refer to HCD Users Guide, SC33-7988, for detailed information regarding the use of HCD. After all definition tasks are completed, the definition data is stored in the I/O definition file (IODF) data set. Any time you make changes to the I/O configuration, the IOCDS has to be updated to ensure consistency between the software and the microcode in the hardware. The newly created IOCDS will be sent to the Support Element of the desired System z server. The Support Element in the server stores the IOCDS on its built-in HDD. The IOCDS stored in the Support Element is loaded into System z storage during a Power-on Reset (POR) of the CPC. Changes or additions of channel paths, CUs, or devices can be performed dynamically. Because no system outage is required for dynamic changes, this is the preferred method for changing I/O definitions. If changes to the I/O definitions are performed dynamically, the changes are directly updated into the Hardware System Area (HSA) of the server. The HSA is a separate storage area in the System z server that holds all I/O specifications and provides status information for all defined I/Os. The operating system uses the HSA to manage the I/O subsystem. The following examples show the CHPID, CNTLUNIT, and IODEVICE statements if all the definition tasks based on the configuration shown in Figure 7-1 on page 134 are finished. First, we explain how the CHPIDs are defined. Example 7-1 displays the CHPID statement and its keywords.
Example 7-1 CHPID definition for cascaded FICON configuration

CHPID PATH=(CSS(0,2),7E),SHARED, * PARTITION=((CSS(0),(A01),(=)),(CSS(2),(A23,A24,A25,A29),* (=))),SWITCH=65,PCHID=5A2,TYPE=FC CHPID PATH=(CSS(1,2),82),SHARED, * PARTITION=((CSS(1),(A11),(=)),(CSS(2),(A23,A24,A25,A29),* (=))),SWITCH=65,PCHID=5D2,TYPE=FC The PATH keyword in the CHPID statement defines CHPID 7E in CSS 2 as SHARED. The LPAR names A23 and A24 are specified in the PARTITION keyword to use CSS 2 to access CHPID 7E. The LPAR names specified in ICODS correspond to system names SC30 and SC31. The SWITCH keyword is used to provide the switch number (65) of the Director to the channel to which it is connected. PCHID 5A2 has been merged to the IOCDS by the CMT, which assigned CHPID 7E to PCHID 5A2. The TYPE of the channel is specified as native FICON (FC).

138

FICON Planning and Implementation Guide

The same definition rules apply to the definition of CHPID 82. With the CHPIDs defined, we illustrate next how the CUs that are attached to the CHPIDs are defined. Example 7-2 displays the CNTLUNIT statement and keywords.
Example 7-2 CNTLUNIT definition for a cascaded FICON configuration

CNTLUNIT CUNUMBR=0065,PATH=((CSS(2),7E,82)), * UNITADD=((00,001)),LINK=((CSS(2),65FE,65FE)),UNIT=2032 CNTLUNIT CUNUMBR=0066,PATH=((CSS(2),7E,82)), * UNITADD=((00,001)),LINK=((CSS(2),66FE,66FE)),UNIT=2032 CNTLUNIT CUNUMBR=D000, * PATH=((CSS(0),50,54,58,5C),(CSS(1),50,54,58,5C),(CSS(2),* 50,54,58,5C,7E,82),(CSS(3),50,54,58,5C)), * UNITADD=((00,256)), * LINK=((CSS(0),2C,2C,0A,0A),(CSS(1),2C,2C,0A,0A),(CSS(2),* 2C,2C,0A,0A,6688,66CF),(CSS(3),2C,2C,0A,0A)),CUADD=0, * UNIT=2107 CNTLUNIT CUNUMBR=D100, * PATH=((CSS(0),51,55,59,5D),(CSS(1),51,55,59,5D),(CSS(2),* 51,55,59,5D,7E,82),(CSS(3),51,55,59,5D)), * UNITADD=((00,256)), * LINK=((CSS(0),28,28,0E,0E),(CSS(1),28,28,0E,0E),(CSS(2),* 28,28,0E,0E,6688,66CF),(CSS(3),28,28,0E,0E)),CUADD=1, * UNIT=2107 There are four control units defined by the CUNUMBR keyword: 0065, 0066, D000, and D100. The LINK keyword defines the destination port address the control unit to which it is attached. The first byte of the link address represents the switch address and the second byte defines the port address in that switch. Control units 0065 and 0066 are the CUP ports (port number FE) in the FICON Directors defined on the LINK statement (in our example, 65FE and 66FE). The destination ports specified for control units D000 and D100 are 6688 and 66CF. The logical CU image number 0 is specified for control unit D000. The logical CU image number 1 for control unit D100 is specified in the CUADD keyword. CHPIDs 7E and 82 in CSS2 have access to all control units, as specified in the PATH keyword for all four CUs. The CHPIDs and the CUs are now defined. The next step is to define the devices owned by the CU. Example 7-3 displays the IODEVICE statement and keywords.
Example 7-3 IODEVICE statement for cascaded FICON configuration

IODEVICE ADDRESS=065,UNITADD=00,CUNUMBR=(0065),STADET=Y, * UNIT=2032 IODEVICE ADDRESS=066,UNITADD=00,CUNUMBR=(0066),STADET=Y, * UNIT=2032 IODEVICE ADDRESS=(D000,113),CUNUMBR=(D000),STADET=Y,UNIT=3390B IODEVICE ADDRESS=(D071,143),CUNUMBR=(D000),STADET=Y,UNIT=3390A IODEVICE ADDRESS=(D100,113),CUNUMBR=(D100),STADET=Y,UNIT=3390B IODEVICE ADDRESS=(D171,143),CUNUMBR=(D100),STADET=Y,UNIT=3390A

Chapter 7. Configuring a cascaded FICON Director topology

139

The IODEVICE statement specifies the characteristics of devices 0065, D000, and D100. Device 0065 belongs to control unit 0065 which is the CUP port in FICON Director. On control unit D000 there are 113 devices defined, starting at device address D000. The same definitions are done for devices associated to control unit D100. These are base devices in the storage unit. In addition, 143 alias devices are defined starting at device address D071 and D171.

Defining a FICON Channel-to-Channel (FCTC) connection

In addition to storage control units, storage devices, and the CUPs in the FICON Directors, we have defined FICON Channel-to-Channel (FCTC) CUs and devices to allow data exchange between LPARs. For considerations regarding FCTCs, refer to Appendix G, Adding FICON CTC connections on page 305. Based on those considerations, we are going to configure FCTCs in a cascaded FICON environment. Figure 7-3 illustrates the desired FCTC configuration. To simplify the illustration, two control units and two LPARs are shown in this sample confirmation, although more than two control units could be defined.

z10 EC
SCZP201
LPAR SC30 (A23) z/OS V1R10
FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E

System Storage SAN384B

Switch @ 66 Switch @ 66
Port 8D Port C4 Port CF

z10 EC
SCZP202

ISL
Port 03 Port 2D Port 34
FICON Express8 LX PCHID 5D2 CHPID 82

CU# 5034
LPAR SC31 (A24) z/OS V1R10

CU# 4034 4044 CU# 4044

Switch # 65 Switch @# 65

CSS2

SAN768B
Figure 7-3 FCTC configuration

Both z10 servers are attached to cascaded FICON Directors: CHPID 7E to port 03 in switch 65, and CHPID 82 to port CF in switch 66. Each LPAR will have access to a logical control unit and their logical devices defined in the corresponding LPAR in different z10 servers. LPAR A23 will communicate with CU 5034, which is accessed by LPAR A24. LPAR A24 will communicate with CU 4044, which is accessed by LPAR A23. Example 7-4 shows the definition of FCTC control units and devices.
Example 7-4 FCTC configuration

CNTLUNIT CUNUMBR=5034,PATH=((CSS(2),7E)),UNITADD=((04,004)), LINK=((CSS(2),66CF)),CUADD=23,UNIT=FCTC IODEVICE ADDRESS=(5034,004),UNITADD=04,CUNUMBR=(5034), STADET=Y,PARTITION=((CSS(2),A24)),UNIT=FCTC 140

FICON Planning and Implementation Guide

* *

CNTLUNIT CUNUMBR=4044,PATH=((CSS(2),82)),UNITADD=((04,004)), LINK=((CSS(2),6503)),CUADD=24,UNIT=FCTC IODEVICE ADDRESS=(4044,004),UNITADD=04,CUNUMBR=(4044), STADET=Y,PARTITION=((CSS(2),A23)),UNIT=FCTC

* *

One control unit is defined for each LPAR: CU 5034 is defined for LPAR A23 and CU 4044 is defined for LPAR A24. CU 5034 resides in server SCZP202. CU 4044 resides in server SCZP201. The LINK keyword specifies the destination port to communicate to the control unit in the other server, which is port 66CF for LPAR 23 to communicate with CU 5034, and port 6503 for LPAR A24 with communicate to CU 4044. The PARTITION keyword in the IODEVICE statement specifies the logical partition that can access the device. This allows data traffic to flow between the LPARs using the corresponding devices, as follows: SC30 4044 4045 4046 4047 <---> <---> <---> <---> SC31 5034 5035 5036 5037

The rules we followed for device numbering are described in FCTC device numbering scheme on page 306. Figure 7-4 illustrates the logical view of the FCTC configuration and the data path between two LPARs. To simplify the illustration, only two logical CUs are shown.

LPAR A23
Data transfer

CU 4044

7E

03

@65

ISL

@66

CF

82

CU 5034

Data transfer

LPAR A24
Figure 7-4 FCTC data transfer

Data is transferred over the FICON link, in both directions, between the CUs and the logical partitions (LPARs). LPAR A23 sends data to A24 via CU 5034 and receives data from A24 via CU 4044. The reverse applies to LPAR A24 when data is sent to or received from A23.

Activating changes (IOCDS) and configuring online

When the definition of the desired configuration is finished, the data must be saved into the IODF data set and the IOCDS file on the Support Element. Up to four IOCDS files can be stored in a System z server. This allows you to plan and define future I/O configurations and store them on the Support Element of the server.

Chapter 7. Configuring a cascaded FICON Director topology

141

The tasks that are required to save definition data and activate the changes dynamically are performed via HCD. Refer to HCD Users Guide, SC33-7988, for detailed descriptions of all the activation procedures. On the HCD start menu, select option 2 (Activate or process configuration data). Figure 7-5 shows the Activate or process configuration data menu.

Figure 7-5 HCD Activate or process configuration data menu

The options on the menu to activate the changes are: Option 1. Build production I/O definition file Option 2. Build ICODS Option 6. Activate or verify configuration dynamically The Build production I/O definition file (IODF) function (option 1) saves definition data into an IODFxx dataset, where xx in the data set name is a suffix to identify the IODF among other IODFs which may already exist. The suffix of the IODF data set is used in the IPL parameter when the z/OS image is brought up. This ensures that the desired IODF used by the operating system image matches the IOCDS that is loaded into the HSA. Using Build IOCDS (option 2) allows you to build an IOCDS and save it to the CPC. The IOCDS is built using the definitions saved in the IODFxx data set. Using Activate or verify configuration dynamically (option 6) allows you to dynamically activate changes in the I/O definitions. There is no Power-on Reset (POR) is required because changes are directly updated to the HSA in the CPC. This is the preferred method of activating changes. Another way to activate changes in I/O definitions is to perform a POR at the CPC with the newly-built IOCDS.Keep in mind, however, that a POR causes a system outage. For this reason, if a POR is required, then we suggest that you plan ahead to make any changes in I/O definitions so that the new IOCDS is ready for use at the next POR. Changes can be

142

FICON Planning and Implementation Guide

made at any time and saved in an IODF data set. The newly built IOCDS based on the I/O definition in the IODF should be sent to the CPCs Support Element and marked for use at the next POR. This ensures that the newly built IOCDS will be used the next time the CPC is activated. To verify which I/O definitions (IODF) are currently active, enter the command D IOS,CONFIG at the z/OS operator console. The response to the command is shown in Example 7-5.
Example 7-5 D IOS,CONFIG command

IOS506I 13.10.28 I/O CONFIG DATA 003 ACTIVE IODF DATA SET = SC30.IODF78 CONFIGURATION ID = TEST2094 EDT ID = 01 TOKEN: PROCESSOR DATE TIME DESCRIPTION SOURCE: SCZP201 09-05-07 09:26:49 SC30 IODF78 ACTIVE CSS: 2 SUBCHANNEL SETS CONFIGURED: 0, 1 CHANNEL MEASUREMENT BLOCK FACILITY IS ACTIVE Example 7-4 on page 140 shows that IODF78 data set is currently active in z/OS image SC30. It also provides information about the CSSs used in this z/OS image. CSS2 is the active CSS in system SC30. Compare the information provided by the D IOS,CONFIG command with the data of the ID statement in the IOCDS, as shown in Example 7-6.
Example 7-6 ID statement in IOCDS

ID

MSG1='IODF78',MSG2='SC30.IODF78 - 2009-05-07 09:26', * SYSTEM=(2097,1),LSYSTEM=SCZP201, * TOK=('SCZP201',00800006991E2094092649000109127F00000000,* 00000000,'09-05-07','09:26:49','SC30','IODF78')

The suffix of the IODF data set that is shown in the IOCDS ID statement and displayed by the D IOS command should be the same number (that is, IODF78). This proves that the IOCDS was built from the current active IODF data set and that the latest changes and new IO definitions are successfully activated.

Configuring the IBM Storage System DS8000

The host adapter ports in the DS8000 storage control unit have to be configured for native FICON protocol. Follow the procedure described in Appendix D, Configuring the DS8000 for FICON on page 259, to configure the host adapter ports. The configuration steps are based on the configuration shown in Figure 7-1 on page 134. For planning and implementation information refer to IBM System Storage DS8000: Architecture and Implementation, SG24-6786.

Configuring the FICON Directors

The FICON Directors must be configured to match the designated configuration, shown in Figure 7-1 on page 134. For the desired cascaded FICON Directors topology, this includes: Specifying the Domain ID in the Directors Enabling the ports Checking the port settings Make sure that the selected ports (LX/SX) in the Director match the FICON node (LX/SX) that will be attached.

Chapter 7. Configuring a cascaded FICON Director topology

143

All the required steps to configure the FICON Directors are described in detail in Chapter 8, Configuring FICON Directors on page 151. Go to 8.1.1, Configuration flowchart on page 152 and follow the procedures described there. Return to this section after the FICON Directors are configured and ready for use.

Connecting the fiber optic cables

After all the definition and configuration tasks are completed, fiber optic cables have to be plugged in to the FICON channels, CU host adapters, and the FICON Directors. All fiber links from the FICON channel to the FICON Director, and from the CU host adapters to the FCION Director, use the longwave (LX) type of laser. For the LX type of laser, a 9m single mode (SM) fiber optic cable is recommended to achieve maximum performance and distance. All fiber optic cables must have an LC duplex connector at both ends. Plug the fiber optic cable connectors to the designated ports (shown in Figure 7-1 on page 134) to the FICON channel, the CU, and the FICON Director, as listed here: PCHID 5A2 to FICON Director 65 port 03 PCHID 5D2 to FICON Director 66 port 1B CU host adapter port 0003 to FICON Director port 28 CU host adapter port 0242 to FICON Director port 3F The fiber optic cables between both FICON Directors (ISL) have to be plugged when you configure the FICON Directors. Make sure the fiber optic cable ends are cleaned before plugging them into the transceiver. Ensure that all fiber optic cables are labeled and documented for later use (for example, when recabling or for problem determination). For further information and considerations regarding fiber cabling and documentation, refer to 4.11, Physical connectivity on page 93 and 4.2, Documentation on page 62.

Configuring the CHPIDs online

After the new I/O definitions are successfully activated, the channel path can be configured online and checked to see if we can communicate to the devices. Note, however, that even if all definitions in the IOCDS are correct, the channel status might be offline. Enter D M=CHP(7E) on the system console to display the current status of CHPID 7E. If the CHPID is offline, enter CF CHP(7E),ON on the z/OS operator console to configure the channel online. The channel status should change to ONLINE, as shown in Example 7-7.
Example 7-7 CHPID 7E online state

CF CHP(7E) IEE502I CHP(7E),ONLINE IEE712I CONFIG PROCESSING COMPLETE To achieve the best performance on the FICON channel, make sure that zHPF is enabled. See 4.9.6, High Performance FICON on page 88 for considerations about exploiting zHPF on System z servers. Enter D IOS,ZHPF at the z/OS console to display zHPF settings for the z/OS image. If zHPF is disabled, enter SETIOS ZHPF,YES. This enables zHPF, as shown in Example 7-8.

144

FICON Planning and Implementation Guide

Example 7-8 Enabling zHPF

D IOS,ZHPF RESPONSE=SC30 IOS630I 13.27.23 ZHPF FACILITY 021 HIGH PERFORMANCE FICON FACILITY IS ENABLED Using the SETIOS ZHPF=YES command enables zHPF temporarily. However, after the next system IPL, the zHPF facility will be reset to the default (disabled). To permanently enable zHPF for z/OS, change the FCX parameter in SYS1.PARMLIB member IECIOSxx to FCX=YES. Now you can query the status and functional details of the channel path by entering D M=CHP(7E) at the operator console. The command output is shown in Example 7-9. It provides information about the channel and the attached devices.
Example 7-9 D M=CHP(7E)

D M=CHP(7E) IEE174I 13.22.46 DISPLAY M 144 CHPID 7E: TYPE=1B, DESC=FICON SWITCHED, ONLINE DEVICE STATUS FOR CHANNEL PATH 7E 0 1 2 3 4 5 6 7 8 9 A B C D E 0006 . . . . . + . . . . . . . . . 0D00 + + + + + + + + + + + + + + + 0D01 + + + + + + + + + + + + + + + 0D02 + + + + + + + + + + + + + + + 0DFD HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFE HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA 0DFF HA HA HA HA HA HA HA HA HA HA HA HA HA HA HA SWITCH DEVICE NUMBER = NONE ATTACHED ND = SLKWRM.DCX.BRD.CA.0AFX0642C00Z PHYSICAL CHANNEL ID = 05A2 FACILITIES SUPPORTED = ZHPF

F . + + + HA HA HA

Example 7-9 shows that CHPID 7E is online and operating. Information is also displayed about the attached devices and the supported facilities (for example, zHPF). To check that communication to the attached devices is working properly, enter the command D M=DEV(xxxx), where xxxx is any device number. For example, to check the status of device 0065 (the CUP in FICON Director 65) and device D000 (a storage device), enter D M=DEV(0065,D000) on a z/OS console, as shown in Example 7-10.
Example 7-10 D M=DEV(0065,D000)

D M=DEV(0065,D000) IEE174I 11.59.11 DISPLAY M 986 DEVICE 0065 STATUS=ONLINE CHP 7E 82 ENTRY LINK ADDRESS 6503 651B DEST LINK ADDRESS 65FE 65FE PATH ONLINE Y Y CHP PHYSICALLY ONLINE Y Y PATH OPERATIONAL Y Y MANAGED N N CU NUMBER 0065 0065
Chapter 7. Configuring a cascaded FICON Director topology

145

MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = SLKWRM.DCX.BRD.CA.0AFX0642C00Z.0003 SCP TOKEN NED = SLKWRM.DCX.BRD.CA.0AFX0642C00Z.0000 SCP DEVICE NED = SLKWRM.DCX.BRD.CA.0AFX0642C00Z.0000 DEVICE D000 STATUS=ONLINE CHP 50 54 58 5C 7E 82 ENTRY LINK ADDRESS 0F 0F 1B 1B 6503 651B DEST LINK ADDRESS 2C 2C 0A 0A 6688 66CF PATH ONLINE N N Y Y Y Y CHP PHYSICALLY ONLINE N N Y Y Y Y PATH OPERATIONAL N N Y Y Y Y MANAGED N N N N N N CU NUMBER D000 D000 D000 D000 D000 D000 MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107.900.IBM.75.0000000BALB1.0300 SCP TOKEN NED = 002107.900.IBM.75.0000000BALB1.0000 SCP DEVICE NED = 002107.900.IBM.75.0000000BALB1.0000 HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW, ZHPF

The response to D M=DEV(0065,D000) displays all available paths to the devices and their status. Information is provided about the device (for example, device type), the control unit number, the link address, and the functions supported by the device (for example, MIDAW and zHPF).

Verifying the installation

After all the configuration tasks in the z10 server, the DS8000 storage, and the FICON Directors are completed, you can verify that the current configuration matches the desired configuration shown in Figure 7-1 on page 134. This verification should be done to ensure that the fiber optic cables are correctly plugged to the designated ports in the z10 server, the DS8000 and the FICON Directors, and to check the status of channels and devices. First, check the status of PCHID 5A2 in the z10 server. From any HMC where the CPC (server) is defined (for example, SCZP201), start a Single Object Operation (SOO) to that CPC. Display the list of channels installed on the CPC. Search the list for the desired PCHID (5A2) and double-click the PCHID icon. This will display the PCHID details panel shown in Figure 7-6 on page 147.

146

FICON Planning and Implementation Guide

Figure 7-6 PCHID details

Essential information for PCHID 5A2 is shown on the PCHID Details panel: The PCHID status is Operating. The PCHID type is FICON Express8. CSS.CHPID 2.7E (and 0.7E) is assigned to PCHID 5A2. The owning image of PCHID 5A2 is A23 and A24 (among other images). The CHPIDs assigned to PCHID 5A2 are shared across images. Similar information can be found on the CHPID Details panel (see Figure 7-7).

Figure 7-7 CHPID details

The CHPID Details panel provides similar information to the PCHID details panel. To display the CHPID details, select the channel list for an operating system image from the Support Element workplace. As shown, information for PCHID 5A2 (CHPID 2.7E) provided on the details window matches the designated configuration. This proves that PCHID 5A2 has CHPID 2.7E assigned and images A23 and A24 can access the channel. Repeat this checking on other channels that have been recently configured.

Chapter 7. Configuring a cascaded FICON Director topology

147

Important: If any of the data displayed on the PCHID or CHPID detail panel does not match the desired configuration, you must correct the definitions in the IOCDS. Next, check that the FICON channels are connected to the correct FICON switch and port. On either the PCHID details or CHPID details panel, click the Channel Problem Determination button. This will display the Channel Problem Determination panel, where you can select which information you want to display; see Figure 7-8.

Figure 7-8 Channel Problem Determination panel

Select Analyze channel information and click OK. The Analyze Channel Information window is displayed as shown in Figure 7-9 on page 149, and it provides information about the node attached to the FICON channel.

148

FICON Planning and Implementation Guide

Figure 7-9 Analyze Channel Information

Information about the nodes is displayed at the bottom part of the window. The lower left side displays information about the node in the System z server. The lower right side displays information about the attached node. Important: Verify that the Node status shown for both nodes is Valid. If any other status is shown, none of the information displayed is valid. Check that the Type/model information, as well as the serial number (Seq. number), is as expected. Next, examine the tag field for each node. The tag provides information about the port number of the node. The two right-most digits of the tag value represent the CHPID number for the channel node (7E), and for the attached node it represents the port number (0003). Be aware, however, that the tag value is provided by the attached device during link initialization and may have different meanings, depending on the vendor. The World Wide Node Name (WWN) and the World Wide Port Name (WWP) are also shown for each port, and may be used to prove that the channel is connected to the correct FICON Director (if the WWN or WWP of the attached device is known). If the node status is not Valid or the tag value and WWPN are not correct, you must check the fiber optic cable link between the z10 server and the FICON Director to ensure that it is plugged correctly.

Chapter 7. Configuring a cascaded FICON Director topology

149

Next, check the Channel link address field, which shows the switch address and the port number of the attached Director. The two left-most digits represent the switch address (65). The two digits in the middle represent the port number (03). In our scenario we are now sure that PCHID 5A2 has connectivity to port number 03 on switch address 65, which matches our desired configuration (see Figure 7-1 on page 134). If the displayed values are not as expected, it means that the fiber optic cables may not be plugged correctly and must be checked. After completing the preceding steps and proving that the physical path to the FICON Director and the logical definitions of the link are correct, check that the path to the control unit image is initialized correctly and properly defined. On the Channel Problem Determination panel, shown in Figure 7-8 on page 148, select Analyze Serial Link Status and click OK. The Analyze Serial Link Status window is displayed, as shown in Figure 7-10.

Figure 7-10 Analyze Serial Link Status

The Analyze Serial Link Status window provides status information about the link to the control unit images defined in the IOCDS. Scroll through the list of CU images and check that status for all CUs is Initialization Complete. In this case, Figure 7-10 shows a link status of Initialization Complete for all defined CU images on link address 65FE and 6628. In a cascaded FICON configuration, you must use a two-byte link address to specify the link address. That means link address 65FE is the CUP port in switch 65, and link address 6688 points to a storage control unit attached to switch 66. Link address 6688 is the destination port where the CU is physically attached. The link to CU address 00 and 01 (among other CU images) is initialized. CU address 00 and 01 correspond to control unit number D000 (CUADD=0) and D1000 (CUADD=1), as defined in the IOCDS. Information is also displayed that PCHID 05A2 (CHPID 2.7E) is connected to port 03 on switch address 65 (Switch number and Channel link address). If the link status Initialization Complete is not shown, you must check that the ports in the FICON Director and CU are correctly configured, and that the fiber optic cable link between the Director and the CU has the correct cable type and plugging.

150

FICON Planning and Implementation Guide

Chapter 8.

Configuring FICON Directors

This chapter explains how to configure IBM 2499 FICON Directors. It provides configuration examples for switched point-to-point and cascaded FICON Director topologies. The following topics are discussed: Configuration overview Installing and using Data Center Fabric Manager Setting up a FICON Director Setting up cascaded FICON Directors FICON Directors in an extended distance environment FICON Directors in an intermixed environment Backing up Director configuration data Backing up DCFM configuration data For information about configuring Cisco MDS FICON Directors, refer to Cisco FICON Basic Implementation, REDP4392.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

151

8.1 Configuration overview

In this chapter, we show our approach to configuring FICON Directors in different environments. All examples are based on an IBM System Storage SAN768B Director (2499-384) and an IBM System Storage SAN384B Director (2499-192). Configuration for all other B-type products is similar. The window captures are based on Data Center Fabric Manager (DCFM) Version 10.1.3 and on Fabric OS (FOS) Version 6.2.0e. For switched point-to-point and cascaded FICON Director environments we illustrate how to set up the following: IP addresses (for switch management) Licensed features Logical switches Domain ID and Insistent Domain ID Port-Based Routing (PBR), Dynamic Load Sharing (DLS) and In-Order Delivery (IOD) Control Unit Port (CUP) Port type and speed Buffer credits Allow/Prohibit matrix Zoning Port fencing Inter-Switch Links (ISLs) including Traffic Isolation (TI) and Quality of Service (QoS) High-integrity fabric and policies (fabric binding) The configuration described in this chapter is based on considerations explained in Chapter 4, Planning the FICON environment on page 59. Terminology: Throughout this chapter the terms FICON Director and switch are used interchangeably. In all cases, the term switch refers to FICON Director.

8.1.1 Configuration flowchart

The flowchart displayed in Figure 8-1 on page 153 illustrates the approach we used to configure our FICON Directors for the environments discussed in Chapter 6, Configuring a switched point-to-point topology on page 115 and Chapter 7, Configuring a cascaded FICON Director topology on page 133. Configuring switched point-to-point and cascaded FICON Directors topologies are very similar. The only differences are the link addressing scheme on the System z server side, the need for the high-integrity fabric function, and the implementation of ISLs. Therefore, 8.3, Setting up a FICON Director on page 164 applies to both topologies. 8.4, Setting up cascaded FICON Directors on page 190 only applies to cascaded FICON Directors.

152

FICON Planning and Implementation Guide

8.3.1, Changing the IP addresses on page 164

8.3.7, Changing port type and speed on page 179

8.3.2, Enabling features (optional) on page 169

8.3.8, Changing buffer credits on page 183

8.3.3, Setting up a logical switch (optional) on page 170

8.3.9, Setting up the Allow/Prohibit Matrix (optional) on page 184

8.3.4, Configuring the Domain ID and Insistent Domain ID on

8.3.10, Setting up zoning (optional) on page 185

8.3.5, Setting up PBR, IOD, and DLS on page 175

8.3.11, Configuring Port Fencing (optional) on page 188

8.3.6, Enabling the Control Unit Port on page 177

8.4.1, Setting up Inter-Switch Links on page 190

8.4.2, Setting up a high integrity fabric on page 192

Figure 8-1 Configuration flowchart

Depending on your requirements, some optional steps may not be needed, and the last two steps are only relevant to a cascaded FICON Directors topology. However, if you plan on installing a single FICON Director and want to use the 2-byte link addressing on the System z server for future use in a two-site configuration, you must configure the Switch Connection Control (SCC) as described in 8.4.2, Setting up a high integrity fabric on page 192. In the SCC policy, you would only have one WWNN entry (from that particular FICON Director). Before we started with the configuration of the FICON Directors, we installed the Data Center Fabric Manager (DCFM). DCFM was used to set up our FICON Directors, as well as to configure some optional functions.

8.1.2 FICON Director management connectivity

The FICON Directors and DCFM servers should always be interconnected on a separate LAN to isolate Director management traffic from other IP traffic. For controlling and maintaining the IBM 2499 FICON Directors, both CP cards in each chassis must have connectivity to the DCFM server. For each FICON Director, a total of three
Chapter 8. Configuring FICON Directors

153

IP addresses are required: two for the CP cards, and one for the Director. One IP address is required for each DCFM server. Note: IP addresses in the range of 10.0.0.1 to 10.0.0.255 are reserved for the internal communication of the IBM 2499 FICON Directors and must not be used for the CP cards or Directors. Because the DCFM server and client applications do direct polling of the fabric to gather configuration and status information, they must be able to access every Director in the fabric. Any Director the DCFM server and client cannot reach will go grey within their respective DCFM application. Our FICON Director management environment is shown in Figure 8-2. It also shows the IP addresses used for our FICON Directors, DCFM server, and firewall/router.
Data Center Fabric Manager Server/Client Data Center Fabric Manager Client

System Storage SAN768B

10.77.77.61 10.1.1.20

SW 65 CP0 CP1
10.1.1.21 10.1.1.31 10.1.1.22 10.1.1.32 10.1.1.10

CP0 CP1 SW 66
10.77.77.62 10.1.1.30

10.1.1.1

Firewall

Corporate Network

System Storage SAN384B

Figure 8-2 Our FICON Director management environment

In this configuration, the DCFM server and all CP cards in the FICON Directors are connected through an Ethernet LAN (known as the service LAN). The first three octets of the subnet mask for the service LAN are set to a value of 255.255.255.x to ensure that it is isolated from the corporate network. The forth octet (x) of the subnet mask can be adjusted based on the number of FICON Directors and DCFM servers that are connected to the service LAN. In our case, we are using 255.255.255.0, which will support up to 254 IP addresses. A router with firewall capabilities was used to allow only remote access to the FICON Directors and the DCFM server from the DCFM client, which resided in the corporate network.

8.2 Installing and using Data Center Fabric Manager

The Data Center Fabric Manager (DCFM) software is used to manage several Directors and switches through one graphical user interface (GUI). With the DCFM you are able to configure and monitor all fabrics, Directors, and switches that are installed and connected via the network to the Management Server.

154

FICON Planning and Implementation Guide

The DCFM is a client-server management software package. After the server software is installed, all other users can download the client part of the package by typing the servers IP address in a Web browser address field. Check the Data Center Fabric Manager (DCFM) data sheet for hardware and software prerequisites at: ftp://ftp.software.ibm.com/common/ssi/pm/sp/n/tsd03064usen/TSD03064USEN.PDF

8.2.1 Installing Data Center Fabric Manager

Follow these steps to install DCFM on the server: 1. Get the installation DVD or the installation package for DCFM. If autorun is enabled, the installation will start. Otherwise, double-click the installation package from the installation source (located in folder Windows\install.exe). 2. Follow the instructions shown on the window until you reach the Installation Complete window shown in Figure 8-3.

Figure 8-3 Installation Complete window

Ensure that Launch DFCM Configuration is selected, then click Done. 3. On the Welcome window, click Next. 4. If you do not need to migrate data from a previous installation, then select No and click Next (see Figure 8-4 on page 156). If you are migrating data, select the appropriate field and then type the location of the installation you wish to import.You can import data from an Enterprise Fabric Connectivity Manager (EFCM), a Fabric Manager (FM), or from a previous Data Center Fabric Manager (DCFM) installation.

Chapter 8. Configuring FICON Directors

155

Figure 8-4 Data migration window

5. Enter the serial number (from the DVD box) and the Server License (from the Key Certificate). If the window does not appear, then your installation does not require a license. The License Key field is not case sensitive. 6. Select Internal FTP Server, as shown in Figure 8-5, and click Next.

Figure 8-5 Select FTP Server

156

FICON Planning and Implementation Guide

Note: You can change the FTP Server settings later in the DCFM by selecting SAN Options. In the Options window, select Software Configuration FTP/SCP and make the required changes. 7. From the pull-down menu in Figure 8-6, select the IP address that will be used by the clients as Return Address (10.1.1.10) and the address that will be used to connect to the Directors and switches as Preferred Address (10.1.1.10). If only one network adapter is installed in the server, then the address will be the same for both pull-down menus. Important: Do not select 127.0.0.1 as Return Address, because if it is selected, clients will not be able to connect to the servers Web application to download the client part of the software. If you would prefer to select the server name, make sure that it is configured in the DNS server. 8. After making your selection, as shown in Figure 8-6, click Next.

Figure 8-6 Select communication interface

9. Enter the appropriate port numbers (see Figure 8-7 on page 158), if you are not using the default values. The Syslog Port # must remain as 514, because this port number cannot be changed in the FICON Director. Note: If a firewall is installed between the DCFM server or client and the FICON Directors, then the configured ports need to be defined in the firewall to allow that traffic flow. Keep in mind that 16 consecutive ports will be needed from the defined Starting Port #. Also, do not use port 2638, because it is used internally by the DCFM server.

Chapter 8. Configuring FICON Directors

157

Figure 8-7 Configure network ports

Tip: You can change these port numbers later in the DCFM by selecting SAN Options. In the Options window, select Software Configuration Server Port to make the changes. To change the SNMP Port #, use Monitor SNMP Traps. 10.Click Next and select the size of your FICON/SAN installation, which will be managed by the DCFM server as shown in Figure 8-8 on page 159.

158

FICON Planning and Implementation Guide

Figure 8-8 Select SAN size

11.Click Next after you have made the selection. 12.Wait until the database is initialized, then click Next at the Server License Summary window. 13.At the Start Server window select Start DCFM Client, then click Finish. 14.Wait until the login window appears (see Figure 8-9). If this is a migration or upgrade from a previous management software installation, then you need to use the user and password from that version. If it is a new install, then the user is administrator and the password is password. Click Login.

Figure 8-9 DCFM Login window

Chapter 8. Configuring FICON Directors

159

15.You will see the DCFM Login Banner, which you can change or remove if it is not needed. This can be done after clicking OK. Figure 8-10 will appear.

Figure 8-10 DCFM first startup

16.Select SAN Options. In the Options window, select Security Misc. and change the text in the Banner Message field or just remove the check mark, if it is not needed.

Installing the Open DataBase Connectivity (ODBC) Driver (optional)

Now you can install the ODBC Driver, to be able to export data from the DCFM database to generate customized reports (for example, to Microsoft Excel). The ODBC Driver needs to be installed on every personal computer that will generate customized reports. Refer to 8.2.2, Using the DCFM on page 161 for an explanation of how to change the default password. Follow these steps to install the ODBC Driver: 1. Get the installation DVD or the installation package for DCFM. 2. Double-click the installation package from the installation source (located in folder odbc\Windows\install.exe). 3. Follow the instructions shown on the windows until Figure 8-11 on page 161 is displayed.

160

FICON Planning and Implementation Guide

Figure 8-11 ODBC configuration window

4. Enter the server IP address from the DCFM installation (in our example, 10.1.1.10). The Port Number is 2638 (this port is automatically assigned during DCFM installation and cannot be changed). The user is guest and the default password is password. 5. Click Next, then Done. The installation is complete. To import data from the DCFM database and create a customized report in Microsoft Excel, follow these steps: 1. Open Microsoft Excel and select Data Import External Data New Database Query. 2. At the Choose Data Source panel, select DCFM* (this entry was created with the installation of the ODBC Driver.) and click OK. 3. Now you can select the required entries and add them to your selection by clicking the arrow pointing to the right. Click Next after your selections are made. 4. In the Filter Data panel you can set filters, if needed. Click Next when done. 5. Select a sort order, if needed. Click Next when done. 6. The last panel will give you the possibility to save your query setup. Click Finish. 7. Select the cell you where want to enter the queried data and click OK. If you need to refresh the queried data of a saved form, select Data Refresh Data.

8.2.2 Using the DCFM

In this section we explain functions in the DCFM that are needed during setup. For other functions, you can always use the help button in the DCFM.

Chapter 8. Configuring FICON Directors

161

Discovering a FICON Director

Before you can manage a FICON Director, it must be discovered. 1. Go to Discover Setup in the menu bar of the DCFM. 2. Click Add to add a Director. A new window appears. 3. Enter the Directors IP address, username (admin), and the password (the default is password). You can also provide a Fabric Name. 4. Select the SNMP tab, and select v3 in the SNMP Version drop-down menu. Accept the default values, then click OK. 5. Click Close to leave the discovery dialog. The FICON Director is now discovered and can be managed via DCFM. During the configuration, it may happen that you will have to rediscover a Director. To do this, select the discovered Director and click Delete. The Director will show up in the lower part of the Discover Setup window. Select the Director and click Discover to rediscover it. Enter the new Fabric Name and click OK. If you want to change the admin password, select Edit and change the password for the selected Director. Click OK after the password is changed. (Note that this password change is only for the communication between DCFM and the Director. The password itself will be changed via the Element Manager of the Director as described in Creating and changing users and passwords.)

Using the DCFM client

To use the client from any personal computer, you must open a Web browser and enter the IP address of the DCFM server in the Web browsers address field. You will reach the welcome window from the DCFM server. You can also download the required Java Version and the MIB tables if needed for external system monitoring software. Click Web Start the DCFM Client. The client will be downloaded to your workstation and then the login window will appear. The handling is exactly the same as at the server. A shortcut will be created in your program list, so that you do not need to download the client again.

Creating and changing users and passwords

To change the password, or create a new user for the DCFM, go to SAN Users at the DCFM menu bar. Click Edit at the Users field if you want to change the password. Click OK after the password is changed. To create a new user, click Add and assign the user to a Resource Group by clicking the arrow pointing to the right. You can also create new user roles or edit an existing user role at the Roles field. If you click Add in the Resource Groups field, you can build a new group that includes only the Directors that you assign to it. Click OK to activate all your changes. To change the passwords on the Director, right-click the Director in the DCFM and select Element Manager Admin. In the Switch Administrator window, click the User tab. Click the Change Password button to change the password for the selected user, then click OK after the password is changed. Click Apply to activate the changes. To change the ODBC password on the DCFM server, go to the DCFM program directory in the command prompt and enter the following command dbpassword <Username> <Password> <New_Password> <Confirm_Password>, as shown in Example 8-1 on page 163.

162

FICON Planning and Implementation Guide

Example 8-1 Changing the ODBC password

C:\Program Files\DCFM 10.1.3\bin>dbpassword guest password test1test test1test DB is updated Successfully C:\Program Files\DCFM 10.1.3\bin>

Now you need to uninstall the ODBC driver on all workstations that use this function and install it again with the new password, as explained in Installing the Open DataBase Connectivity (ODBC) Driver (optional) on page 160.

Renaming the Director and ports

To change the Director name, right-click the Director in the DCFM and select Properties. The properties window opens. Double-click in the Name field and type the new name. You can also change the port names by clicking the Port tab in the properties window and clicking in the Name field to change the port name. Click OK after all changes are done. The port name can also be changed by right-clicking the Director and selecting Element Manager Ports. Select the Port in the Port Administration window and click the Rename button. Type the name for the port and click Rename to save the changes. To change the Fabric Name, right-click the Fabric and select Properties. You can change the Name, Description, Location and Contact by double-clicking the desired field. Click OK after the changes are done.

Changing the view for FICON management

To change the view for FICON, click SAN Options. The Option window appears. Select Display and click the Set Up FICON Display button. Click OK to close the window.

Setting the display to hex

To view the display in hexadecimal, select the hex display option from the pull-down menu on the toolbar as illustrated in Figure 8-12.

Figure 8-12 Set to hex

Viewing node descriptors

If you have changed the view for FICON management, you will see the node descriptor details in the DCFM for each port on the left side of the window. To change the Port view, right-click a Director (on the left of the window) and select Port Display. Check-mark the details you want to display. You can also display the node descriptor for a Director, by right-clicking it and selecting Element Manager Hardware. In the Directors Element Manager, click Name Server on the left side of the window. The Name Server window opens and you will see all connected devices. By selecting the list entry and clicking Detail View, you will get a detailed view for the selected device.

Chapter 8. Configuring FICON Directors

163

8.3 Setting up a FICON Director

In this section we illustrate how to set up FICON Directors, which applies to Directors in both topologies. We also include optional steps for the most commonly used functions. Additional configuration steps for a cascaded FICON Director are covered in 8.4, Setting up cascaded FICON Directors on page 190.

8.3.1 Changing the IP addresses

Before connecting the FICON Director to the network, you need to change the default IP address to match your network layout and rules. The steps to configure the Directors IP addresses are as follows: Connecting a laptop or personal computer to the Directors active Control Processor (CP) Configuring the laptop or personal computer to be in the same network Altering the Directors IP addresses Connecting the Director to the customer network Note: If you are redeploying a FICON Director and do not know its IP addresses, then the serial port will have to be used. Steps to configure the Directors IP addresses using the serial port are described in IBM System Storage SAN768B Installation, Service, and User's Guide, GA32-0574 and in IBM System Storage SAN384B Installation, Service, and User's Guide, GA52-1333. Follow these steps to change the IP addresses in the Director: 1. Connect a personal computer or laptop to the active Control Processor Card (CP0 in Slot 6 or CP1 in Slot 7). Use the Management Port to connect the network cable. The active CP is indicated by the blue LED on the card. 2. Alter the personal computer or laptop IP address to be in the same network as the Director. Go to Start Control Panel Network Connections (use the classic view in the control panel). 3. Right-click the Network Adapter and click Properties, as shown in Figure 8-13 on page 165.

164

FICON Planning and Implementation Guide

Figure 8-13 Select network interface

4. Select Internet Protocol (TCP/IP) and click Properties (see Figure 8-14).

Figure 8-14 Select Protocol

5. Select Use the following IP address and enter the address (10.77.77.70, in this case) to be in the same network as the Director; see Figure 8-15 on page 166. 6. Click OK on both Properties windows.

Chapter 8. Configuring FICON Directors

165

Figure 8-15 Setting up a laptop IP address

7. Connect the cable to the active CP Card in the Director and open a Web browser. Enter the default IP address in the Web browsers address field (10.77.77.77, in this case) and press Enter. The default IP address for CP0 in Slot 6 is 10.77.77.75. The default IP address for CP1 in Slot 7 is 10.77.77.74. 8. In the login window, enter the default user (admin) and the default password (password), then click OK. (The Mozilla Firefox Web browser will ask for open or save. Select open with and click OK.) If you receive a message like the one shown in Figure 8-16, you will need to install a new Java version (Version 1.6 or later). If you do not receive this message, proceed to step 9 on page 167.

Figure 8-16 Wrong Java version message

If Version 1.6 is installed, go to Start Control Panel Folder Options (in classic view). Select the File Types tab, then scroll down to and select JNLP. Click Advanced. In the Edit File Type window, select Launch and click Edit. The window shown in Figure 8-17 on page 167 will appear.

166

FICON Planning and Implementation Guide

In the field Application used to perform action you must change the location of your Java version; for example, C:\Program Files\Java\jre1.6.0_05\bin\javaws.exe" "%1.

Figure 8-17 Assign Java version to JNLP files

Click OK after the changes are made. The login window should appear after you enter the IP address in the Web browsers address field. Type in the user (admin) and password (password) and click OK. 9. At the Directors graphical user interface, select Manage Switch Admin. Select the Network tab at the Switch Administration window to reach Figure 8-18 on page 168.

Chapter 8. Configuring FICON Directors

167

Figure 8-18 Setup management IP address

10.Fill in the required fields. Our example is based on Figure 8-2 on page 154. Table 8-1 lists the IP addressing scheme we used.
Table 8-1 IP address example Virtual address Ethernet IP Ethernet mask Gateway IP 10.1.1.20 255.255.255.0 10.1.1.1 CP0 address 10.1.1.21 255.255.255.0 n/a CP1 address 10.1.1.22 255.255.255.0 n/a

Important: The Virtual address is later used to connect to the Director. This will guarantee a connection even if a Control Processor switchover has occurred (for example, during a firmware upgrade). When connecting with the IP address of CP0 or CP1, only the active CP is manageable. IP addresses between 10.0.0.0 to 10.0.0.255 cannot be used because they are already used internally in the Director. 11.To activate the new IP addresses, scroll down to the bottom of the window and click Apply. 12.Connect each Control Processor Management Port to the network (Ethernet switch). The Director needs to be discovered by DCFM, as described in Discovering a FICON Director on page 162.

168

FICON Planning and Implementation Guide

Using the FICON wizard

The FICON wizard provides you with a simple means to set up the following configuration options: Insistent Domain ID Set high integrity fabric Enable port-based routing Enable in order delivery Disable dynamic load sharing Enable the FICON Management Server (FMS) for Control Unit Port (CUP) management You can access the FICON wizard via DCFM by right-clicking the Director and selecting Configure FICON Configure Fabric. If you decide not to use the FICON wizard or you need to configure other options not included in the FICON wizard, refer to the remaining parts of this section.

8.3.2 Enabling features (optional)

To use a licensed feature, it must first be enabled. 1. Go to the Switch Admin window in the Element Manager of the Director. This can be done via DCFM by right-clicking the Director and selecting Configure Element Manager Admin. 2. Select the License tab, as shown in Figure 8-19.

Figure 8-19 License overview

3. Click Add, then enter the license key for the feature you want to enable, as shown in Figure 8-20 on page 170. Note: The license key is case sensitive and will only work on the Director to which it is assigned. Therefore, you need to compare the WWN on the license with your Director.

Chapter 8. Configuring FICON Directors

169

Figure 8-20 Add a license

4. Click Add License to activate the license. If you have more than one license, then repeat step 3 on page 169 for each. The feature is now enabled and can be used or configured.

8.3.3 Setting up a logical switch (optional)

A logical switch in the Director can be used to separate the physical ports for different System z environments or for mixed environments (FICON and Fibre Channel). A maximum of eight logical switches can be created. Two of these switches can have FICON Management Server and CUP enabled. The default logical switch has Logical Fabric ID (FID) 128 assigned. If you want to assign port IDs (PIDs), then you can configure a logical switch and assign all ports from the default logical switch to the new logical switch. The default logical switch is always present if Virtual Fabric is enabled, but it does not need to have ports assigned to it. If the Director uses 48 port cards, you must create two logical switches with Zero-based Area Assignment. If the Director was delivered with Firmware level 6.2.0 (required for logical switches), then you have by default one logical switch, thereby allowing you to create more. Important: If the Firmware was updated to level 6.2.0, then you need to enable Virtual Fabric first, which is disruptive. Follow these steps to create a logical switch: 1. If Virtual Fabric is enabled (indicated by the blue V on the Director in the DCFM Switch view), then proceed to step 5 on page 171. Otherwise, select the Director in DCFM that you want to enable for Virtual Fabric in the Chassis Group. 2. Right-click the Director and select Enable Virtual Fabric, as shown in Figure 8-21 on page 171.

170

FICON Planning and Implementation Guide

Figure 8-21 Enable Virtual Fabric

3. Read the warning message and select OK. Important: Clicking OK will immediately reboot the Director and create a logical switch, which is disruptive to the Director. 4. Wait for the Director to complete the reboot. You need to discover the Director again after the reboot, because the Director was changed and appears as a new Director. Refer to Discovering a FICON Director on page 162 for more details about this topic. 5. Select Configure Logical Switches at the DCFM menu bar. 6. Wait for the Logical Switches configuration window to appear, as shown in Figure 8-22 on page 172. 7. In the Chassis drop-down menu, select the chassis where you want to create a new logical switch.

Chapter 8. Configuring FICON Directors

171

Figure 8-22 Create a Logical Switch

8. Select Undiscovered Logical Switch, then click the New Switch button. 9. Enter the desired settings for the fabric using the Fabric tab; see Figure 8-23.

Figure 8-23 Logical switch configuration

The Logical Fabric ID (FID) can be a number from 1 to 128 (decimal). The FID must be unique in the selected chassis. Note: Keep in mind that only logical switches with the same FID can be cascaded. Do not select Base Switch or Base Fabric for Transport, and do not set the 256 Area Limit to Disable, because this is not supported by FICON. The 256 Area Limit can be set to Zero-Based Area Assignment. Ports will be numbered in ascending order, starting from 00 to FF, for each logical switch. Ports with a Port Index above 255 can be added, but only in Brocade Native Interoperability Mode or Port Based 172
FICON Planning and Implementation Guide

Area Assignment (ports with a Port Index above 255 cannot be added; instead, the default Port numbering will be used). You can change the port numbering for logical switches (except for the default logical switch) later in the Port Admin panel in the Element Manager by using the Bind PID and Un-Bind PID buttons. 10.Select the Switch tab and assign the Domain ID and the logical switch name. Also, check-mark the Insistent Domain ID field. 11.Click OK after all settings are complete. 12.Now you need to assign the desired ports to the new logical switch. Select the ports on the left side, and click the arrow pointing to the right; see Figure 8-24.

Figure 8-24 Assign ports to a logical switch

13.After the ports are assigned, click OK. 14.Verify the changes in the confirmation window, and select Re-Enable Ports after moving them (if not selected, all ports must be enabled manually), then click Start. 15.Wait for the changes to complete. After the status Success displays, click Close. 16.At this point you need to discover the new logical switch. Instructions that explain this process are provided in Discovering a FICON Director on page 162. The logical switch is now created and can be used like any other physical switch or Director.

8.3.4 Configuring the Domain ID and Insistent Domain ID

To set up the Domain ID, the Director must be disabled, as explained here. We also explain how to set the Insistent Domain ID flag. This will allow you to later merge two Directors without disruption. Only Domain IDs between 1 and EF in hex are possible, which is decimal 1 to 239. The default Domain ID is 1.

Chapter 8. Configuring FICON Directors

173

Follow these steps to change the Domain ID to the desired value: 1. Go to the Switch Admin window in the Element Manager of the Director. This can be done via the DCFM by right-clicking the Director and selecting Configure Element Manager Admin. 2. Select the Switch tab, as shown in Figure 8-25.

Figure 8-25 Domain ID configuration

3. Select Disable in the Switch Status field and click Apply. Important: Setting the Director to disable is disruptive. 4. Read the warning message, and confirm it by clicking Yes. 5. Now you can change the Domain ID field to the desired value. In our example it will be 65 in hex, which is 101 in decimal. The value has to be entered in decimal. You can also change the Name of the Director so it can be identified easily. Note: The Domain ID must match the switch address configured in HCD/IOCP on the System z side. Be aware that the switch address is configured in hex, but the Domain ID in the Director is defined with a decimal value. 6. Click Apply. After reading the warning message, click Yes to confirm the action. 7. If not already done, click Show Advanced Mode in the upper right corner of the Switch Administration window. 8. Click the Configure tab. 9. Click Insistent Domain ID Mode to check-mark it, as shown in Figure 8-26 on page 175.

174

FICON Planning and Implementation Guide

Figure 8-26 Configure Insistent Domain ID

10.Click Apply to save the changes. 11.To enable the Director at the Switch Status field, return to the Switch tab. (You may also skip this step and configure Port-Based-Routing (PBR), Dynamic Load Sharing (DLS) and In-Order Delivery (IOD), before you enable the Director.) 12.Select Enable in the Switch Status field and click Apply. 13.Read the warning message and click Yes to confirm the action. The Domain ID is now set. Note that it cannot be automatically changed if another Director joins the Fabric with the same Domain ID. Instead, the new Director will be segmented until another Domain ID is set at the joining Director.

8.3.5 Setting up PBR, IOD, and DLS

Port-Based Routing (PBR) and In-Order Delivery (IOD) must be set in a FICON environment. Dynamic Load Sharing (DLS) can be used in a FICON environment if lossless DLS is also enabled; otherwise, it must be set to disabled. Lossless DLS will ensure that no frames are dropped during rebalancing of the Inter-Switch Links (ISLs), which will be the case if an ISL is added or removed. However, lossless DLS is not supported in long distance environments. When lossless DLS is used in long distance environments, you must disable creditrecovery at all ISLs. Follow these steps to set up PBR, IOD, and DLS: 1. Return to the Switch Admin window in the Element Manager of the Director. This can be done via the DCFM by right-clicking the Director and selecting Configure Element Manager Admin. 2. If the Director is already disabled, go to step 6 on page 176. Otherwise, select the Switch tab. 3. Select Disable in the Switch Status field and click Apply.
Chapter 8. Configuring FICON Directors

175

Important: Setting the Director to disable is disruptive. 4. Read the warning message, and confirm it by clicking Yes. 5. Click Show Advanced Mode in the upper right corner. 6. Select the Routing tab, as shown in Figure 8-27.

Figure 8-27 Set up Port-Based Routing, DLS and IOD

7. Select Port-Based-Routing in the Advanced Performance Tuning (APT) Policy field. Set Dynamic Load Sharing (DLS) to Off if lossless DLS should not be used. Set In-Order Delivery (IOD) to On. Click Apply to save the changes. 8. If you have set DLS to On, then you need to enable lossless DLS also via the command-line interface (CLI). Otherwise, proceed to step 10. The following steps explain one method of using the CLI: a. In the DCFM server or client, right-click the Director and select Telnet. A new window opens. (You can also use an ssh connection to the Director, if needed.) b. Enter the user name (in our case, admin) and press Enter. c. Enter the password and press Enter (the default is password). d. Enter, for example, setcontext 1 to connect to a logical switch with FID 1. e. Type the command iodset --enable -losslessdls and press Enter To check the settings, enter iodset --show and dlsshow. f. For long distance environments, execute portcfgcreditrecovery --disable slot/port for each ISL. g. Type logout and close the window. 9. To enable the Director at the Switch Status field, return to the Switch tab. (You may also skip this step and configure the Control Unit Port (CUP) before enabling the Director.) 10.Select Enable in the Switch Status field and click Apply.

176

FICON Planning and Implementation Guide

11.Read the warning message and click Yes to confirm the action. The Director is now ready to be used in a FICON environment.

8.3.6 Enabling the Control Unit Port

If the Director has a CUP license, then the Control Unit Port (CUP) can be enabled. In this case, ports FE and FF in a 256 Port Director must be disabled because those port addresses are used by the CUP server. Refer to 8.3.7, Changing port type and speed on page 179 for more information. The Director must be enabled before you can disable the ports. The CUP can be enabled while the Director is enabled. A Registered State Change Notification (RSCN) will be sent to inform the System z server that the CUP is now enabled. However, because our example is in the process of setting up the Director, you can do it while the Director is disabled. Follow these steps to enable the CUP: 1. Go to the Switch Admin window in the Element Manager of the Director. This can be done via the DCFM by right-clicking the Director and selecting Configure Element Manager Admin. 2. If the Director is already disabled, proceed to step 6. Otherwise, select the Switch tab. 3. Select Disable in the Switch Status field and click Apply. Important: Setting the Director to disable is disruptive. 4. Read the warning message, and confirm it by clicking Yes. 5. Click Show Advanced Mode in the upper right corner. 6. Select the FICON CUP tab, as shown in Figure 8-28 on page 178.

Chapter 8. Configuring FICON Directors

177

Figure 8-28 Enable the CUP

7. Select Enable in the FICON Management Server Mode field. 8. Click Apply. 9. Read the message and click Yes to confirm the action. 10.Now you can alter the FICON Management Server Behavior Control (Mode Register) field. Select the desired functions you want to use with the CUP. 11.Click Apply after the changes are made. 12.Read the message and click Yes to confirm the action. 13.Return to the Switch tab, to enable the Director at the Switch Status field. 14.Select Enable in the Switch Status field and click Apply. 15.Read the warning message and click Yes to confirm the action. The CUP Port is now usable from the System z server.

Missing Interrupt Handler Process Timeout (MIHPTO)

The default timeout value for the CUP is 180 seconds. To check the value, or to change it, you need to connect via telnet or ssh to the Directors IP address. Here is one method for accomplishing this task: 1. In the DCFM server or client, right-click the Director and select Telnet. A new window opens. (You can also use an ssh connection to the Director, if needed.) 2. Enter the user name (in this case, admin), then press Enter. 3. Enter the password and press Enter (the default is password); see Example 8-2 on page 179. 4. Enter, for example, setcontext 1 if you need to connect to a logical switch with FID 1. 178
FICON Planning and Implementation Guide

Example 8-2 FICON CUP MIHPTO

IBM_SAN768B login: admin Password: ----------------------------------------------------------------IBM_SAN768B:FID128:admin> ficoncupshow mihpto MIHPTO for the CUP: 180 seconds IBM_SAN768B:FID128:admin> ficoncupset mihpto 181 The input value of the MIHPTO will be rounded down to the nearest value divisible by 10. This is due to IBM's MIHPTO specification. MIHPTO has been changed to 180 seconds IBM_SAN768B:FID128:admin>logout 5. To see the timeout that is currently set, enter ficoncupshow mihpto. 6. To change the value, enter ficoncupset mihpto xxx, where xxx is the new value. This value must be divisible by 10 if over 63, or will be rounded down. A value between 15 and 600 seconds can be used. 7. Enter logout and close the window. The MIHPTO for the CUP is now set and active.

8.3.7 Changing port type and speed

To change a specific port type, speed, or Extended Distance Link, you must modify the port configuration. The default port type is a U_Port, which automatically detects if an F_Port, E_Port, or L_Port device is connected. The default speed is set to auto-negotiate. Important: Changing the port configuration is disruptive to port operations. Follow these steps to configure a port: 1. Go to the port configuration window of the Directors Element Manager. Right-click the Director in the DCFM and select Element Manager Ports. 2. The Port Administration window will appear. Click Show Advanced Mode in the upper right corner to see the configuration options. 3. Select the port you want to configure and click Disable. Important: This is disruptive for the traffic on the selected port. 4. Read the information at the confirmation panel and click Yes. 5. Click Edit Configuration for the selected port; see Figure 8-29 on page 180.

Chapter 8. Configuring FICON Directors

179

Figure 8-29 Port type configuration

6. Select the Port Type you want to configure, or accept the defaults if you only want to change the speed. Note the following points: - E_Port is for ISLs. - L_Port is for loop devices. - F_Port is for all other devices. Be aware that EX_Port is not selectable, because it is only available at the Base Switch in a logical switch configuration, which is not supported for a FICON environment. 7. Click Next and Figure 8-30 on page 181 will appear.

180

FICON Planning and Implementation Guide

Figure 8-30 Port speed and distance configuration

Note the following points: - At the Speed drop-down menu, you can select Auto (for auto-negotiate) or 1, 2, 4, and 8 Gbps. - The Ingress Speed Limit (MBps) drop-down menu is a Quality of Service function, to limit the throughput of the port (200, 400, 600, 800, 1000, 1500, 2000, 3500, 4000, 5000, 7000 and 8000 Mbps are possible). - With Long Distance Mode, you can configure the distance of the link (L0:normal is the default value). For more detailed information about this topic, refer to 8.3.8, Changing buffer credits on page 183. 8. Click Next after all selections are done. 9. At the Confirmation window, click Save after verifying the settings. 10.Click Close after the configuration complete message. 11.Click Enable to enable the port after the configuration is complete. After the port type and speed are configured, it is useful to give the port a name by clicking Rename.

Verifying and changing the fill word

A fill word is sent when there is no traffic on the link, in order to keep it online. For 8 Gbps SFPs, the fill word should be ARB(FF), according to the specifications. However, some interfaces were developed before the final specifications were approved. These interfaces use idles to fill the link, which is the standard for 1 Gbps, 2 Gbps, and 4 Gbps. The default (when a Director is delivered with FOS 6.2) is set to ARB(FF), which is only valid for 8 Gbps SFPs. If 4 Gbps SFPs are present, idles will be sent even if ARB(FF) is configured. The mode can only be verified and changed via the CLI.

Chapter 8. Configuring FICON Directors

181

Important: Changing the mode is disruptive for the port. The following method illustrates one way to change the fill word for a 8 Gbps interface: 1. In the DCFM server or client, right-click the Director and select Telnet. A new window opens. (You can also use an ssh connection to the Director, if needed.) 2. Enter the user name (in this case, admin) and press Enter. 3. Enter the password and press Enter (the default is password). 4. Enter, for example, setcontext 1 if you need to connect to a logical switch with FID 1. 5. The portcfgfillword 1/8 0 command changes the fill word for Slot1, Port8 to Idles. The portcfgfillword 1/8 1 command changes the fill word for Slot1, Port8 to ARB(FF). 6. The portcfgshow command will list all ports of the Director. The portcfgshow 1/8 command will only list the settings for Slot1, Port8 (see Example 8-3).
Example 8-3 Changing fill words

Fabric OS (DCXBOT) Fabos Version 6.2.0e_rc1_bld01

DCXBOT login: admin Password: ----------------------------------------------------------------IBM_SAN768B:FID128:admin> portcfgfillword 1/8 0 IBM_SAN768B:FID128:admin> portcfgshow Ports of Slot 1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 -----------------+--+--+--+--+----+--+--+--+----+--+--+--+----+--+--+-Speed AN AN AN AN AN AN AN AN AN AN AN AN AN AN AN AN Fill Word 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 ~~~ IBM_SAN768B:FID128:admin>logout 7. You can also use a script that changes all ports on a given port card, as shown in Example 8-4.
Example 8-4 Script to change the fill word for a 16-port card in slot 2

for ((i=0;i<16;i++)); do (echo 2/$i;portcfgfillword 2/$i 1); done This script will change the settings for the first 16 ports at the card in Slot 2 to ARB(FF), which is specified by the 1 in the script. If a 32-port card is used, you must replace 16 with 32. The slot also needs to be changed in both positions in the script. For a 32-port card in slot 3, it will look as shown in Example 8-5.
Example 8-5 Script to change the fill word for a 32-port card in slot 3

for ((i=0;i<32;i++)); do (echo 3/$i;portcfgfillword 3/$i 1); done 8. Enter logout and close the window. The fill word is now changed, and the Director will use the configured mode.

182

FICON Planning and Implementation Guide

8.3.8 Changing buffer credits

Buffer credits are used as a flow control mechanism to avoid overwriting stored frames received in a port. In most cases, extended distances require that the number of buffer credits be increased at the port; refer to Buffer credits on page 39 for more information. The default is 8 buffer credits per F-Port. Follow these steps to change the buffer credits for an F_Port: 1. In the DCFM server or client, right-click the Director and select Telnet. A new window opens. (You can also use an ssh connection to the Director, if needed.) 2. Enter the user name (in this case, admin) and press Enter. 3. Enter the password and press Enter (the default is password). 4. Enter, for example, setcontext 1 if you need to connect to a logical switch with FID 1. 5. Disable the port by entering: portdisable 1/8. 6. This command will change the buffer credits for a given port: portcfgfportbuffers --enable 1/8 80 This command will change the buffers for port 8 in the card in slot 1 to 80: portcfgfportbuffers --disable 1/8 7. Enable the port by entering: portenable 1/8. 8. To display the assigned buffers use the portbuffershow 1 command. This will show the assigned buffers for each port at the card in Slot 1, as well as the remaining buffers for the port group. To change the buffer credits for an ISL, use the Element Manager of the Director as follows: 1. Follow the steps in Changing port type and speed on page 179. 2. Open the drop-down menu for the Long Distance Mode. 3. Select one of the following options: LE:<=10 km (only use if the link is shorter than 10 km). A total of 46 buffers will be allocated for a 8 Gbps link. A total of 26 buffers will be allocated for a 4 Gbps link. A total of 16 buffers will be allocated for a 2 Gbps link (a 1 Gbps link uses the default of 8 buffers). LD:Auto - and enter the maximum link length in the Desired Distance (km) field. The buffer credits will be calculated during link initialization of the E_Port and automatically allocated. The value you enter in the Desired Distance (km) field will determine the number of buffers allocated to the port based on a 2 KB frame size. LS:Static - and enter the link length in the Desired Distance (km) field. The buffer credits will be set for the given distance and the selected speed. For an explanation of how buffer credits work, refer to Buffer credits on page 39. 4. Click Next after all selections are done. 5. At the Confirmation window, click Save after verifying the settings. 6. Click Close after you receive the configuration complete message. 7. Click Enable to enable the port after the configuration is done.

Chapter 8. Configuring FICON Directors

183

8.3.9 Setting up the Allow/Prohibit Matrix (optional)

To prohibit traffic between selected ports, you can use the Allow/Prohibit Matrix. To use the Allow/Prohibit Matrix, zoning must be active. If you use the default zone, then create a new zone with all ports, as shown in 8.3.10, Setting up zoning (optional) on page 185. Follow these steps to change the Allow/Prohibit Matrix: 1. Go to the Switch Admin window in the Element Manager of the Director. This can be done via the DCFM by right-clicking the Director and selecting Configure Element Manager Admin. 2. Click Show Advanced Mode in the upper right corner. 3. Select the FICON CUP tab. 4. Select the CUP Port Connectivity tab in the lower left corner of the FICON CUP tab. The window shown in Figure 8-31 opens. 5. Perform one of the following actions: - To make changes directly, select the Active Configuration and click Edit. - To create a new matrix, click New. - To copy an existing configuration, click Copy; you must provide a new Name and Description in the pop-up window and then click OK.

Figure 8-31 Allow/Prohibit configurations

6. After you make your selection and click Edit or New, the window shown in Figure 8-32 on page 185 appears.

184

FICON Planning and Implementation Guide

Figure 8-32 Allow/Prohibit Matrix

At this point, you can modify the matrix by clicking in the fields to allow or prohibit a connection. Figure 8-32 shows that port 03 (hex) is prohibited to connect to ports 00, 01, and 02. All other ports are allowed to connect to port 3. Note the following points: - The Block button will block the selected port. It is the same function as a port disable. - The Prohibit button will prohibit all connections to and from the selected port. If you do not want to activate the changes right away, then click Save (not available for the active matrix), or Save As, to create a new configuration. 7. After making the desired changes, you can activate the changes by clicking Activate. 8. Click Yes to confirm the action.

8.3.10 Setting up zoning (optional)

Because FICON connectivity is defined in the System z server based on a port address in the FICON Director, you do not need to set up zoning. The default zone in the Director is sufficient. However, if you want to separate some FICON channels and control units (for a test environment), follow these steps: 1. In the DCFM, select Configure Zoning. The window shown in Figure 8-33 on page 186 appears.

Chapter 8. Configuring FICON Directors

185

Figure 8-33 Zoning dialog overview

2. In the Zoning Scope drop-down menu in the upper left corner of the window, select the Director for which you want to create Zoning. 3. To create a new Zone, click the New Zone drop-down menu and select New Zone. 4. Enter a name for the new Zone (in this case, Test_LPAR). 5. Ensure that Domain Port Index is selected in the Type drop-down menu, because port-zoning is used in FICON environments. 6. Select the ports you want to include in the zone you just created at the left side of the window. (You may select more than one port at a time by pressing and holding Ctrl key on the keyboard while selecting the ports with the mouse. If you want to select a range of ports, use the Shift key.) 7. To place the selected ports in the new zone in the middle of the window, click the arrow pointing to the right. Repeat steps 3 to 7 for all zones, you want to create. In our example, we created a Rest_of_the_world zone, which includes all other ports of the Director (except ISLs). The result is shown in Figure 8-34 on page 187.

186

FICON Planning and Implementation Guide

Figure 8-34 Creating zones

8. After all zones are created, you must put them in a Config. Click the New Config button. Enter a name for the new Zoneconfig (in this case, FICON_Zoneset). 9. Select the zones in the middle of the window that you want to place in the newly created Zone Config, then click the arrow pointing to the right. The result is shown in Figure 8-35.

Figure 8-35 Creating a Zone Config

10.Now you must activate the new Zone Config. Select the Zone Config (in this case, FICON Zoneset) and click Activate. 11.You will reach the Activate Zone Config window. Verify that all ports are in the correct Zones. Click OK after you verify all zones this will activate the newly created Zone Config. 12.Click Yes when you reach the confirmation window. 13.After activation you will see that all active zones, as well as the active Zone Config, have green markers attached (see Figure 8-36 on page 188).

Chapter 8. Configuring FICON Directors

187

Figure 8-36 Zoning activated

To verify the active zone, click the Active Zone Config tab. You will see the active Zone Config, as well as the active zones with all zone members. 14.To leave the Zoning dialog, click OK. 15.If you get a confirmation window, read the warning and click OK. Zoning is now completed and active in the Director.

8.3.11 Configuring Port Fencing (optional)

Port Fencing is used to force a port offline if a specified error rate is reached. A bad link will cause error recovery on the attached device, which could lead to degraded performance. If the link is an ISL, this will lead to recovery for many devices. When Port Fencing is configured, the port will be set offline automatically to keep the fabric stable. After the link is checked or repaired, the port will have to be manually set online. To configure Port Fencing, follow these steps: 1. In the DCFM, select Configure Port Fencing from the menu bar. The window shown in Figure 8-37 on page 189 appears.

188

FICON Planning and Implementation Guide

Figure 8-37 Port Fencing dialog

2. Select the Fencing Policy from the Violation Type drop-down menu (in this case, the Invalid CRCs policy). 3. Select the policy on the left part of the window, or click Add to create a new one. If you want to change or verify the policy, click Edit. 4. Two selections are presented: m-EOS (which is for McData Directors) and FOS. In this case, we selected FOS for our SAN768B Director. There are two possible policy types in FOS: the Default and the Custom Policy. We selected the default policy. You can also change the policy name in this panel. (Some policies are only for FOS or m-EOS as indicated at the selection from the drop down menu.) Click OK after the settings are done. 5. Assign the policy by highlighting it on the left site, and highlighting the Director on the right side (or the fabric), then clicking the arrow pointing to the right. 6. The policy is added to the Director as indicated by the green plus (+) sign. 7. Click OK to activate the policy on the Director (or fabric). You have now assigned the policy to the Director. When the threshold is reached for the assigned error type, the port will go offline. Refer to Fabric OS (FOS) Administration Guide, the Data Center Fabric Manager (DCFM) User Manual, and the Fabric OS Command Reference Manual for other functions required for your installation, but not covered in this section.

Chapter 8. Configuring FICON Directors

189

8.4 Setting up cascaded FICON Directors

A cascaded FICON Director is one that is connected to another Director via ISLs. In a cascaded environment, the System z FICON channel has to use the 2-byte addressing scheme, which includes the switch address (first byte) and the port address (second byte). The FICON Directors that will be merged have the same setup as any other FICON Director. Therefore, you must first complete the tasks in 8.3, Setting up a FICON Director on page 164 for both FICON Directors.

8.4.1 Setting up Inter-Switch Links

Inter-Switch Links (ISLs) are used to connect the two FICON Directors. Each ISL port in the FICON Director is configured as an E_Port. Both sides of the ISL must have the same setup for them to function properly. This means that the port type (E-Port), port speed, Long Distance Mode, Trunking, and Quality of Service must be the same on both sides of the ISL. See 8.3.7, Changing port type and speed on page 179 for more detailed information.

Trunking (optional)
If trunking is enabled for an ISL port, it will automatically form a trunk. For trunking, all links in the trunk (maximum 8) must be in the same port group and must be set up with the same speed (for example, 8 Gbps). All fiber optic links in the trunk must be the same length (maximum of 30m difference). If two of four ports have Quality of Service (QoS) enabled, two trunks will be formed with two ports each (even if they are in the same port group): one trunk with QoS, and one trunk without QoS.Trunking will only work in Brocade native Interoperability Mode. To enable trunking, follow these steps: 1. Right-click the Director in the DCFM and select Element Manager Ports. 2. In the Port Administration window, click Show Advanced Mode in the upper right corner. 3. From the list, select the port that you want to enable for trunking. 4. Click the Enable Trunking button. Important: This will directly enable trunking for the selected port. If this port is online, the link will go offline for a short moment, which will interrupt the traffic on the link.

Traffic Isolation (TI) zoning (optional)

Using Traffic Isolation (TI) zoning, you can isolate traffic for some devices, thereby forcing them to use a specific ISL or trunk exclusively. If an ISL is selected that is a member of a trunk, the whole trunk will be assigned to the TI zone. The trunk or ISL cannot be used by any other TI zone. TI zoning is set up like regular zoning, but ISLs are included. Follow these steps to create a TI zone: 1. Perform steps 1 to 7 in Setting up zoning (optional) on page 185, but select New TI Zone instead of New Zone. Add both sides of the ISL or trunk to the new TI zone, which should be named with TI at the beginning of the zone name. The TI zone will not be included in the Zone Config on the right side of the Zoning window (see Figure 8-38 on page 191).

190

FICON Planning and Implementation Guide

Figure 8-38 TI zoning setup

2. To activate the TI zone, right-click the TI zone and select Configured Enabled. You can also select Configured Failover. If Failover is enabled, the devices in the TI zone will use another available ISL (which is not in the TI Zone). When the configured ISL is available again, it will automatically fall back. Important: When Failover is disabled, then the devices in the TI zone will not use another available ISL outside its TI zone. 3. To activate one or more created TI zones in the Directors, click Activate in the Zoning window. 4. Compare the settings at the confirmation window and select OK to activate the zoning changes. 5. If a second confirmation window appears, read the message and select Yes. 6. Click OK after the changes are activated. 7. To verify that the TI zone is active, click the Active Zone Config tab.

Quality of Service (QoS) (optional)

Using Quality of Service (QoS), you can decrease the usable bandwidth for a device. This is done with the Ingress Rate Limit (MBps) setup already shown in 8.3.7, Changing port type and speed on page 179. To enable QoS, follow these steps: 1. Right-click the Director in the DCFM and select Element Manager Ports. 2. In the Port Administration window, click Show Advanced Mode in the upper right corner. 3. From the list, select the port that you want to enable for QoS. 4. Click QoS Enable for the selected port. There is also a possibility to use QoS zones. However, you must use WWN zoning instead of port zoning.
Chapter 8. Configuring FICON Directors

191

To use Quality of Service zones, follow these steps: 1. Complete the steps in Setting up zoning (optional) on page 185, but select WWN instead of Domain, Port Index in the Type drop-down menu. 2. Right-click the created zone and select QoS Priority QoS_High or QoS_Low as shown in Figure 8-39. The default is QoS_Medium, for all zones.

Figure 8-39 QoS priority selection for WWN Zoning

3. The name of the zone will be changed to QoSH (for high priority) or QoSL (for low priority) in front of the given zone name. 4. To activate the QoS zone, click Activate in the zoning window. 5. Compare the settings at the confirmation window and select OK to activate the zoning changes. 6. If a second confirmation window appears, read the message and select Yes. 7. Click OK after the changes are activated. 8. To verify that the zone is active, click the Active Zone Config tab.

8.4.2 Setting up a high integrity fabric

To merge the two Directors, the Switch Connection Control (SCC) Policy must be set up to form a high integrity fabric. Therefore, you need to set the policy to Strict, which will distribute the policy only to FOS 5.2.0 or later Directors, resulting in a stable and secure fabric. In the

192

FICON Planning and Implementation Guide

SCC policy, we show how include the World Wide Node Name (WWNN) of the two Directors in the fabric. To do this, follow these steps: 1. Go to the Switch Admin window in the Element Manager of the Director. This can be done via the DCFM by right-clicking the Director and selecting Configure Element Manager Admin. 2. Click Show Advanced Mode in the upper right corner. 3. Select the Security Policies tab. 4. On the left side of the window, click FWCP, which is the Fabric Wide Consistency Policy. 5. Select Strict at the SCC Consistency Behavior drop-down menu. 6. Click Apply. Note: This will change the security settings of the Directors. From now on, each System z server needs to use the 2-byte addressing scheme to initialize the link. 7. Select ACL on the left side of the window. 8. Click Edit to set up the SCC Policy. 9. Select only SCC and click Next. 10.Click Modify to add the WWNNs of the two Directors. 11.Select the Director on the left side, and click Add Switch (see Figure 8-40). This will add the WWNN of the Director you are connected to.

Figure 8-40 Creating SCC policy

Chapter 8. Configuring FICON Directors

193

12.In the Add Other Switch field, enter the WWNN of the second Director and click Add. (The WWNN of the Director is shown on top of the Switch Admin window). 13.Click OK, then click Next. 14.Verify the settings you just made, and click Finish. 15.Back in the Switch Administration window, you now see Defined Policy Set. Click Activate to activate the Policy (see Figure 8-41).

Figure 8-41 SCC policy activated

16.Perform steps 1 on page 193 through 6 on page 193 on the second FICON Director. (Set the Fabric Wide Consistency Policy for SCC to Strict.) 17.Connect the ISLs. The ISLs will be initialized, and a cascaded fabric will be created. Remember to set up the ISLs on both sides to the same settings as described in Setting up Inter-Switch Links on page 190. The SCC Policy will be distributed to the second Director automatically. In DCFM, you will now see that the two Directors are merged and connected via ISLs, as shown in Figure 8-42 on page 195.

194

FICON Planning and Implementation Guide

Figure 8-42 FICON Cascaded Director DCFM overview

To verify the ISLs, right-click them in the DCFM Directors View and select Properties. You will see all active ISLs and their settings on both Directors, as shown in Figure 8-43.

Figure 8-43 ISL properties window

Chapter 8. Configuring FICON Directors

195

If ISLs are missing, go to the Port Admin window of the two Directors and check the settings on both sides. Use the Advanced View to see all settings. Scroll to the right side and check the Additional Port Info field (this field also displays some error reasons for the ISL).

8.5 FICON Directors in an extended distance environment

In an extended distance environment, using a Wave Division Multiplexing (WDM) platform will require you to set up the Director to match the WDM configuration. Check which WDM platforms are IBM System z qualified and ensure that you are using the correct SFPs; otherwise, you can overdrive them in the WDM. You will need to configure the ISL ports as E-Ports (not auto-detect U_Ports) and use a fixed speed (not auto-negotiate). Also, the Long Distance Mode must to be configured for the correct distance. See 8.3.7, Changing port type and speed on page 179 and 8.3.8, Changing buffer credits on page 183 for configuration information. By default, the Director uses VC_Link_Init with ARB Fillword to initialize the ISL. However, some Dense WDM platforms do not support this mode. For those WDM platforms, you must use Idle Fillwords, or disable it and use ISL R_RDY Mode. Also, you must not use Quality of Service and Credit Recovery in this mode. They must be disabled before IDLE Fillwords are used, or ISL R_RDY Mode is enabled. ISL R_RDY does not support trunking. The following commands are needed: portcfgqos --disable slot/port to disable QoS. portcfgcreditrecovery --disable slot/port to disable credit recovery. portcfglongdistance slot/port LS 1 100 to enable Long Distance in LS Mode for 100 km with VC_Link_Init enabled for ARB Fillword (1). portcfglongdistance slot/port LS 0 100 to enable Long Distance in LS Mode for 100 km with VC_Link_Init enabled for IDLE Fillword (0). portcfgislmode slot/port, 1 to enable ISL R_RDY Mode (1). portcfgislmode slot/port, 0 to disable ISL R_RDY Mode (0). portcfgfillword slot/port 0 to set the Fillword to Idles (0). portcfgfillword slot/port 1 to set the Fillword to ARB(FF) (1). portcfgshow slot/port to verify the settings. Example 8-6 illustrates how to set up Long Distance Mode to 100 km (LS Mode = static) with ISL R_RDY Mode and Idles as Fillword, which is required for some WDM platforms.
Example 8-6 DWDM example for Port 1/15

portcfgqos --disable 1/15 portcfgcreditrecovery --disable 1/15 portcfglongdistance 1/15 LS 0 100 portcfgislmode 1/15, 1 portcfgfillword 1/15 0 portcfgshow 1/15

196

FICON Planning and Implementation Guide

8.6 FICON Directors in an intermixed environment

If you use FICON and Fibre channel in the same Director, we strongly recommend that you create two logical switches: one logical switch for FICON, and the other for Fibre Channel. For configuration details, refer to 8.3.3, Setting up a logical switch (optional) on page 170. This will allow you to configure a FICON logical switch with all required settings. and a Fibre Channel logical switch with its own settings.

8.7 Channel swap

A channel swap may be useful if you need to move a connected device to another physical port in the Director without changing the definition in the System z server or the zoning. The port will have the same address as it had before the swap. To perform a channel swap, right-click the director in the DCFM and select Element Manager Ports. In the Port Administration window, click Show Advanced Mode in the upper right corner to see all possible functions. Disable both the port you want to swap and the port you want to swap to. Note: If you have intermittent errors to a System z server, set the attached CHPID offline before disabling the port on the Director side. Set the CHPID online after the swap is complete and the port is enabled again on the Director side. Select the port you want to swap to and click Port Swap. Enter the Slot Number and the Port Number of the defective port and click Swap. After the swap is complete, plug the cable to the new port in the Director. Enable the port after the cable is connected and check the Node Descriptor for a valid connection, as shown in 10.7, Node descriptor on page 226. Set the port online in z/OS.

8.8 Backing up Director configuration data

To save the Director configuration data, right-click any Director and select Configuration Save. In the Save Switch Configurations dialog, select the Directors you want to back up and add them to the Selected Switches list by clicking the arrow pointing to the right. Click OK to save the Director configurations. You can also schedule a backup by right-clicking any Director and selecting Configuration Schedule Backup. Make the desired settings in the Schedule Backup dialog and click OK. The backup will be performed automatically according to your settings. You can view the backups by right-clicking any Director and selecting Configuration Configuration Repository. We recommend that you have a backup in case of a disaster (the configuration file can be restored in a replacement Director). You should also perform a backup before a firmware upgrade.

Chapter 8. Configuring FICON Directors

197

To restore, right-click the Director you want to restore and select Configuration Restore. You will obtain a list of available backup files. Select the backup you want to restore and click OK. Important: The restore of the Directors configuration data is disruptive, because the Director will perform a reboot to activate the new configuration.

8.9 Backing up DCFM configuration data

You should also backup the DCFM configuration. Select SAN Options from the DCFM menu bar. In the Options window, select Backup on the left and provide the necessary information. Click OK after the settings are done. Note: We recommend that you save the backup on a network drive in order to be able to restore the data in case of a disaster or a hard disk failure. The data can be restored via the DCFM Server console (located in the program directory). Therefore, you need to stop the DCFM services by clicking Stop. Select the Restore tab and enter the location of the backup, or browse the filesystem. After you select the backup location, click Restore.

198

FICON Planning and Implementation Guide

Part 4

Part

Managing the FICON environment

This part discusses FICON operations and problem determination tools and techniques. Common issues are examined and descriptions are provided that explain how to resolve these issues using proven methods when working with FICON environments.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

199

200

FICON Planning and Implementation Guide

Chapter 9.

Monitoring the FICON environment

This chapter discusses verification and monitoring tools and techniques for FICON environments. The following topics are discussed: System Activity Display (SAD) RMF Reporting DCFM performance monitor

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

201

9.1 System Activity Display

The System Activity Display (SAD) shows individual system activity for System z servers (also known as Central Processor Complexes or CPCs), or for a group of Central Processor Complexes. System activity includes the channel activity and physical processing activity that has been defined in the system activity profiles, which are stored in the selected CPCs. For more information about assigning and customizing activity profiles, refer to HMC Operations Guide, SC28-6830. Using the SAD, you can create a list of channels that you want to look at and monitor, if a performance problem is suspected to be caused by one or more FICON channels. Figure 9-1 shows a System Activity Display that is customized for high usage CHPIDs on processor SCZP201. However, for performance troubleshooting, we recommend that you use the Resource Measurement Facility (RMF).

Figure 9-1 SAD frame displaying high usage channel activity

Note: The utilization reported by the Activity task for most channel types will correspond with the utilization reported by the Resource Measurement Facility (RMF). For Fibre Channels, however, this task considers the channel to be busy any time an operation is pending, even if the channel is waiting for a device to respond. RMF, in contrast, looks at the amount of work performed versus the amount of work that could be performed by the channel. This means that if you have devices that are relatively slow to respond, leaving the channel waiting for a response but otherwise idle, activity will show a utilization that is significantly higher than that reported by RMF.

202

FICON Planning and Implementation Guide

9.2 Resource Measurement Facility monitoring

Resource Measurement Facility (RMF) is a performance management tool that measures selected areas of system activity, including different varieties of the FICON channel, and presents the data collected in the form of System Management Facility (SMF) records. This data is essential for any kind of FICON channel performance monitoring. For more information, refer to the IBM RMF Web site at: http://www.ibm.com/servers/eserver/zseries/zos/rmf/

9.3 Introduction to performance monitoring

Before you undertake any performance analysis, it is imperative to establish your performance objectives. The goal of performance management is to make the best use of your current resources to meet your objectives, without excessive tuning efforts. An essential ingredient in performance monitoring is understanding the components of response time. The measurements we calculate to report I/O response time and that RMF reports on are: Queue time Pend time Connect time Disconnect time Figure 9-2 on page 204 indicates the I/O performance measurement points. It depicts a configuration with multiple servers exploiting a cascaded FICON Director environment; however, the measurement points are also valid for the other supported topologies (by excluding the PEND time for the second switch or both switches).

Chapter 9. Monitoring the FICON environment

203

1. Queue time (IOS queuing activity)

z/OS 2 z/OS 1 z/OS 3

z/OS 4 z/OS 5

z/OS 6

2. PEND time (channel path activity)

z10 EC Server

z9 EC Server

z10 BC Server

3. PEND time (switch port activity)

Switch 1

Switch 2

4. PEND time (CU activity) 5. Connect time 6. Disconnect time

FICON DASD CU

Figure 9-2 I/O performance indicators

FICON exploits CCW chaining, so there will be fewer channel ends and device ends. Also, PEND time ends at the FICON channel when the first command response (CMR) is received from the CU.

9.4 Introduction to Resource Measurement Facility

The Resource Measurement Facility (RMF) provides an interface to your System z environment that facilitates reporting and detailed measurements of your critical resources. RMF is comprised of several components: Monitor I, Monitor II, Monitor III Postprocessor RMF Performance Monitoring Client/Server Reporter Sysplex Data Server Distributed Data Server LDAP back-end These components complement each other to provide the infrastructure for performance management: Gathering data Reporting data

204

FICON Planning and Implementation Guide

Accessing data across the sysplex We introduce data gathering and RMF reporting in the next sections.

9.4.1 Data gathering

RMF gathers data using three monitors: Short-term data collection with Monitor III Snapshot monitoring with Monitor II Long-term data gathering with Monitor I and Monitor III The system operator starts all monitors as non-interactive (background) sessions with a variety of options that determine what type of data is collected and where it is stored. The data gathering functions run independently on each system, but each monitor can be started sysplex-wide by one operator command.

9.4.2 RMF reporting

RMF has three monitors and a postprocessor for reporting performance statistics: RMF Monitor I produces interval reports that are created at the end of a measurement interval, for example, 30 minutes. You can obtain Monitor I session reports during or at the end of RMF processing, or they can be generated at a later time by the postprocessor. RMF Monitor II is a snapshot reporting tool used to obtain very fast information about how specific address spaces or system resources (processor, DASD, volumes, storage) are performing. Monitor II has two modes for reporting on system performance: A Monitor II display session, which can be invoked from either an ISPF dialog or directly with a TSO/e command (RMFMON) A Monitor II background session, which is a non-interactive session to create a report for printing RMF Monitor III runs as a TSO/e session under ISPF and provides sysplex or system performance reports by: Displaying your current system status in real-time mode Showing previously collected data that is still available in either storage buffers or preallocated VSAM data sets Monitor III offers a wide spectrum of reports for answering questions that arise during the various performance management tasks. All reporting is available within one TSO/e session, so there is no need to log on to different systems in the sysplex to get all performance data. The postprocessor is usually invoked via a batch job, but it can also be run under a TSO/e session. It offers the following types of reports: Interval reports, which reflect a picture of performance for each interval for which the data has been gathered Duration reports, which summarize data over longer periods of time with a maximum value of 100 hours Summary, plot, exception, and overview reports In addition, the postprocessor can create overview records that are the optimal base for further spreadsheet processing on the workstation.

Chapter 9. Monitoring the FICON environment

205

9.5 RMF example reports

The following examples have been produced using the RMF postprocessor. It should be noted that with any RMF reports, the I/O information is reported from one z/OS system only. If you are sharing I/O across multiple systems, you will need to review reports from each of the sharing systems to see the complete I/O activity.

9.5.1 DASD Activity Report

The DASD Activity Report (see Example 9-1) is the first starting point for any I/O analysis and consequently one we should focus on in some detail.
Example 9-1 DASD Activity Report
1 z/OS V1R10 TOTAL SAMPLES = STORAGE GROUP ACIRLS DEV NUM D308 D309 D30A D30D D30E D30F D310 D311 ... D320 DEVICE TYPE 33909 33909 33909 33909 33909 33909 33909 33909 33909 NUMBER OF CYL 10017 10017 10017 10017 10017 10017 10017 10017 600 IODF = 07 CR-DATE: 06/19/2009 DEVICE VOLUME PAV LCU ACTIVITY SERIAL RATE ACIV01 1.0H 0028 0.008 TBIG08 1.0H 0028 0.000 SAPTL3 1.0H 0028 0.000 NWD30D 1.0H 0028 0.000 IDGC00 1.0H 0028 0.000 #@$#X2 1.0H 0028 0.000 FKDBU3 1.0H 0028 0.000 COMST4 1.0H 0028 448.748 CR-TIME: AVG AVG RESP IOSQ TIME TIME .973 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .250 .000 10.00.10 AVG AVG CMR DB DLY DLY .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .018 .000 ACT: ACTIVATE AVG AVG AVG PEND DISC CONN TIME TIME TIME .128 .000 .845 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .144 .002 .105 % DEV CONN 0.00 0.00 0.00 0.00 0.00 0.00 0.00 4.63 0.00 0.05 % DEV UTIL 0.00 0.00 0.00 0.00 0.00 0.00 0.00 4.69 0.00 0.05 % DEV RESV 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 AVG NUMBER ALLOC 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.5 0.0 1.5 % % ANY MT ALLOC PEND 100.0 0.0 100.0 0.0 100.0 0.0 100.0 0.0 100.0 0.0 100.0 0.0 100.0 0.0 100.0 0.0 100.0 0.0 100.0 0.0 D I R E C T A C C E S S D E V I C E DATE 06/23/2009 TIME 15.39.37 A C T I V I T Y PAGE SYSTEM ID SC33 RPT VERSION V1R10 RMF INTERVAL 10.00.036 CYCLE 1.000 SECONDS 43

10017 TBIG26 1.0H 0028 0.000 .000 .000 .000 .000 .000 .000 .000 LCU 0028 448.780 .250 .000 .018 .000 .144 .002 .105

The fields of most interest are: AVG RESP TIME - This is the average response time, in milliseconds, for an I/O to this device or LCU. This can be considered as the overall measurement that reflects the health of the device. The measurement is composed of the sum of the following metrics: AVG CMR (command response) Delay - This delay indicates the time between a successful I/O operation being initiated on the channel and the device acknowledging the connection. These delays could be caused by contention in the fabric and at the destination port. AVG DB (device busy) Delay - This delay is due to the device being busy because of I/O from another sharing z/OS system. Channel path delay - Any time that is not accounted for in the preceding two measurements is due to a delay at the SAP or channel, or a CU busy. AVG DISC time - This reflects the time when the device was in use but not transferring data. AVG CONN time - This is the time measured by the channel subsystem during which the device is actually connected to the CPU through the path (channel, control unit, DASD) and transferring data. Figure 9-3 on page 207 illustrates when and how IOSQ, Pend, and Connect times appear in a DASD I/O operation when a CACHE hit occurs.

206

FICON Planning and Implementation Guide

Application I/O Request

Start I/O (IOS-CSS)

FICON DASD I/O Operation CU CACHE hit

I/O Operation Ends

IOSQ
UCB Busy PAV/HiperPAV reduces it with extra UCBs

Pend
Channel Busy FICON Director Port Busy # of Buffer Credits CU/Device Busy Open Exchange limit

Connect
Working Transferring Data

FC frame Multiplexing allows for better link utilization, but may extend some connect times FICON connect time is not as predictable as ESCON, more an awareness than a problem

Connect
Figure 9-3 FICON DASD I/O operation times with a CACHE hit

Figure 9-4 illustrates when and how IOSQ, Pend, ESS Logical time, and Connect times appear in an DASD I/O operation when a CACHE miss or an Extent conflict has to be resolved by the control unit.
Application I/O Request Start I/O (IOS-CSS)

FICON DASD I/O Operation CU CACHE miss / Extent conflict

I/O Operation Ends

IOSQ
UCB Busy PAV/HiperPAV reduces it with extra UCBs

Pend
Channel Busy FICON Director Port Busy # of Buffer Credits CU/Device Busy Open Exchange limit

Connect
Cache miss

ESS
Logical Disconnect

Connect
Working Transferring Data

Extent Confict

No Reconnect Port Busy (buffer credits)

Connect minus Logical Disconnect = Reported Connect

Figure 9-4 FICON DASD I/O operation times with a CACHE miss or an Extent Conflict

9.5.2 I/O Queueing Report

The I/O Queueing Report highlights percentages of activity rates and queue lengths for I/O components grouped by Logical Control Unit (LCU). A sample report is shown in Example 9-2 on page 208.

Chapter 9. Monitoring the FICON environment

207

Example 9-2 I/O Queuing Activity

1 I/O Q U E U I N G A C T I V I T Y SYSTEM ID SC30 DATE 06/23/2009 INTERVAL 10.00.037 RPT VERSION V1R10 RMF TIME 15.39.37 CYCLE 1.000 SECONDS SAMPLES = 600 IODF = 07 CR-DATE: 06/19/2009 CR-TIME: 10.00.10 ACT: ACTIVATE - INITIATIVE QUEUE ------- IOP UTILIZATION -------- % I/O REQUESTS RETRIED --------- RETRIES / SSCH --------ACTIVITY AVG Q % IOP I/O START INTERRUPT CP DP CU DV CP DP CU DV RATE LNGTH BUSY RATE RATE ALL BUSY BUSY BUSY BUSY ALL BUSY BUSY BUSY BUSY 1454.535 0.00 3.14 1454.535 1338.536 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0.00 0.00 0.00 1452.405 0.00 1.17 1452.405 1538.245 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0.00 0.00 0.00 1453.399 0.00 1.33 1453.397 2342.267 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0.00 0.00 0.00 1451.525 0.00 1.47 1451.525 2460.060 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0.00 0.00 0.00 1452.490 0.00 1.42 1452.490 2161.580 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0.00 0.00 0.00 1451.754 0.00 2.33 1451.754 1575.881 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0.00 0.00 0.00 8716.105 0.00 1.81 8716.105 11416.57 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0.00 0.00 0.00 AVG AVG DELAY AVG AVG DATA CU DCM GROUP CHAN CHPID % DP % CU CUB CMR CONTENTION Q CSS HPAV OPEN XFER MIN MAX DEF PATHS TAKEN BUSY BUSY DLY DLY RATE LNGTH DLY WAIT MAX EXCH CONC 2080 02 0.127 0.00 0.00 0.0 0.0 * 0.127 0.00 0.00 0.0 0.0 0.000 0.00 0.0 20A0 03 0.127 0.00 0.00 0.0 0.0 * 0.127 0.00 0.00 0.0 0.0 0.000 0.00 0.0 20C0 04 0.133 0.00 0.00 0.0 0.0 * 0.133 0.00 0.00 0.0 0.0 0.000 0.00 0.0 20E0 05 0.127 0.00 0.00 0.0 0.0 * 0.127 0.00 0.00 0.0 0.0 0.000 0.00 0.0 3080 1C 0.127 0.00 0.00 0.0 0.0 * 0.127 0.00 0.00 0.0 0.0 0.000 0.00 0.0 30A0 1E 0.127 0.00 0.00 0.0 0.0 * 0.127 0.00 0.00 0.0 0.0 0.000 0.00 0.0 4B00 4C 0.317 0.00 0.00 0.0 0.0 * 0.317 0.00 0.00 0.0 0.0 0.000 0.00 0.1 0.00 0.00 4B08 4E 0.013 0.00 0.00 0.0 0.0 * 0.013 0.00 0.00 0.0 0.0 0.000 0.00 0.1 0.00 0.00 z/OS V1R10

-TOTAL IOP 0 00 01 02 03 04 05 0 SYS LCU 0 0001 0 0002 0 0003 0 0004 0 0007 0 0008 0 000B 0 000C

In a PR/SM environment (such as our configuration), the report is split into two sections. The top section reports PR/SM system activity. The rest of the report applies to I/O activity for the z/OS system being measured by RMF. The fields we focus on are: PR/SM fields from the I/O Queuing Report INITIATIVE QUEUE: This queue reflects the I/O initiation activity for each I/O processor (IOP). Average queue length reflects the average number of entries on the initiative queue for each IOP. The activity rate reflects the assignment of I/O requests. IOP UTILIZATION shows fields to measure the I/Os started and interrupted on each IOP and an overall percentage busy indicator. Note: IOP or I/O Processors refer to the System Assist Processors (SAPs) on your server.

I/O REQUESTS RETRIED reflects the ratio of retried I/O requests. Retries/SSCH (start subchannel) reflects the retried I/O requests initially started. Reasons for each: CP - Channel path busy DP - Director port busy CU - Control unit busy DV - Device busy

z/OS System activity fields from the I/O Queuing Report LCU and CONTROL UNITS show the logical assignment to the physical resources. The Dynamic Channel Path Management (DCM) GROUP reports the minimum, maximum, and initially defined number of DCM channels for each LCU in this interval. CHPID taken shows how evenly I/O requests are spread across the available paths to the LCU.

208

FICON Planning and Implementation Guide

 The percentage busy fields refer to Director Port (DP) and control unit (CU). The value reflects the ratio of the number of times an I/O request was deferred due to either resource being busy. AVG CUB DLY indicates the time (in milliseconds) that an I/O operation was delayed due to the control unit being busy. AVG CMR DLY indicates the time (in milliseconds) between a successful I/O operation being initiated and the device acknowledging the connection. CONTENTION RATE indicates the rate that the I/O processor places the request on the control unit header for this interval. This occurs when all paths to the subchannel are busy and at least one path to the control unit is busy. DELAY Q LENGTH reflects the average number of delayed requests on the control unit header for this interval. AVQ CSS DLY reflects a time (in milliseconds) that an I/O is delayed from the acceptance of a start on the subchannel to the point that the channel subsystem actually initiates the operation on this LCU.

9.5.3 Channel Path Activity Report

Example 9-3 reflects a subset of a Channel Path Activity Report. The purpose of this report is to reflect channel path activity for the requested interval. Example 9-3 Channel Path Activity Report
1 z/OS V1R10 IODF = 07 CHANNEL ID TYPE 51 FC_S 52 FC_S 53 FC_S 54 FC_S 55 FC_S 56 FC_S 57 FC_S 58 FC_S 59 FC_S 5A FC_S 5B FC_S 5C FC_S 5D FC_S 5E FC_S CR-DATE: 06/19/2009 CR-TIME: 10.00.10 ACT: ACTIVATE MODE: LPAR CPMF: EXTENDED MODE PATH UTILIZATION(%) READ(MB/SEC) WRITE(MB/SEC) FICON OPERATIONS ZHPF OPERATIONS G SHR PART TOTAL BUS PART TOTAL PART TOTAL RATE ACTIVE DEFER RATE ACTIVE DEFER OFFLINE 4 Y 9.39 15.66 1.30 0.36 0.75 1.96 3.23 1176.8 1.5 0.0 0.0 0.0 0.0 4 Y 0.01 0.29 0.05 0.01 0.15 0.00 0.08 18.1 1.1 0.0 0.0 0.0 0.0 OFFLINE OFFLINE 4 Y 9.37 15.63 1.30 0.35 0.75 1.97 3.24 1176.3 1.5 0.0 0.0 0.0 0.0 4 Y 0.01 0.23 0.05 0.01 0.16 0.01 0.09 16.0 1.1 0.0 0.0 0.0 0.0 4 Y 2.58 3.45 0.25 0.18 0.71 0.00 0.16 219.3 1.0 0.0 90.6 1.0 0.0 4 Y 2.46 4.09 0.33 0.17 1.14 0.00 0.05 221.0 1.1 0.0 86.8 1.0 0.0 4 Y 9.41 15.67 1.30 0.35 0.76 1.96 3.22 1178.5 1.5 0.0 0.0 0.0 0.0 4 Y 0.01 0.40 0.07 0.01 0.20 0.01 0.12 21.3 1.1 0.0 0.0 0.0 0.0 4 Y 1.77 2.51 0.20 0.13 0.58 0.00 0.14 153.4 1.0 0.0 67.7 1.1 0.0 4 Y 1.65 3.28 0.30 0.13 1.15 0.00 0.06 157.1 1.1 0.0 65.4 1.0 0.0 4 Y 9.39 15.66 1.30 0.36 0.77 1.96 3.23 1178.4 1.5 0.0 0.0 0.0 0.0 CSSID: 2 C H A N N E L SYSTEM ID SC30 RPT VERSION V1R10 RMF P A T H A C T I V I T Y PAGE DATE 06/23/2009 TIME 15.39.37 INTERVAL 10.00.037 CYCLE 1.000 SECONDS 2

The fields of interest are: CHANNEL PATH ID reflects the CHPID number. TYPE reflects that the type (FC_S) is FICON switched. G reflects the generation type (2 denotes a 2 GBps channel - FICON). SHR indicates whether the channel is shared between one or more partitions.

UTILIZATION PART reflects the utilization for this partition for this interval. TOTAL reflects the utilization from the entire processor for this interval. BUS reflects the percentage of cycles the bus was found to be busy for this channel in relation to the potential limit.

Chapter 9. Monitoring the FICON environment

209

READ(MB/SEC) PART shows the data transfer rate in MBps for this channel from this partition to the control unit in this interval. TOTAL shows the data transfer rate in MBps for this channel from the entire processor to the control unit for this interval. WRITE(MB/SEC) PART shows the data transfer rate in MBps for this channel from this partition to the control unit in this interval. TOTAL shows the data transfer rate in MBps for this channel from the entire processor to the control unit for this interval. Note: On a machine running in LPAR mode, but with only one LPAR defined, the PART columns of this report will reflect a zero (0) value for the READ, WRITE, and UTILIZATION displays for FICON channels. FICON and zHPF Operations (Physical Channel) RATE is the number of native FICON or zHPF operations per second at the physical channel level. ACTIVE is the average number of native FICON or zHPF operations that are concurrently active. Often referred to as the number of open exchanges. DEFER is the number of deferred native FICON or zHPF operations per second. This is the number of operations that could not be initiated by the channel due to lack of available resources.

9.5.4 FICON Director Report

The measurements in this report reflect statistics regarding the ports that are performing I/O for all systems. Example 9-4 provides detailed configuration about each FICON Director and information pertaining to the connectivity and activity of the ports.
Example 9-4 FICON Director Report
1 z/OS V1R10 0 0 IODF = 07 CR-DATE: 06/19/2009 SWITCH DEVICE: 0062 SWITCH ID: PORT -CONNECTIONAVG FRAME ADDR UNIT ID PACING CU 8400 CU 8200 CU 8000 0D SWITCH ---0 0E CHP 86 0 0F CHP 54 0 10 CU ---0 11 CU ---0 12 CHP 87 0 13 CU ---0 14 CU ---0 15 CU ---0 16 CHP 8C 0 17 CHP 55 0 18 SWITCH ---0 19 CU ---0 1A CHP 8D 0 1B CHP 45 0 1C CU ---0 1D ------ ---0 F I C O N D I R E C T O R A C T I V I T Y PAGE SYSTEM ID SC30 DATE 06/23/2009 INTERVAL 10.00.037 RPT VERSION V1R10 RMF TIME 15.39.37 CYCLE 1.000 SECONDS CR-TIME: 10.00.10 ACT: ACTIVATE 62 TYPE: 006064 MODEL: 001 MAN: MCD PLANT: 01 SERIAL: 0000000119D2 AVG FRAME SIZE PORT BANDWIDTH (MB/SEC) ERROR READ WRITE -- READ -- -- WRITE -COUNT 1

56 72 388 796 683 71 94 973 747 72 397 106 455 68 0 957 0

56 279 965 329 459 107 40 284 397 449 934 76 686 171 0 111 0

0.00 0.00 0.05 0.02 0.02 0.00 0.00 0.04 0.14 0.00 0.11 0.00 0.07 0.00 0.00 0.10 0.00

0.00 0.00 0.17 0.00 0.01 0.00 0.00 0.01 0.07 0.01 0.27 0.00 0.17 0.00 0.00 0.01 0.00

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

210

FICON Planning and Implementation Guide

1E 1F 20 21

CHP CHP CHP SWITCH

8E 95 42 ----

0 0 0 0

72 0 0 56

169 0 0 58

0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00

0 0 0 0

Fields of significance are: AVG FRAME PACING - This is the average time (in microseconds) that a frame had to wait before it could be transmitted due to no buffer credits being available. AVG FRAME SIZE READ/WRITE - The average frame size (in bytes) used to transmit and receive data during this interval. PORT BANDWIDTH READ/WRITE - The rate (in MBps) of data transmitted and received during the interval. ERROR COUNT - The number of errors that were encountered during the interval.

9.5.5 ESS Link Statistics Report

This report should be used to analyze the external link usage and capacity planning of Peer-to-Peer-Remote-Copy (PPRC) links. For each adapter (SAID) of the ESS, this report contains statistics by link type. Example 9-5 reflects the activity on a selected adapter.
Example 9-5 ESS Link Statistics
1 z/OS V1R10 E S S L I N K S T A T I S T I C S PAGE SYSTEM ID SC31 DATE 05/13/2009 INTERVAL 10.00.237 RPT VERSION V1R10 RMF TIME 21.49.36 CYCLE 1.000 SECONDS 0SERIAL NUMBER 00000BALB1 TYPE-MODEL 002107-922 CDATE 05/13/2009 CTIME 21.49.37 CINT 10.00 0------ADAPTER-------LINK TYPE-BYTES BYTES OPERATIONS RESP TIME I/O SAID TYPE /SEC /OPERATION /SEC /OPERATION INTENSITY 00201 FIBRE 2Gb NO DATA TO REPORT OR ZERO 00202 FIBRE 2Gb ECKD READ 1.7M 5.4K 315.1 0.1 19.1 ECKD WRITE 142.4K 1.8K 78.5 0.0 3.7 -----22.8 00203 FIBRE 2Gb ECKD READ 899.4K 5.2K 174.0 0.1 9.8 ECKD WRITE 83.7K 2.3K 37.1 0.1 2.1 -----11.9 00230 FIBRE 2Gb ECKD READ 2.4M 618.8 3808.8 0.0 65.1 ECKD WRITE 661.5K 11.4K 58.1 0.1 7.6 -----72.7 00231 FIBRE 2Gb ECKD READ 2.1M 553.0 3799.4 0.0 63.7 ECKD WRITE 633.1K 12.9K 49.2 0.2 8.7 -----72.3 00232 FIBRE 2Gb NO DATA TO REPORT OR ZERO 00233 FIBRE 2Gb NO DATA TO REPORT OR ZERO 00240 FIBRE 2Gb ECKD READ 168.4K 9.6K 17.5 0.1 1.0 ECKD WRITE 857.2K 20.3K 42.3 0.2 6.7 -----7.6 00241 FIBRE 2Gb ECKD READ 45.4K 3.8K 12.0 0.0 0.4 ECKD WRITE 9.2K 1.8K 5.0 0.1 0.5 -----0.9 00242 FIBRE 4Gb ECKD READ 5.5K 2.6K 2.1 0.0 0.1 ECKD WRITE 655.4 1.9K 0.3 0.1 0.0 -----0.1 00243 Undefined NO DATA TO REPORT OR ZERO 00300 FIBRE 2Gb ECKD READ 2.6M 676.2 3843.2 0.0 67.8 ECKD WRITE 745.8K 10.2K 72.9 0.1 9.2 -----77.0 2

Chapter 9. Monitoring the FICON environment

211

The fields are: ADAPTER SAID reflects the System Adapter Identifier address. TYPE reflects the characteristics of the connection (2/4 Gb fiber, in our example). LINK TYPE will be one of the following: PPRC send and receive SCSI read and write BYTES/SEC reflects the average number of bytes transferred for all operations on this link. BYTES/OPERATION reflects the average number of bytes transferred for each individual operation. OPERATIONS/SEC reflects the average number of operations per second. RESP TIME/OPERATION reflects the average response time (in milliseconds). I/O INTENSITY is measured in milliseconds/second. It reflects the duration of the interval for which the adapter was active. A value of 1000 indicates that the link was busy for the entire time period. A value greater than 1000 would indicate that concurrent operations had been active.

9.5.6 General performance guidelines

The following guidelines are basic practices that can be used to improve the performance of your FICON devices. Be aware that their applicability may vary based on vendor specifications and model type.

DASD
Here we discuss basic practices for DASD.

High disconnect time

The following points reflect typical approaches to reduce disconnect time: Cache the volume, if it is a suitable cache candidate. Un-cache the volume, if it is a poor cache candidate. Tune cache. Used Dynamic Cache Management (DCME) to dynamically manage cache for you. If necessary, manually tune cache by turning off poor cache users. Review blocksize. Increase where applicable. Add paths (if not already 8-path). Use faster devices. Move or copy data sets to other volumes to reduce contention. Reduce arm movement via data set placement (seek analysis required).

High IOSQ time

The following points reflect typical approaches to reduce IOSQ time: Implement Parallel Access Volumes. Decrease disconnect, connect, and PEND times. Decreasing the other response time components will also decrease IOSQ time proportionally. Cache the volume, if it is a suitable cache candidate. Tune cache. Use Dynamic Cache Management (DCME) to manage cache for you. 212
FICON Planning and Implementation Guide

Use I/O priority queuing (IOQ=PRTY in IEAIPSxx). Move or copy data sets to other volumes to reduce contention. Run requests sequentially to reduce contention.

High connect time

The following points reflect typical approaches to reduce connect time: Cache the volume, if it is a suitable cache candidate. Use indexed VTOCs. In a cached environment, use faster channels. Use PDSEs if you suspect directory search time is a problem.

High PEND time

The following points reflect typical approaches to reduce PEND time: Change the mix of data on the volume to reduce contention. If you can, identify one data set that is contributing to most of the problem. This may be eligible to be moved to a custom volume or moved into storage. Check channel utilization, especially if not a small number of channels per controller. Exploit multiple allegiance if available on your DASD subsystem. Implement Parallel Access Volumes. Check CMR time and estimate CU host adapter utilization to see if CU is being overloaded. Low channel utilization with high pend times means you might have to add more CU host adapters.

9.5.7 Tape
The following list items reflect typical approaches to improving tape performance. These are basic recommendations. The significant developments in tape technology (such as the IBM Virtual Tape Server) will mitigate many of these points, and you should refer to the specific performance recommendations of your chosen solution for more guidance. Review blocksize; increase where applicable. Allocate data sets to DASD. Tape mount management (TMM) is a method that you can use to accomplish this, and thereby significantly reduce tape mount requirements. Increase host buffers for priority jobs. Increase buffers from the default of 5, to as high as 20 (but no more than 20). Use VIO for temporary data sets. Add channels (if not already one channel per control unit function). Reduce mount delay (for example, with Automated Cartridge Loader). Use enhanced-capacity cartridges and tape drives. Restructure for more parallelism (for example, to allow an application to use multiple drives).

9.6 DCFM performance monitor

With the DCFM performance monitor, you can analyze the data flow from end-to-end or for selected ports. The performance monitor must be configured and it consists of an End-To-End Monitor and a Real-Time Monitor.
Chapter 9. Monitoring the FICON environment

213

9.6.1 End-to-End Monitor

To set up an End-to-End Monitor, follow these steps: 1. Select Monitor Performance End-to-End Monitors from the DCFM menu bar. 2. In the Set End-to-End Monitors dialog, select on the left side of the window an initiator port (from the upper window) and a target port (from the lower window) as shown in Figure 9-5.

Figure 9-5 End-to-End Monitor setup

3. To create the monitor pair, click the arrow pointing to the right. 4. Click Apply to save the changes. 5. Now you can select the End-to-End monitor from the Monitored Pairs part of the window and select the Real-Time Graph or Historical Graph button. 6. You will get the realtime or historical data of the monitored device pair. You can also select more measurements from the Additional Measures drop-down menu. The End-to-End Monitor is now set up and can be used.

9.6.2 Real-Time Monitor

To set up and use a Real-Time Monitor for some ports, follow these steps: 1. Select Monitor Performance Real-Time Graph from the DCFM menu bar. 2. Select the port type from the Show drop-down menu to view the ports displayed at the left part of the window (for example, Device Ports or ISL Ports). 3. Select the port on the left and click the arrow pointing to the right to add the port to the monitored list, as shown in Figure 9-6 on page 215. 214

FICON Planning and Implementation Guide

Figure 9-6 Real-Time Monitor setup

4. Click OK to save the changes. 5. You will get the real-time data of the monitored devices. You can also select more measurements from the Additional Measures drop-down menu. The Real-Time Monitor is now set up and can be used.

Chapter 9. Monitoring the FICON environment

215

216

FICON Planning and Implementation Guide

10

Chapter 10.

Debugging FICON problems

This chapter describes the problem determination (PD) tools and techniques that you can use in a FICON environment. The following topics are discussed: Preparing for problem determination activities Environmental Record, Editing, and Printing program FICON Purge Path Extended Helpful z/OS commands Node descriptor DCFM logs DCFM data collection Common z/OS FICON error message - interpretation

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

217

10.1 Preparing for problem determination activities

The problem determination (PD) information provided here assumes that the environment adheres to the architecture specification and implementation levels for the System z servers, the FICON Director, and control units. When the procedures discussed in this section are implemented in a z/OS z environment, they can simplify some problem determination tasks.

10.1.1 Using the D M - Display system Matrix command

After your system has been IPLed, execute a D M command to create an entry in the z/OS system log that will function as a footprint to record the processor and I/O hardware configuration at IPL time. This command can be automatically performed during the IPL process; the command can be added to member IEACMDxx in SYS1.PARMLIB to automate this process. The output of this command can be used later during problem determination to help you understand why the status of a hardware resource is in a certain condition.

10.1.2 Creating a CONFIG member

A config-id member can be created by HCD from an existing IODF dataset. When a config-id member is created in the z/OS PARMLIB, the command D M=CONFIG(config-id) is used to show any deviation from the required operating configuration. The D M=CONFIG(config-id) command can be invoked by Operations on a periodic basis (once every eight hours, at a minimum), and at the operator shift changeover to determine and be aware of any deviations from the initial configuration status. The D M=CONFIG(config-id) command is also useful after any of the following events: I/O-related repair activity I/O-related physical change activity Changes to cabling FICON Director settings Add, remove, customize, or recustomize a CU or device Dynamic I/O re-configuration activity for either software or hardware changes During scheduled I/O-related change activity, more than one config-id may be needed to represent the I/O configuration at a particular time or interval. The D M=CONFIG(xx) command is used to call a member CONFIGxx in SYS1.PARMLIB. The CONFIGxx data set contains a predefined hardware configuration that should reflect all the resources critical to system operation. In response to the D M=CONFIG(xx) command, the system compares the contents of the CONFIGxx member and the current configuration. It then displays the differences (if they exist) to the operator, who can take the corrective actions. Example 10-1 on page 219 shows a CONFIGxx data set member.

218

FICON Planning and Implementation Guide

Example 10-1 SYS1.PARMLIB(CONFIGxx)

CPU(0),ONLINE CPU(1),ONLINE CPU(2),ONLINE CHP(00-3F),ONLINE STOR(E=0),ONLINE STOR(E=1),ONLINE STOR(E=2),ONLINE DEV(0200-02FF),ONLINE DEV(100O-2FFF),ONLINE DEV(5100-6FFF),ONLINE DEV(9A00-BF00),ONLINE ... Example 10-2 shows the output of a D M=CONFIG(xx) command where deviations from a desired configuration (S1) exist. If the current running configuration matches the definitions in the CONFIGxx SYS1.PARMLIB member, the output of the z/OS command would contain the message NO DEVIATION FROM REQUESTED CONFIGURATION.
Example 10-2 D M=CONFIG(xx) output

D M=CONFIG(S1) IEE097I 11.38.13 DEVIATION STATUS 568 FROM CONFIGS1 DEVICE DESIRED ACTUAL D2A6 ONLINE OFFLINE D080 BASE HIPERPAV E200 OFFLINE ONLINE F890 PAV UNBOUND

10.1.3 Most common I/O-related problems

Most common problems are due to changes made to the configuration. Therefore, it is very important to be aware of all planned and unplanned changes. Examples of situations that have caused problems are listed here.

Problems since last Power-on Reset (POR), Activate, or IPL

Wrong HMC CPC reset profile used for the CPC for the Activate. This will cause the wrong IOCDS to be selected and used. Wrong IOCDS associated to the Reset Profile used to activate. I/O configuration definition changes performed that do not match the physical configuration. Wrong LOAD profile used, or wrong LOAD device or IODF device specified in the LOAD profile. The wrong LOAD device used for the IPL may cause the wrong OSCONFIG to be used. (The OSCONFIG member contains the z/OS software I/O configuration definition.) The wrong IODF device (volume) used may point to a different I/O configuration.

Chapter 10. Debugging FICON problems

219

Problems since the last physical configuration change

Physical cabling changes at System z Server, FICON Director, patch panels, DWDMs or Control Unit sides. Addition or removal of FICON Directors or control units.

Problems since the last control unit customization changes

Problems since the last FICON Director changes. FICON Director matrix changes (prohibit/allow). FICON Director customization changes: Fabric, FICON Director, or Port bindings. FICON Director port status changed: ports being blocked or taken offline. FICON Director zoning customization changes. FICON Director port cabling changes.

10.2 Problem determination approach for FICON

Before you initiate any problem determination tasks, you need to have a clear understanding of the problem or condition. Begin by taking note of IOS messages that are issued (see 10.10, Common z/OS FICON error message - interpretation on page 230 for examples). IOS messages provide useful information and are an excellent starting point for problem determination. Determining the status of all components in the affected I/O path is also key; techniques to use for this task are shown in Appendix F, Useful z/OS commands on page 287 and in Appendix E, Using HMC and SE for problem determination information on page 265. You also need current and accurate information and documentation to resolve problem situations, so try to have the information listed here available. The symptom or z/OS message associated with the problem that needs to be addressed The device number or device numbers of the problem device or devices The CHPID numbers of the problem channels or the channel paths to the device or devices The CPC and logical partition names that the z/OS is running on A configuration diagram of the complete I/O connectivity (all paths and FICON Director connections for a multi-path device) to the problem device The physical location of all involved hardware equipment and their respective consoles, when applicable For a FICON CTC-related problem: Target CPC Target Channel Subsystem Target Image (LPAR) Target matching device number (or the source/target CTC device pair) z/OS offers an extensive list of commands that will provide most of the required PD information listed. The HMC and SE panels include specific Problem Determination functions which may also be required to use when investigating FICON channel problems. Both the z/OS commands and the HMC/SE problem determination facilities are discussed in Appendix F, Useful z/OS commands on page 287 and Appendix E, Using HMC and SE for problem determination information on page 265. For FICON channel problem determination, a significant amount of information is available in various areas. The following sections provide an overview of what can be used based on four

220

FICON Planning and Implementation Guide

levels of tools. Figure 10-1 illustrates the different levels of software and hardware, as well as the tools provided from a verification perspective.
z/OS commands for status z/OS system log and LOGREC EREP and Purge Path Extended HMC and SE provide:
CHPID status CU and device status Node descriptor information SAD display

z/OS level
Problem Determination Approach

System z Server Hardware level

FICON Director

FICON Director tools provide:

Port status Node descriptor information logs

FICON Control Unit

Figure 10-1 Sources of information

Storage tools provide:

Port status Node descriptor information logs

Most common FICON failures

There are two common failures seen on FICON channels: Failure to complete initialization or establish a logical path Failure detected in the communications of command and data by either the channel or the control unit Initialization failures may or may not create a Hardware Message on the HMC/SE. The SE Problem Determination panels, shown under Displaying the Analyze Serial Link Status PD panel on page 281, will clearly show initialization failures on a per-FICON channel basis. There is also a panel available under the Service logon user ID called IFCC and Other Errors that will display information related to initialization types of failures. Channel Control Check (CCC) errors, initialization failures, Channel Detected Errors and Interface Control Checks (IFCC) can be caused by failures in FICON hardware or in the FICON definition setup.

FICON hardware-related failures

The FICON channel card can cause initialization IFCC and CCC failures. The FICON I/O (CU) Host Bay Adapter (HBA) can cause initialization and IFCC failures. The FICON Director can cause initialization and IFCC failures. Fibers in general can cause initialization and IFCC failures.

FICON definition-related failures

IOCP definition errors can cause initialization failures. Fabric definitions can cause initialization and IFCC failures. Control Unit definition errors can cause initialization and IFCC failures.

Chapter 10. Debugging FICON problems

221

FICON zoning-related failures

Zoning issues in fabrics can cause problems that are very difficult to diagnose. If zoning has been implemented, the zoning tables in the fabric must be updated whenever a new WWPN is installed and connected to a zoned fabric. Note: A common zoning problem occurs when swapping or changing FICON channel cards and the channels are defined to a switch that has all the FCP, FICON and Storage devices defined to their respective zones.

10.3 Environmental Record, Editing, and Printing program

When an error occurs, the system records information about the error in the LOGREC data set or the LOGREC log stream. This information provides you with a history of all hardware failures, selected software errors, and selected system conditions. Use the Environmental Record, Editing, and Printing program (EREP) to print reports about the system records, to determine the history of the system, or to learn about a particular error. For further information about this topic, refer to z/OS MVS Diagnosis: Tools and Service Aids, GA22-7589.

10.4 FICON link incident reporting

To improve the ability to capture data for link error analysis, z/OS can now register to receive FICON link incident records, which are referred to as Link Maintenance Information (LMI) records. z/OS will display this information on consoles, as well as save it in the system log and in LOGREC. A link incident record consists of detailed node information that is gathered for particular link-degraded and link-failure incidents. The node descriptor information is gathered during the Fabric Login or Port Login when the logical path is established. For additional information pertaining to LMI records, refer to z/OS MVS Diagnosis: Reference, GA22-7588.

10.5 FICON Purge Path Extended

When the extended subchannel logout facility is installed and enabled, all Fibre Channel (FC) channel paths can create extended subchannel logouts, under applicable error conditions, and generate I/O interruptions that indicate such a logout is pending retrieval. If a logout is pending for an I/O interruption, IOS will retrieve the logout data and include the data in the subchannel logout record that it records in SYS1.LOGREC. Note: FICON Purge Path Extended must be supported by the control unit and FICON Director to receive error counters. Check with your product vendor for details about this topic. With FICON Purge Path Extended (PPE) enabled, Link Error Statistical Buffer (LESB) information is collected from each Fibre Channel port in the path from host to control unit. This information then results in the abort reason code being built. The channel subsystem

222

FICON Planning and Implementation Guide

passes this information to z/OS as Extended Subchannel Logout Data (ESLD), and an SLH record is cut in the LOGREC medium. The overall process is shown in Figure 10-2.

SYS1.LOGREC SLH

I/O interrupt IRB ESLD

z/OS
FICON - FC
LESBs abort reason ESLD model-dependent Channel data Channel Abort reason code Channel port LESB Channel attached Port LESB

FICON
LESBs abort reason

Director
CUattached Port LESB data CU port LESB CU Abort error code CU model dependent data

FICON - CU

LESBs abort reason

Figure 10-2 FICON Purge Path Extended error recording

This function provides advanced problem determination of Fibre Channel bit errors. A subset of a sample Subchannel Logout Handler (SLH) record, formatted by EREP, is included in Figure 10-3 on page 224.

Chapter 10. Debugging FICON problems

223

DEVICE NUMBER: DEVICE TYPE:

0C07A 3390

REPORT: SCP:

SLH EDIT VS 2 REL.

DAY YEAR DATE: 275 08 HH MM SS.TH TIME: 00 17 17.30

JOB

CPU MODEL: 2084XA CHANNEL PATH ID: 57 LOGICAL CPU ID: 01796A PHYSICAL CHAN ID: 0131 PHYSICAL CPU ADDRESS: 31 CC CA FL CT FAILING CCW 00 00000000 00 0000 K FLAGS CA US SS CT 04 824017 00000008 00 02 0000 SUB-CHANNEL STATUS

VOLUME SERIAL SUBCHANNEL ID NUMBER ERROR TYPE

SCSW

---UNIT STATUS----

ATTENTION 0 PGM-CTLD IRPT 0 STATUS MODIFIER 0 INCORRECT LENGTH 0 CONTROL UNIT END 0 PROGRAM CHECK 0 BUSY 0 PROTECTION CHECK 0 CHANNEL END 0 CHAN DATA CHECK 0 DEVICE END 0 CHAN CTL CHECK 0 UNIT CHECK 0 I/F CTL CHECK 1 UNIT EXCEPTION 0 CHAINING CHECK 0 ----SOFTWARE RECOVERY STATUS----HARD FAIL 0 DEGRADE FAIL 0 SOFT FAIL 1 PASSED 0 CHANNEL ERROR ANALYSIS IRB STORED BY INTERRUPT TERMINATION BY -- SELECTIVE RESET SEQ CODE *** INVALID *** VALIDITY OF RECORDED DATA COUNT INVALID TERMINATION CODE VALID SEQUENCE CODE INVALID DEVICE STATUS INVALID CCW ADDRESS INVALID DEVICE NUMBER VALID SENSE DATA NOT STORED CHANNEL LOGOUT DATA 0000 00000000 00000000 00000000 0020 00000001 00000000 000023C7 0040 00000000 00613613 00006106 0060 01600CE6 50050764 00C1796A 0080 2200002A 00000000 02000000 00A0 00200100 30303231 30353830 00C0 00000000 00000000 00000000 00E0 00000000 00000000 00000000 CONTROL UNIT LOGOUT DATA 0000 00000000 00000000 0020 00000000 00000000 0040 00000000 00000000 0060 00000000 00000000 0080 00000000 00000000 00A0 00000000 00000000

--------------------------SCSW FLAGS--FLAG 0 FLAG 1 CCW FORMAT 1 RESERVED 0 PRE-FETCH CCW 0 SSCH FUNCTION 1 INIT STATUS 0 HSCH FUNCTION 0 ADDR LIMIT 0 CSCH FUNCTION 0 SUPP SUSPEND INT 0 RESUME PENDING 0 ZERO COND CODE 0 START PENDING 0 EXTENDED CONTROL 1 HALT PENDING 0 PATH NOT OPER 0 CLEAR PENDING 0

--

CODE

00000000 00000000 10000800 00190002 18100020 3049424D 00000000 00000000

00000000 0C000000 88A082F4 00000000 50050763 31333030 00000000 00000000

00000000 9000AB00 10000800 00000000 00CB945C 30303030 00000000 00000000

00000002 80000000 88A082F4 0800003E 50050763 30323232 00000000 00000000

00000000 80000000 50050764 00000000 00C0945C 31320024 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

Figure 10-3 Formatted Subchannel Logout Handler record

Figure 10-4 on page 225 shows the CHANNEL and CONTROL UNIT LOGOUT DATA fields extracted from the EREP SLH record shown in Figure 10-3. Some fields are interpreted to give you an idea of how this information can be broken down.

224

FICON Planning and Implementation Guide

CHANNEL LOGOUT DATA 0000 00000000 00000000 0020 00000001 00000000 0040 00000000 00613613 0060 01600CE6 50050764 0080 2200002A 00000000 00A0 00200100 30303231 00C0 00000000 00000000 00E0 00000000 00000000 CONTROL UNIT LOGOUT DATA 0000 00000000 00000000 0020 00000000 00000000 0040 00000000 00000000 0060 00000000 00000000 0080 00000000 00000000 00A0 00000000 00000000
Entry port on the LESB for the Director FICON Director entry port. Hex count of a specific error type

00000000 000023C7 00006106 00C1796A 02000000 30353830 00000000 00000000

00000000 00000000 10000800 00190002 18100020 3049424D 00000000 00000000

00000000 0C000000 88A082F4 00000000 50050763 31333030 00000000 00000000

00000000 9000AB00 10000800 00000000 00CB945C 30303030 00000000 00000000

00000002 80000000 88A082F4 0800003E 50050763 30323232 00000000 00000000

00000000 80000000 50050764 00000000 00C0945C 31320024 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

00000000 00000000 00000000 00000000 00000000 00000000

Channel F-Port LESB logout The Director entry port (Error counters are shown)

Error Code = 0C Indicates an ABORT was received from the Channel

Channel N-Port LESB logout (clean in this example)

Control Unit LESB counters (clean in this example)

D C
FICON CHPID FICON Control Unit N-Port

F-Port

F-Port

N-Port

C
Device C07A

Figure 10-4 CHANNEL and CONTROL UNIT LOGOUT details

For a detailed description regarding the debugging of these records, refer to the ANSI standard architecture, which can be found at:
http://www.t11.org

10.6 Helpful z/OS commands

The display results of the following commands, if issued from the affected logical partition (LPAR), can be very helpful for documenting the configuration and analyzing the errors in a FICON environment. Therefore, before you begin to perform problem determination that involves a FICON Director, use these commands. Also, obtain a copy of the z/OS syslog. D M=CHP displays the status and type of CHPIDs D M=CHP(cc) where cc is the affected CHPID D U,,,dddd,1 where dddd is the affected device address DS P,dddd where dddd is the affected device address D M=DEV(dddd) where dddd is the affected device address D M=DEV(dddd,(cc)) where dddd is the affected device address and cc is the failing CHPID D M=SWITCH(xx) displays the ports configured and in use by the specified FICON Director number.

Chapter 10. Debugging FICON problems

225

Examples of the z/OS commands can be found in Appendix F, Useful z/OS commands on page 287. Example 10-3 shows the output of a z/OS D M=SWITCH(xx) command.
Example 10-3 Output of the D M=SWITCH(65) z/OS command

D M=SWITCH(65) IEE174I 12.45.15 DISPLAY M 215 SWITCH 0065, PORT STATUS 0 1 2 3 4 5 6 7 8 9 A B C D E F 0 u u u c u u u u u u u u u u u u 1 u u u u u u u u u u u c u u u u 2 u u u u u u u u u u u u u x u u 3 u u u u x u u u u u u u u u u u 4 . . . . . . . . . . . . . . . . 5 . . . . . . . . . . . . . . . . 6 . . . . . . . . . . . . . . . . 7 . . . . . . . . . . . . . . . . 8 . . . . . . . . . . . . . . . . 9 . . . . . . . . . . . . . . . . A . . . . . . . . . . . . . . . . B . . . . . . . . . . . . . . . . C . . . . . . . . . . . . . . . . D . . . . . . . . . . . . . . . . E . . . . . . . . . . . . . . . . F . . . . . . . . . . . . . . . . ***************** SYMBOL EXPLANATION ***************** + DCM ALLOWED - DCM NOT ALLOWED BY OPERATOR x NOT DCM ELIGIBLE p DCM NOT ALLOWED DUE TO PORT STATE c CHANNEL ATTACHED $ UNABLE TO DETERMINE CURRENT ATTACHMENT u NOT ATTACHED . DOES NOT EXIST The D M=SWITCH(xx) command provides important information about the configured and in-use ports at the specified switch, along with their current state. Notice in the example that Switch 65, ports 03 and 1B are in use and attached to a FICON Channel. Ports 2D and 34 display an x state; these ports are used as ISL ports.

10.7 Node descriptor

The node descriptor will provide you with useful information such as the manufacturer, node type, node status, serial number, Tag information, WWNN, and WWPN. For an explanation about how to obtain this information from a System z server, refer to Displaying the Analyze Channel Information panels on page 272.

10.7.1 View node descriptors from DCFM

If not already done, you can change the view for FICON in the DCFM by clicking SAN Options. In the Option window, select Display and click the Set Up FICON Display button. Click OK to close the window.

226

FICON Planning and Implementation Guide

You will see the Node Descriptor details in the DCFM for each port at the left hand side of the window. To change the port view, right-click a Director (on the left of the window) and select Port Display. Check mark the details you want to see. You can also see the Node Descriptor for a Director by right-clicking and selecting Element Manager Hardware. In the Directors Element Manager, click Name Server at the left side of the window. This will open the Name Server window displaying all connected devices, as shown in Figure 10-5.

Figure 10-5 Node Descriptor overview

Clicking the Accessible Devices button displays all devices that have access to the selected list entry (that is, each device that is in the same Zone as the selected list entry). Selecting the list entry and clicking Detail View produces a detailed view for the selected device, as shown in Figure 10-6 on page 228.

Chapter 10. Debugging FICON problems

227

Figure 10-6 Node descriptor detailed view

The Tag field for the System z server is a four-digit number, and it consists of two parts. The last two digits are the CHPID and the first digit gives you information about the Channel Subsystem (CSS) in which the CHPID is defined. In our example, the Tag is a07e, which means that the CHPID is 7E and is defined in CSS 0 and 2 (see Table 10-1).
Table 10-1 CSS mapping First digit 1 2 3 4 5 CSS 3 2 2 and 3 1 1 and 3 First digit 6 7 8 9 A CSS 1 and 2 1,2 and 3 0 0 and 3 0 and 2 First digit B C D E F CSS 0,2 and 3 0 and 1 0,1 and 3 0,1 and 2 0,1,2 and 3

The Tag field details for a DS8000 are shown in DS8000 port layout on page 262.

10.8 DCFM logs

The DCFM has several logs that include information which will help you during problem determination. The Master Log is displayed in the lower part of the DCFM and it includes, by default, all events. You can also set a filter by clicking Filter and selecting the events you want show in the Master Log. If you want a specific log, select Monitor Logs, as shown in Figure 10-7 on page 229.

228

FICON Planning and Implementation Guide

Figure 10-7 Selecting a Log

You can display the log from a selected Director or logical Switch by right-clicking the Director and selecting Events.

10.9 DCFM data collection

If you need to request service, you may be asked to collect data. To collect FICON-specific data, you must configure it for each Director. This is done via the command-line interface at the Director. One possibility is as follows: 1. In the DCFM server or client, right-click the Director and select Telnet. A new window will open. (You can also use an ssh connection to the Director, if needed.) 2. Enter the user name (in our case, admin) and press Enter. 3. Enter the password (the default is password) and press Enter. 4. Enter, for example, setcontext 1 if you need to connect to a logical switch with FID 1. 5. Enter the command supportshowcfgenable ficon and press Enter. To check the settings, enter supportshowcfgshow. 6. Enter logout to close the window. To perform data collection, select Monitor Technical Support Collect Data in the DCFM menu bar. The Technical Support Data window appears, as shown in Figure 10-8 on page 230.

Chapter 10. Debugging FICON problems

229

Figure 10-8 Data collection window

Select the Director or logical switch at the left side of the window and click the arrow pointing to the right, to put it in the list of selected switches for data collection. Click OK to start the data collection. Read the message at the confirmation window and then click OK. After data collection is complete (which may take up to 30 minutes, depending on the size of your configuration), you can select the zip file by going to Monitor Technical Support View Repository. Select the correct zip file by looking at the date and time stamp included in the file name (if the size is 0 KB, it means that data collection is not finished yet and you need to wait a few minutes longer). Click the zip file and save it to your hard disk. If you are at the DCFM Server, you can also go into the following directory to get the data: C:\Program Files\DCFM 10.1.3\data\ftproot\technicalsupport\. Data collection is finished now and can be sent to the Support Center.

10.10 Common z/OS FICON error message - interpretation

This section describes the most common z/OS FICON messages (IOS) and their respective meanings, explanations, and required actions when applicable. For specific message details and actions, refer to z/OS MVS System Messages, SA22-7637. In 10.1.3, Most common I/O-related problems on page 219, the two major types of problems are discussed: Failure to complete initialization or establish a logical path Failure detected in communications of command and data by either the channel or the control unit This section explains the messages you may receive when encountering these problems. 230
FICON Planning and Implementation Guide

Note: The messages covered here may change, depending on the z/OS release and version. APARs may also implement architectural changes regarding how messages are generated and displayed. We recommend that readers refer to the current z/OS Messages library for reviewing the output whenever a discrepancy is noted. This section provides insight to the most common FICON problem reporting messages.

IOS001E dddd, INOPERATIVE PATH pp<pp..>

This message indicates that one or more paths went inoperative (indicated by the PNOM UCW field) which were previously operational (Path Operational Mask, POM was ON). For more information about Path Not Operational Mask, PNOM and POM fields, refer to Displaying the Analyze Subchannel Information PD panel on page 275. The message text shows: dddd pp This was the device in use by an I/O operation when the failure occurred. This was the path in use by the device that become inoperative.

Tip: Check for a disabled Control Unit interface, an altered FICON director configuration, or a disabled channel path. Refer to the explanations of messages IOS2001I or IOS2002I in this section for possible reasons why you might receive message IOS001E.

IOS002A dddd, NO PATHS AVAILABLE

This message indicates that all paths to the device were INOPERATIVE (DCC=3, Deferred Condition Code 3), which was possibly caused by power drop of the CU or FICON Director. dddd represents the device in use. The I/O requester will be kept waiting until: Path connection is physically restored (a V dddd,ONLINE command may be required). The device is BOXed by a V dddd,OFFLINE,FORCE command. On all paths that present Deferred Condition Code 3 (DCC=3), Dynamic Pathing Switch Validation returns RC=12, to indicate to the caller that the no paths bit is to be turned on in the device UCB and wait for the CU to become operational, generating a State-Change Ready interrupt. Tip: Power up or Enable the Control Unit (or FICON Director) and then issue a VARY ONLINE command, or enter V dddd,OFFLINE,FORCE to box the device.

IOS050I CHANNEL DETECTED ERROR ON dddd,pp,cmd,stat <PCHID=xxxx>

This message indicates that a Channel Error (Interface Control Check - IFCC, Channel Control Check - CCC, Channel Data Check - CDC, and so on) was detected. The message text shows: dddd This was the device in use by an I/O operation when the failure occurred.

Chapter 10. Debugging FICON problems

231

pp cmd stat PCHID: xxxx

This is the channel path identifier (CHPID), if known. Otherwise, this field will be set to asterisks (*). This is the failing command being executed by the device, if known. Otherwise, this field will be set to asterisks (*). This is the device and subchannel status, if known. Otherwise, this field will be set to asterisks (*). This is the System z server Physical Channel IDentifier.

Tip: Run EREP, or examine the SE CHPID work area for the IFCC Threshold Exceeded CHPID status. Alternatively, use the IFCC and Other Errors service task on the SE to further investigate possible reasons why this message was issued.

IOS051I INTERFACE TIMEOUT DETECTED ON dddd,chp,cmd,stat <PCHID=xxxx>

This message indicates that the channel subsystem detected an interface timeout IFCC. The message text shows: dddd pp cmd stat PCHID: xxxx This was the device in use by an I/O operation when the failure occurred. This is the channel path identifier (CHPID), if known. Otherwise, this field will be set to asterisks (*). This is the failing command being executed by the device, if known. Otherwise, this field will be set to asterisks (*). This is the device and subchannel status, if known. Otherwise, this field will be set to asterisks (*). This is the System z server Physical Channel IDentifier.

Tip: The IOS will check the state of the timeout bit (the timeout indication is reported by the channel in the Extended Status Word Timeout (ESWT) before generating the IOS051I message for FICON channels. If timeout is not on, then IOS050I will be issued instead. Note that IOS050I and IOS051I are the most common z/OS FICON error messages. Various enhancements have been made via APARs OW47845 and OW57038 to provide more information about the possible error root cause. Both messages should be addressed at the hardware level.

IOS052I CHANNEL EVENT REPORTED FOR dddd,pp,op,stat <PCHID=xxxx>

This message indicates that the Channel detected an error while the system was operating a device. A Channel card hardware error has occurred. (Cross check is the most common Channel card hardware error). The message text shows: dddd pp This was the device in use by an I/O operation when the failure occurred. This is the channel path identifier (CHPID), if known. Otherwise, this field will be set to asterisks (*).

232

FICON Planning and Implementation Guide

cmd stat PCHID: xxxx

This is the failing command being executed by the device, if known. Otherwise, this field will be set to asterisks (*). This is the device and subchannel status, if known. Otherwise, this field will be set to asterisks (*). This is the System z server Physical Channel IDentifier.

Tip: This message indicates that the channel error has been recovered by the channel and it is logged only to SYSLOG (not to consoles), and will be logged in SYS1.LOGREC. (No operator console message). The message appears if the channel error has occurred with Extended Subchannel Logout Data pending and the Logout-only bit is ON.

IOS071I dddd,pp,jobname, START PENDING

This message indicates that the Missing Interrupt Handler (MIH) detected one of these conditions: dddd,pp,jobname, MISSING (CHANNEL AND) DEVICE END dddd,pp,jobname, IDLE WITH WORK QUEUED dddd,pp,jobname, HALT SUBCHANNEL INTERRUPT MISSING dddd,pp,jobname, CLEAR SUBCHANNEL INTERRUPT MISSING Tips: The jobname can be *MASTER*. In this case, IOS most likely is recovering I/O such as Unconditional Reserve (U/R) or Dynamic Path Selection (DPS) and the operation is generating the MIH condition. If Missing CE/DE for CTC devices, probably the system on the other end was not properly running (in a loop or wait condition; very busy with I/O interrupts being disabled; or in a thrashing state for paging I/Os). Look at the RMF report for the paging I/O rate. An IDLE WITH WORK QUEUED message may occur if an Immediate Execution (IMEX) Request is queued while a sense operation is pending for a device. This occurs because the IMEX request cancels the pending sense operation so it can execute first. However, IOS will not be able to start the IMEX request because the sense I/O is pending. Thus, MIH will find this condition and declare it as an IDLE WITH WORK QUEUED condition. For Start-Pending, Channel Path pp may not be correct. It is just the previous successfully selected path. To identify the correct path, refer to the EREP MIH record. Refer to APAR OW50355, which corrects many MIH-related problems, for further guidance. For example, if the default device class MIH is 00.00, the device gets MIH for IOS-issued I/O after 1 second (typically, CTC devices).

IOS075E dddd, RECURRING MIH CONDITION FOR THIS DEVICE

Chapter 10. Debugging FICON problems

233

This message indicates that the Recursive MIH condition was not cleared by the previous MIH Clear Subchannel (CSCH) recovery sequence. It is possible that the original MIH message may have been lost. Tips: If you specify an MIH HALT interval that is too short (such as 1 second), change it to at least 2 seconds because the interval of the FICON Interface-Timeout IFCC (No Response from CU) is 2 seconds. CSCH was rescheduled due to the recursive MIH. This means that the recursive MIH condition was detected before the original MIH recovery completed. As a result, the original MIH message may not have been issued and this message was issued instead. The difference from message IOS077E is that this message is issued (instead of IOS077E) when the MIH message type is unknown (whether IOTIMING, Primary Status Missing, Secondary Status Missing, Start Pending, Halt Interrupt Missing, or Clear Interrupt Missing), as indicated by the CONDITON flag being OFF in the MIH message module. CSCH may not complete quickly if the CSS-selected CHPID (based on the LPUM of the subchannel) is in a permanent busy condition.

IOS076E dddd,ch,jobname, CLEAR SUBCHANNEL INTERRUPT MISSING

This message indicates that one of these conditions was detected: dddd,pp,jobname, MISSING CHANNEL AND DEVICE END dddd,pp,jobname, MISSING DEVICE END dddd,**,jobname, START PENDING dddd,**,jobname, IDLE WITH WORK QUEUED dddd,pp,jobname, HALT SUBCHANNEL INTERRUPT MISSING The message indicates missing interrupt conditions where the MIH recovery did not reset the device (that is, it did not issue the CSCH); for example: The MIH was for CLEAR Subchannel interrupt missing. Or the TAPE MIH exit requested no CSCH to be issued when the current active I/O contains a long-running CCW, such as Forward-Space File and so on. For a DASD device, there is no case where this occurs. Tip: This message is issued when there is an active I/O operation that was not cleared by the CSCH recovery action sequence that was performed.

IOS077E dddd,ch,jobname, CLEAR SUBCHANNEL INTERRUPT MISSING

This message indicates that one of these conditions was detected: dddd,pp,jobname, MISSING CHANNEL AND DEVICE END dddd,pp,jobname, MISSING DEVICE END dddd,**,jobname, START PENDING

234

FICON Planning and Implementation Guide

dddd,**,jobname, IDLE WITH WORK QUEUED dddd,pp,jobname, HALT SUBCHANNEL INTERRUPT MISSING Tip: This message means that another missing interrupt condition (MIH) was detected before the previous MIH recovery completed.

IOS100I DEVICE dddd BOXED, LAST PATH pp LOST (text)

This message indicates that one of these conditions was detected, where (text) is one of the following: RESERVE LOST CANNOT RE-RESERVE STILL RESERVED RESERVE MAY BE LOST or MAY BE RESERVED ASSIGN LOST CANNOT RE-ASSIGN STILL ASSIGNED SELF-DESCRIPTION RECOVERY Tip: In the Channel Path recovery, the last path (CHPID) to the device has been varied offline. The device has been Boxed. Any new I/O request to this device will result in Permanent I/O Errors. If other channel paths are available to the device, vary them online. This condition is generally associated to a hardware failure in the CU or device.

IOS101I DEVICE dddd FORCED OFFLINE {BOXED} (text)

This message indicates that one of these conditions was detected, where: dddd is the device and (text) is one of the following: RESERVE LOST CANNOT RE-RESERVE RESERVED ON FAILING PATH pp RESERVE MAY BE LOST MAY BE RESERVED ON PATH pp ASSIGN LOST CANNOT RE-ASSIGN WAS RESERVED WAS ASSIGNED STILL ASSIGNED STILL RESERVED ASSIGNED IN FAILING PATH pp

Chapter 10. Debugging FICON problems

235

Tip: The Device Recovery routine invoked in the Channel Path Recovery detected one of these conditions as it tried to Re-Reserve/Assign the device after RCHP - Reset Channel Path (System Reset). Any new I/O request to the device will result in a Permanent I/O error. This condition is generally associated to a hardware failure in the CU or device. Note that this condition could result from BOX_LP=(xxx) being specified for HOTIO or TERMINAL in the IECIOSxx PARMLIB member. BOX_LP=(ALL) is the default from z/OS 1.4. (with ASSIGN or RESERVE LOST text).

IOS102I DEVICE dddd BOXED, (text)

This message indicates that one of these conditions was detected, where: dddd is the device and (text) can be one of the following; OPERATOR REQUEST (via the V dddd,OFFLINE,FORCE command) PERMANENT ERROR (for example, IOS291I on all paths, and so on) WAS RESERVED (BUT IT HAS BEEN RELEASED NOW) STILL RESERVED - MAY BE RESERVED - WAS ASSIGNED STILL ASSIGNED - NO PATHS - RESERVE LOST C.U.I.R. REQUEST - ALIAS UNBOUND DEFERRED BOX PROCESSING COMPLETE The device had its BOX processing deferred and this message is issued to indicate that deferred BOX processing is now complete and the device is now in the boxed state. See IOS071I on page 233IOS107I for more information. This message itself may not show the reason of the device BOX. It is usually accompanied by another IOS message (for example, IOS451I or IOS109E and so on) that may contain the reason for the device being boxed. Also, be aware that if these messages are issued within the same time window (milliseconds), then the SYSLOG message order may sometimes be out of sequence (reversed). Tip: A permanent error on all Paths, or Reserve/Assign Lost in DPS Array, was detected during DPS Validation with message IOS451I.

IOS107I DEVICE dddd BOX PROCESSING DEFFERRED

Where: dddd is the device. This message indicates that the system deferred the device dddd box processing to allow recovery processing to run. Tip: Certain recovery processing, such as HyperSwap operations, may be able to provide continuous availability to the device by swapping the faulty device (that is, the primary device) with a fully operational backup device (that is, the secondary device) prior to permanently boxing the faulty device. At the end of this process, an IOS102I message will indicate that the boxing process has completed.

236

FICON Planning and Implementation Guide

IOS162A CHPID pp ALERT, UNSOLICITED MALFUNCTION INTERRUPT

Where: pp is the path. While trying to service a request from a device, the channel subsystem found an error before the system determined the device number. (Channel Report Word received - Channel Path, Temporary Error). Check EREP.

IOS163A CHPID pp ALERT, NO ASSOCIATED SUBCHANNEL FOR DEVICE

Where: pp is the path ID.

IOS2001I dddd, INOPERATIVE PATHS

Where: dddd is the device and pp is the path. This message appears only in the SYSLOG. When such a condition is present, the D M=DEV will report it via the text messages presented. Example 10-4 shows how the message text appears in the D M=DEV(dddd,(pp)) output. Note that we entered the subject path in the DM command.
Example 10-4 D M=DEV with information about inoperative paths

D M=DEV(D003,(7E)) IEE174I 18.00.58 DISPLAY M 907 DEVICE D003 STATUS=ONLINE CHP 7E ENTRY LINK ADDRESS 6503 DEST LINK ADDRESS 6688 PATH ONLINE Y CHP PHYSICALLY ONLINE Y PATH OPERATIONAL Y MANAGED N CU NUMBER D000 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107.900.IBM.75.0000000BALB1.0300 SCP TOKEN NED = 002107.900.IBM.75.0000000BALB1.0000 SCP DEVICE NED = 002107.900.IBM.75.0000000BALB1.0003 CHANNEL LINK LEVEL FACILITY DETECTED A LINK ERROR (90) LINK RCVY THRESHOLD EXCEEDED FOR ALL LOGICAL PATHS OF DEST LINK (60) HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW, ZHPF When message MSGIOS001E, MSGIOS002A, or MSGIOS450E is issued, IOS will retrieve the not-operational reason from the Channel Subsystem and issue message MSGIOS2001I (for MSGIOS001E or MSGIOS450E) or MSGIOS2002I (for MSGIOS002A), to SYSLOG only, with text describing the reason. Tip: The message displays the STATUS FOR PATH(S) chp, chp,... not-operational reason text. The second line of text, if it exists, contains the specific reason for the not-operative path (as shown in Example 10-4 on page 237).

Chapter 10. Debugging FICON problems

237

The paths specified were found to be not operational for the specified device. Some FICON-relevant Status Code message texts are listed here: 50: CHANNEL SUBSYSTEM DETECTED A LINK FAILURE CONDITION) 10: LOSS OF SIGNAL OR SYNCHRONIZATION CONDITION RECOGNIZED 20: NOT OPERATIONAL SEQUENCE RECOGNIZED 30: SEQUENCE TIMEOUT RECOGNIZED 40: ILLEGAL SEQUENCE RECOGNIZED 60: CHANNEL LINK LEVEL FACILITY IN OFFLINE RECEPTION Note: This is the case when the Channel-to-CU logical path (H/W link-level) has been de-established (broken). 70: PORT REJECT WAS ENCOUNTERED 10: ADDRESS INVALID ERROR 11: UNDEFINED DESTINATION ADDRESS ERROR 12: DESTINATION PORT MALFUNCTION 13: DYNAMIC SWITCH PORT INTERVENTION REQUIRED 80: LINK LEVEL REJECT WAS ENCOUNTERED 01: TRANSMISSION ERROR 05: DESTINATION ADDRESS INVALID ERROR 07: RESERVED FIELD ERROR 08: UNRECOGNIZED LINK CONTROL FUNCTION 09: PROTOCOL ERROR 0A: ACQUIRE LINK ADDRESS ERROR 0B: UNRECOGNIZED DEVICE LEVEL 90: CHANNEL LINK LEVEL FACILITY DETECTED A LINK ERROR 10: CONNECTION ERROR 20: TRANSMISSION ERROR 30: PROTOCOL ERROR 40: DESTINATION ADDRESS INVALID ERROR 50: DEVICE LEVEL ERROR (50) 60: LINK RCVY THRESHOLD EXCEEDED FOR ALL LOGICAL PATHS OF DEST LINKS. Note: For System z server FICON channels, this is the case where the CU logical path connection has been fenced because the up/down (flapping) error count has been exceeded. (Channel Path or Link level threshold exceeded). You must enter the V PATH (dddd,pp),ONLINE command to issue a Reset Link Recovery Threshold CHSC command, or Config CHPID OFF then ON (using a CF z/OS command or via the HMC). A0: LOGICAL PATH IS REMOVED OR NOT ESTABLISHED 01: PACING PARAMETERS ERROR 02: NO RESOURCES AVAILABLE 04: DESIGNATED CONTROL UNIT IMAGE DOES NOT EXIST

238

FICON Planning and Implementation Guide

 05: LOGICAL PATH PRECLUDED BY CONFIGURATION AT CONTROL UNIT IMAGE 06: LINK RECOVERY THRESHOLD EXCEEDED FOR LOGICAL PATH Note: For System z servers FICON channels, this is the case where CU logical path connection has been fenced due to up/down (flapping) error count exceeded. (Channel Path or Link level threshold exceeded). You must enter the V PATH (dddd,pp),ONLINE command to issue a Reset Link Recovery Threshold CHSC command, or Config CHPID OFF then ON (using a CF z/OS command or via the HMC). B0: IN PROCESS OF INITIALIZING PATH 10: CONTROL UNIT DEVICE LEVEL INITIALIZATION IS NOT COMPLETE 20: LINK BUSY CONDITION LAST ENCOUNTERED 30: PORT BUSY CONDITION LAST ENCOUNTERED 30: PORT BUSY CONDITION LAST ENCOUNTERED 30: PORT BUSY CONDITION LAST ENCOUNTERED FF: NO FURTHER INFORMATION AVAILABLE OR UNKNOWN CONDITION

C0: SUBCHANNEL PATH AVAILABLE, BUT DEVICE NOT OPERATIONAL

IOS2002I dddd, NO PATHS AVAILABLE

Where: dddd is the device and pp is the path. This message appears only in the SYSLOG. Tips: All paths to the specified device were found not operational. Path status code is created by Path State Info routine issuing Store Subchannel Path Information (SSPI). Note the following examples: FICON cables were disconnected at the CU side through FICON SW: STATUS FOR PATH(S) pp IN PROCESS OF INITIALIZING PATH (B0) NO FURTHER INFORMATION AVAILABLE OR UNKNOWN CONDITION (FF) FICON Director Power-Down (by power button): CHANNEL LINK LEVEL FACILITY IN OFFLINE RECEPTION STATE (60)

IOS202E CHANNEL PATH pp FORCED OFFLINE. DEVICE IS: dddd / UNKNOWN

Where: dddd is the device and pp is the path. The CHPID has been configured OFFLINE from the Hardware Management Console (HMC), or it was forced offline when Channel Path Recovery (RCHP) failed to clear the path in an error condition.

IOS203I CHANNEL PATH pp SUCCESSFULY RECOVERED. DEVICE IS: dddd / UNKNOWN

Where: dddd is the device and pp is the path. Channel Path Recovery was successfully completed. Device Address 'dddd' is the trigger for Channel Path Recovery.

Chapter 10. Debugging FICON problems

239

This is caused by the following conditions: CRW w/ Channel Path, Terminal condition (I/F Hung) Reset-Event-Notification (Path Verification) Hot I/O & Action specified for its message is CHPK

IOS288A SYSTEM-INITIATED OFFLINE | ONLINE RECONFIGURATION IS IN PROGRESS FOR THE FOLLWING CHPIDS: CC, CC-CC,...
Where: cc is the channel path. System z servers allow the use of the Hardware Management Consoles (HMCs) to Configure CHPIDs On and Off without stealing the channel from the configuration. This facility, called System Initiated CHPID Reconfiguration, uses a Support Element (SE) component to request a config off or on of all the CSS.CHPIDs associated with a particular channel ID. Normally, in a repair action, the operator has to take the channels offline using the z/OS commands to configure off or on all of the associated CHPIDs in all partition. With this facility, z/OS handles the SE requests configuring all the CSS.CHPIDs, with the exception of the last path to a device.

IOS289I SYSTEM-INITIATED OFFLINE | ONLINE RECONFIGURATION HAS COMPLETED.

This message reports the results of reconfiguration actions requested and reported by IOS288A. The results can be: RECONFIGURATION WAS SUCCESSFUL FOR THE FOLLOWING CHPIDS: cc,cc-cc RECONFIGURATION FAILED FOR THE FOLLOWING CHPIDS: cc,cc-cc.

240

FICON Planning and Implementation Guide

Part 5

Part

Appendixes

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

241

242

FICON Planning and Implementation Guide

Appendix A.

Example: planning workflow

This appendix shows an example of a workflow using the steps described in Chapter 4, Planning the FICON environment on page 59.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

243

Step 1 - Documentation
Creating and maintaining documentation throughout all the planning, design, and implementation phases is very important. Throughout this book we provide information regarding the various features and functions offered by System z servers and the FICON Directors (and under which conditions they should be applied). That information can be used as a starting point for your documentation. For documentation of our implementation steps for a cascaded FICON Director environment see Chapter 7, Configuring a cascaded FICON Director topology on page 133, and Chapter 8, Configuring FICON Directors on page 151. Configuration worksheets for your FICON Director and FICON CTC environments can be found in Appendix B, Configuration worksheets on page 251.

Step 2 - Requirements
We based our requirements on the need to access non-business-critical data in a remote location from two z/OS LPARs in the main location. Isolating the new configuration from the existing one is essential to the solution. High availability is not a key requirement. Based on our physical and logical inventory, these were the components involved in the solution: An IBM System z10 Enterprise Class in the main location Two z/OS V1R10 LPARs Two FICON Express8 channels An IBM System Storage DS8000 in the remote location Two storage control units and two devices Two 4 Gbps ports that will be upgraded to 8 Gbps in the near future Two IBM System Storage SAN b-type family (FICON Directors) SAN768B with four 8 Gbps ports (in the main location) SAN384B with four 8 Gbps ports (in the remote location) Figure A-1 displays the components used to create our new FICON environment.

z10 EC
SCZP201
LPAR A23 z/OS V1R10
FICON Express8 PCHID 5A2

System Storage

SAN384B
Port 88 Port 8D Port C4 Port CF

DS8000

Port 0003

LPAR A24 z/OS V1R10

FICON Express8 PCHID 5D2

Port 03

Port 2D

Port 34

Port 0242
Port 1B

SAN768B
Figure A-1 Results of the inventory effort

244

FICON Planning and Implementation Guide

Step 3 - Context
Taking into consideration the existing and planned components (System z, FICON Director, and storage control device), we determined that the scenarios 4.4.2, Moving to a high bandwidth environment (FICON Express8) on page 65 and 4.4.3, Migrating from a single site to a multi-site environment on page 66, best fit our requirements. In Figure A-2, ports 2D and 34 in SAN768B and ports 8D and C4 in SAN384B will be used for ISLs. All ports will be running at date link rate of 8 Gbps, except for the ports in the DS8000; they will run at a date link rate of 4 Gbps. Ports 0003 and 0242 in the DS8000 will be upgraded at a later point.

z10 EC
SCZP201
LPAR A23 z/OS V1R10
FICON Express8 PCHID 5A2

DS8000
Port 03 Port 2D

Port 1B

Port 34

SAN384B
Port 8D Port C4 Port 88 Port CF

Port 0003

LPAR A24 z/OS V1R10

FICON Express8 PCHID 5D2

SAN768B

Port 0242

Figure A-2 Our FICON components in the two locations

Step 4 - Topologies and supported distances

We implemented a cascaded FICON Director topology (see 4.5.3, Cascaded FICON Directors on page 68 for more information). The two locations were less than the maximum distance of 10 km apart, therefore a WDM environment was not required. For more information about supported distances, refer to Table 4-1 on page 67. The transceiver type required for the ISLs in a cascaded FICON Director topology is long wavelength (LX). LX transceivers are generally recommended for all connections, which we used for all our ports; see Figure A-3.

z10 EC DS8000
FICON Express8 LX

LX

LX

SAN384B
FICON Express8 LX

LX

LX

LX LX LX

LX LX

SAN768B

LX

Figure A-3 Our cascaded FICON Director topology

Appendix A. Example: planning workflow

245

Step 5 - Convergence
In our FICON environment we did not set up an intermix environment. However, the FICON Directors do offer this possibility, if needed. Because we were using existing environments, other relevant factors for the FICON fabric such as power consumption, cooling, and space inside the data center were not a concern. Our FICON environment was connected to an isolated network within a secured area, therefore we only used username and password authentication and authorization. However, we highly recommend that you change all the default passwords for all default accounts. To isolate the new configuration from the existing configuration, we secured the fabric using zoning and the fabric configuration server with strict SCC policies.

Step 6 - Management
For management purposes, we used the Data Center Fabric Manager (DCFM) for the entire FICON environment including setup and monitoring, as illustrated in Figure A-4.

Figure A-4 FICON management via DCFM

Our FICON Director management environment is shown in Figure A-5. The figure also shows the IP addresses that we used for our FICON Directors, DCFM server, and firewall/router.
Data Center Fabric Manager Server/Client Data Center Fabric Manager Client

System Storage SAN768B

10.77.77.61 10.1.1.20

SW 65 CP0 CP1
10.1.1.21 10.1.1.31 10.1.1.22 10.1.1.32 10.1.1.10

CP0 CP1 SW 66
10.77.77.62 10.1.1.30

10.1.1.1

Firewall

Corporate Network

System Storage SAN384B

Figure A-5 Our FICON Director management environment

246

FICON Planning and Implementation Guide

We also used the CUP and SA for I/O to manage and control System z FICON connectivity.

Step 7 - Virtualization
In System z, we had two z/OS LPARs. Although the two z/OS LPARs use the same CSS, each FICON channels will be defined to an additional CSS (spanned) for future use. The channels were on two different FICON features. We defined redundant paths from both z/OS LPARs to the storage control unit.

z10 EC
SCZP201
LPAR SC30 A23 z/OS V1R10
FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E

System Storage SAN384B

Switch # 66 Switch @ 66
Port 8D Port C4 Port 88 Port CF

DS8000
CU# D000
LX 0003

D0xx

LPAR SC31 A24 z/OS V1R10

FICON Express8 LX CSS2 PCHID 5D2 CHPID 82

Port 03

Port 2D

Port 34

Switch # 65 Switch @ 65

LX 0242

D1xx

CU# D100
Port 1B

SAN768B
Figure A-6 Virtualization

For the DS8000, we set up two ports for redundant access to all devices. We had two logical control units inside the DS8000, and we defined the DS8000 channels to different FICON features. Our FICON Directors were, by default, set up with virtualization support and we used this technology to define port addresses beyond the FICON port addressing range. We created a new logical switch to be able address the 48-port line cards. We also assigned our ports to two different cards for redundancy. The two virtual FICON Directors were interconnected with ISLs. The following two figures show the worksheets that were used to configure the FICON connectivity for the z10 EC, the two FICON Directors, and the DS8000. Figure A-7 on page 248 displays the configuration worksheet for FICON Director 1.

Appendix A. Example: planning workflow

247

FICON Director Configuration Worksheet

FICON Director Manufacturer: IBM Type: 2499 Model: 384 S/N: 02-8D014

HCD Defined Switch ID x65 FICON Director Domain ID x65

FICON Director Ports
Slot Number Port Number Port Address Laser Type: LX / SX

(Switch ID) (Switch @)

Cascaded Directors No Yes x Corresponding Cascaded Director Domain ID x66 Fabric Name Cascaded Fabric 1

Attached N_Ports / E_Ports (CU, CPC, or ISL)

Port Name Node Type
CU/CHNL/FD

Machine Type

Model

Serial Number

ISL CU I/F CPC CHPID

1 2 3 4

3 11 13 4

6503 651B 652D 6534

LX LX LX LX

Z10 EC PCHID 5A2 Z10 EC PCHID 5D2 ISL1 to SAN384B ISL2 to SAN384B

CHNL CHNL FD FD

2097 2097 2499 2499

E26 E26 192 192

02-1DE50 02-1DE50 02-2C00Z 02-2C00Z

CPC CHPID CPC CHPID ISL ISL

Figure A-7 Configuration worksheet for FICON Director 1

Figure A-8 displays the configuration worksheet for FICON Director 2.

FICON Director Configuration Worksheet

FICON Director Manufacturer: IBM Type: 2499 Model: 192 S/N: 02-2C00Z

HCD Defined Switch ID x66 FICON Director Domain ID x66

FICON Director Ports
Slot Number Port Number Port Address Laser Type: LX / SX

(Switch ID) (Switch @)

Cascaded Directors No Yes x Corresponding Cascaded Director Domain ID x65 Fabric Name Cascaded Fabric 1

Attached N_Ports / E_Ports (CU, CPC, or ISL)

Port Name Node Type
CU/CHNL/FD

Machine Type

Model

Serial Number

ISL CU I/F CPC CHPID

7 8 7 8

8 15 13 4

6688 66CF 668D 66C4

LX LX LX LX

DS8K 10003 DS8K 10242 ISL1 to SAN768B ISL2 to SAN768B

CU CU FD FD

2107 2107 2499 2499

E26 E26 384 384

75-BALB1 75-BALB1 02-8D014 02-8D014

CU I/F CU I/F ISL ISL

Figure A-8 Configuration worksheet for the FICON Director 2

For high integrity, we set up Insistent Domain IDs and strict SCC policy. We used one port-based zone for all our FICON channels.

Step 8 - Performance
The short distance between the two locations allowed us to keep the default buffer credit assignment in the FICON Directors. For high performance, we used the zHPF feature in both the System z server and the DS8000. For better performance of the fabric links, we used lossless DLS with frame-based trunking. We also used port fencing to ensure that no loss of performance occurs due to excessive errors on the interfaces, and to make certain that the fabric remains stable.

248

FICON Planning and Implementation Guide

To monitor and evaluate the utilization of the paths and the traffic patterns, we used RMF on our System z server.

Step 9 - Prerequisites and interoperability

We ensured that all System z, DS8000, and FICON Director prerequisites and interoperability requirements are fulfilled, for example: System z10 Feature code for FICON Express8 LX FC3325 z/OS 1.7 with IBM Lifecycle Extension for z/OS V1.7 (5673-A01) DS8000 Feature code for zHPF FC0709 and FC7092 SAN768B and SAN384B Feature code for the FICON CUP management FC7886 See references in 4.10, Prerequisites and interoperability on page 92.

Step 10 - Physical connectivity

Our setup was based on LX FICON transceivers with standard LC duplex connectors, therefore we used 9 m single-mode fiber optic cables (OS2 fiber type); see Figure A-9.

z10 EC
FICON Express8 LX PCHID 5A2 CHPID 7E

System Storage SAN384B

LX 88 LX 8D LX C4 LX CF

DS8000
SM

9 m

9 m

SM
LX 03

LX 0003
m 9 SM

m 9 SM
LX 2D LX 34

FICON Express8 LX PCHID 5D2 CHPID 82

SM
LX 1B

LX 0242

SAN768B
Figure A-9 Physical connectivity

All cable connectors are LC Duplex type

For the cabling infrastructure we used IBM Fiber transport service products with single mode fiber.

Conclusion
After all these steps were completed, our final FICON design appeared as shown in Figure A-10 on page 250.

Appendix A. Example: planning workflow

249

z10 EC
SCZP201
LPAR SC30 A23 z/OS V1R10
FICON Express8 LX CSS2 PCHID 5A2 CHPID 7E

SAN384B
Switch # 66 Switch @ 66
Port 8D Port C4 Port 88 Port CF

DS8000
9 m SM

CU# D000
LX 0003

9 m

SM

D0xx

m 9 SM
Port 2D

L IS

L IS

9 m SM

LPAR SC31 A24 z/OS V1R10

FICON Express8 LX CSS2 PCHID 5D2 CHPID 82

Port 03

Port 34

9 m

Switch # 65

LX 0242

D1xx

SM

Switch @ 65

CU# D100
Port 1B * All cable connectors are LC Duplex type

SAN768B
Figure A-10 Our final FICON design

250

FICON Planning and Implementation Guide

Appendix B.

Configuration worksheets
This appendix contains the following worksheets: FICON Director Configuration Worksheet Use this worksheet to document the layout of your FICON Director. It can be applied as a tool to help you understand how the ports are allocated for configuration and problem determination purposes. FICON CTC Image-ID Worksheet Use this worksheet as an aid in planning and documenting the association between LPARs, MIF IDs, Logical Channel subsystems, and CTC Image-IDs.

Copyright IBM Corp. 2005, 2006, 2009, 2006. All rights reserved.

251

252

FICON Director Configuration Worksheet

Model: ____ S/N: _____________

FICON Director Manufacturer: ______________ Type: ______

Cascaded Directors No ___ Yes ___ Corresponding Cascaded Director Domain ID _______ Fabric Name ________________________________

HCD Defined Switch ID _____ (Switch ID) FICON Director Domain ID _____ (Switch @)

FICON Director Ports

Port Address Port Name Model Node Type CU / CHNL Machine Type Laser Type: LX / SX

Attached N_Ports / E_Ports (CU, CPC, or ISL)

Serial Number ISL CU I/F CPC CHPID

FICON Planning and Implementation Guide

Slot Number

Port Number

Server: Type:
LPAR Name CSS MIF ID CTC-Image ID

Server: Type:
LPAR Name CSS MIF ID CTC-Image ID

Appendix B. Configuration worksheets

253

254

FICON Planning and Implementation Guide

Appendix C.

Configuration and definition tools

This appendix provides an overview of tools that you can use to define the path between a System z server and a DASD control unit and a device. The tools are: Hardware Configuration Definition (HCD) Hardware Configuration Manager (HCM) CHPID Mapping Tool (CMT) I/O Configuration Program (IOCP) HCD is a program that runs in a z/OS image. HCM and CMT are PC-based programs. IOCP is a stand-alone program running in a partition on a System z server. HCD, HCM, CMT, and IOCP are briefly described in this appendix. Also included are descriptions of statements and keywords that are used to define a FICON switched Point-to-Point topology. Unlike other platforms, with System z all paths to an I/O device must be defined (there is no auto-detection or auto-configuration). The characteristics of the channel paths, the control units, and the devices have to be specified and saved in an I/O Configuration Data Set (IOCDS), which is used by a System z server to load the configuration into its storage. The IOCDS file is stored on the Support Element. Although HCD and HCM can create an IOCDS for each physical CPC present in a sysplex environment, IOCP can create an IOCDS only for the physical system it is running on. The IOCDS contains statements and keywords which specify the characteristics of a channel, control unit, and device. These statements and keywords are described in IOCDS statements and keywords used for FICON on page 257. Refer to IBM System z10 Enterprise Class Configuration Setup, SG24-7571, for additional information about how to set up the System z server.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

255

Hardware Configuration Definition

The Hardware Configuration Definition (HCD) supplies an interactive dialog to generate your I/O definition file (IODF) and subsequently your Input/Output Configuration Data Set (IOCDS). It is strongly recommended that you use HCD to generate your IODF, as opposed to writing your own IOCP statements. The validation checking that HCD performs as you enter data helps to eliminate errors before you implement your I/O configuration. Refer to HCD Users Guide, SC33-7988, for detailed information about HCD.

Hardware Configuration Manager

The z/OS Hardware Configuration Manager (HCM) is a PC-based client/server interface to HCD that combines the logical and physical aspects of hardware configuration management. In addition to the logical connections, you can also manage the physical aspects of a configuration. For example, you can effectively manage the flexibility offered by the FICON infrastructure. All updates to your configuration are done via the HCM intuitive graphical user interface and, most importantly, all changes of the logical I/O configuration are written into the IODF and fully validated and checked for accuracy and completeness by HCD. Refer to HCM Users Guide, SC33-7989, for detailed information about HCM.

CHPID Mapping Tool

The CHPID Mapping Tool (CMT) provides a mechanism for customizing the CHPID assignments for System z servers, to avoid connecting critical paths to single points of failure. Additional enhancements have been built into the CMT to help with the requirements of z9 and z10 servers, thereby providing a mechanism to map CHPIDS to PCHIDS, and the best availability recommendations for installed features and defined configurations. More details and informations about the CHPID Mapping Tool are available on IBM Resource Link at the following URL: https://www.ibm.com/resourcelink The CMT program and the CMT users guide can be downloaded to a workstation from IBM Resource Link.

I/O Configuration Program

The I/O Configuration Program (IOCP) is a stand-alone program running in partition on a System z server. It resides on the support element of a System z server and is loaded into an active partition when the Input/output (I/O) Configuration icon is clicked in the support element workplace. The stand-alone IOCP can be used to manage and modify the input/output (I/O) configuration source files and data sets for the selected central processor complex (CPC) only. Refer to IOCP Users Guide, SB10-7037, for detailed information about IOCP.

256

FICON Planning and Implementation Guide

IOCDS statements and keywords used for FICON

This section describes the required keyword definitions and rules for a FICON configuration. Refer to IOCP Users Guide, SB10-7037, for a detailed description of statements and keywords. The latest version of this publication is available on IBM Resource Link: https://www.ibm.com/resourcelink

ID statement
The ID statement is used to specify identification information of the server and IODF used to build the IOCDS. In addition to other keywords allowed in the ID statement, only keywords used later in this chapter are described here. SYSTEM The SYSTEM keyword specifies the machine limits and rules that are used for verification of input data set. The device type of the server using the IOCDS has to be specified in the ID keyword. The TOK keyword provides information about the source file used to build the IOCDS.

RESOURCE statement
The RESOURCE statement provides information about logical partitions. The partition names and their MIF IDs and associated CSSs are defined here. PART The PART keyword allows you to specify an LPAR name and to assign the MIF ID to a partition. Optionally, a CSS number can be assigned to the logical partition.

CHPID statement
The CHIPD statement contain keywords that define the characteristics of a channel. Note that not all keywords allowed in the CHPID statement are described here; only keywords used to define a FICON channel (FC) are listed. PATH The PATH parameter allows you to specify the CSSs and the CHPID number that will be assigned to a channel path. The PCHID keyword will be used to identify the location of the physical channel in the processor. The CHPID type for FICON channels in a point-to-point configuration is required and must be specified as TYPE=FC. This indicates that the FICON channel will operate in FICON native (FC) mode. The Switch keyword is required, and it specifies an arbitrary number for the FICON Director to which the channel path is assigned. Note that the switch number defined in the IOCDS is not the same as the switch address that is set up in the FICON Director. The switch address is the Domain ID specified in the FICON Director. It is recommended that you have the same value for the switch number and the switch address.

TYPE

SWITCH

CNTLUNIT statement
The CNTLUNIT statement contains keywords that define the characteristics of a control unit. Note that not all the keywords allowed in the CNTLUNIT statement are described here; only keywords used to define a control unit connected to a FICON channel (FC) are listed. PATH LINK The PATH keyword specifies the channel path in each channel subsystem (CSS) attached to the control unit. The LINK keyword is required to define the destination port address where the control unit is connected to the FICON Director. The link address can be

Appendix C. Configuration and definition tools

257

specified with one or two bytes. Specifying a two-byte link address requires a special feature (High Integrity Fabric) and setup in the FICON Director. CUADD For FICON native (FC) channel paths, the logical address is specified as two hexadecimal digits in the range 00FE. Not all FICON control units support logical addressing. To determine a products logical addressing information, contact your System Service Representative supporting the FICON control units. UNITADD The UNITADD keyword specifies the unit addresses of the I/O devices that the control unit recognizes.

IODEVICE statement
The IODEVICE statement specifies the characteristics of an I/O device and defines to which CU the devices is attached. CUNUMBR The CUNUMBR keyword specifies the control unit number to which the device is attached.

258

FICON Planning and Implementation Guide

Appendix D.

Configuring the DS8000 for FICON

This appendix describes how to configure the I/O ports on a DS8000 to communicate with a z10 EC server using the native FICON protocol. The configuration steps described are applicable to point-to-point, switched point-to-point, and cascaded FICON Director topologies.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

259

Getting started
The LX ports 0003 and 0242 in our DS8000 are used to connect to either z10 FICON Express8 ports or to FICON Director ports. The following steps describe how to set up the ports for FICON use: 1. Start a Web browser on a workstation that has network connectivity to the DS8000 HMC. 2. Type the address of the DS8000 in the Web browser address field, for example http://xxx.xxx.xxx.xxx:8451/DS8000/Login (where xxx.xxx.xxx.xxx is the IP address of the DS8000 HMC), and then press Enter. 3. When the logon panel is displayed, type in the user ID and password. The welcome window will appear as shown in Figure D-1.

Click Storage Images

Figure D-1 Welcome to DS8000

4. Click Storage images to continue. The Storage images panel is displayed next (see Figure D-2 on page 261).

260

FICON Planning and Implementation Guide

Select Storage Image

Click on Configure I/O Ports

Figure D-2 DS8000 Storage Images

5. On the Storage Images window click the check box on the left of the image name to select the Storage Image. 6. In the Select Action pull-down menu, click Configure I/O Ports. The Configure I/O Ports panel will be displayed as shown in Figure D-3.

Select Change to FICON Select Port

Figure D-3 DS8000 Configure I/O Ports

7. Click the check box of the port number to select the port you want to change. 8. In the Select Action pull-down menu, click Change to FICON. This will specify the port to use native FICON protocol for communication. 9. Complete these steps for all ports desired for native FICON protocol.

Appendix D. Configuring the DS8000 for FICON

261

10.Finally, check that all ports using native FICON protocol are specified correctly. See Figure D-4 for an example of the Configure I/O Ports panel (port 0003 is set up and online).

FICON is selected for port 0003

Figure D-4 DS8000 I/O ports configured for FICON

For our configuration we need to have I/O ports 0003 and 0242 set up for FICON in the controller.

DS8000 port layout

The DS8000 has two port layouts: a physical port layout and a logical port layout. The physical port layout is written in the following way: R1-I1-C1-T0. A DS8000 can have two racks with I/O drawers installed. Rack 1 (R1) is the rack with the controllers and the HMC in it. The rest of the notation (shown in Table D-1 on page 263) is the same for rack 2 (R2). A rack can have four I/O drawers (I1-I4), and it consists of six card slots (C1-C6). Each card can have four ports (T0-T3).

262

FICON Planning and Implementation Guide

Table D-1 Physical port numbering (rear view) I1 T0 T1 T2 T3 C1 T0 T1 T2 T3 C1 T0 T1 T2 T3 C2 T0 T1 T2 T3 C2 T0 T1 T2 T3 C3 T0 T1 T2 T3 C3 I3 R I O R I O T0 T1 T2 T3 C4 T0 T1 T2 T3 C4 T0 T1 T2 T3 C5 T0 T1 T2 T3 C5 T0 T1 T2 T3 C6 T0 T1 T2 T3 C6 T0 T1 T2 T3 C1 T0 T1 T2 T3 C1 T0 T1 T2 T3 C2 T0 T1 T2 T3 C2 T0 T1 T2 T3 C3 T0 T1 T2 T3 C3 I4 R I O R I O I2 T0 T1 T2 T3 C4 T0 T1 T2 T3 C4 T0 T1 T2 T3 C5 T0 T1 T2 T3 C5 T0 T1 T2 T3 C6 T0 T1 T2 T3 C6

The logical port numbering is written as I0000 (it can also be written without the I in front). Table D-2 shows the last three digits (I0xxx) of the logical port layout for rack 1.
Table D-2 Logical port numbering for rack1 (rear view) I1 000 001 002 003 C1 200 201 202 203 C1 010 011 012 013 C2 210 211 212 213 C2 020 021 022 023 C3 220 221 222 223 C3 I3 R I O R I O 030 031 032 033 C4 230 231 232 233 C4 040 041 042 043 C5 240 241 242 243 C5 050 051 052 053 C6 250 251 252 253 C6 100 101 102 103 C1 300 301 302 303 C1 110 111 112 113 C2 310 311 312 313 C2 120 121 122 123 C3 320 321 322 323 C3 I4 R I O R I O I2 130 131 132 133 C4 330 331 332 333 C4 140 141 142 143 C5 340 341 342 342 C5 150 151 152 153 C6 350 351 352 353 C6

Table D-3 on page 264 shows the last three digits of the logical port layout for rack 2.

Appendix D. Configuring the DS8000 for FICON

263

Table D-3 Logical port numbering for rack 2 (rear view) I1 400 401 402 403 C1 600 601 602 603 C1 410 411 412 413 C2 610 611 612 613 C2 420 421 422 423 C3 620 621 622 623 C3 I3 R I O R I O 430 431 432 433 C4 630 631 632 633 C4 440 441 442 443 C5 640 641 642 643 C5 450 451 452 453 C6 650 651 652 653 C6 500 501 502 503 C1 700 701 702 703 C1 510 511 512 513 C2 710 711 712 713 C2 520 521 522 523 C3 720 721 722 723 C3 I4 R I O R I O I2 530 531 532 533 C4 730 731 732 733 C4 540 541 542 543 C5 740 741 742 742 C5 550 551 552 553 C6 750 751 752 753 C6

264

FICON Planning and Implementation Guide

Appendix E.

Using HMC and SE for problem determination information

This appendix explains how you can use the HMC and SE panels to obtain information when performing problem determination or verifying the FICON environment.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

265

HMC and SE versions and user IDs

The HMC and SE provide System z and FICON channel path information from a hardware point of view. All panels we discuss here are based on the following: The hardware being used for window captures is a z10 EC Hardware Management Console (HMC) and z10 EC Support Element (SE), both running at version 2.10.1 which corresponds to Driver level 76D. Users are logged on in SYSPROG mode (User role) using the CLASSIC User Interface (UI) style. A System z server can also be set to a User Interface style called Tree-Style User Interface (UI). This UI style provides the exact same information, but looks quite different. For details about this UI style or about how to change the UI style, refer to HMC Operations Guide, SC28-6873. The panels may differ if you are looking at another processor generation or LIC level. However, the information and data you can gather are the same.

Logging on to the HMC and SE

To log on to the HCM and SE, do the following: 1. Log on at the HMC using SYSPROG (user ID) 2. Double-click Groups. 3. Double-click Defined CPCs. 4. Highlight the CPC that you want to work with. 5. Double-click the Single Object Operations (SOO) icon in the CPC Recovery task list.

Displaying System z server CPC details

The CPC details work area can be opened by performing the following steps: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266 for more information about this topic). 2. On the SE workplace, double-click Groups. 3. Double-click CPC. 4. Right-click CPC and select CPC Details. Figure E-1 on page 267 shows the CPC status of a CPC called SCZP101. The panel has four selectable tabs. We discuss the contents of the Instance Information and Product Information tabs.

266

FICON Planning and Implementation Guide

Figure E-1 CPC Details - Instance Information panel

As highlighted in the CPC Instance Information panel, the following information can be retrieved (the values shown reflect our scenario): The IOCDS slot used to Activate the CPC: A2 The IOCDS name used during Activation: IODF00 The Activation Profile associated with the CEC: SCZP201 The Activation Profile last used to Activate the CEC: SCZP201 Figure E-2 on page 268 shows CPC product information.

Appendix E. Using HMC and SE for problem determination information

267

Figure E-2 CPC Details - CPC Product Information

As highlighted in the CPC Product Information panel shown in Figure E-2, the following information can be retrieved from this panel: The machine type and hardware model: 2097 / E28 The machine serial number: 02-001DE50 The Model-Capacity Identifier: 714 The Model-Temporary Capacity Identifier:714 The Model-Permanent Capacity Identifier: 714

Finding a physical resource on a System z server

The following steps explain how to find a physical resource such as a FICON channel in the System z server. We discuss two SE panels: The View Frame Layout panel shows the location of the installed I/O cages where the FICON channels are normally installed. The Channel to PCHID Assignment panel shows important information about the PCHIDs and CHPIDs and the adaptor types and their physical locations.

Displaying the System z server View Frame Layout panel

Follow these steps to display the View Frame Layout panel on the SE: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266). 2. On the SE workplace, double-click Groups. 3. Double-click CPC. 4. Click the CPC icon and select View Frame Layout from the CPC Configuration task.

268

FICON Planning and Implementation Guide

Note: The HMC has a similar panel under CPC configuration tasks that allows the Frame Layout panel to be viewed.

Displaying the View Frame Layout panel

Figure E-3 shows the View Frame Layout SE panel. Clicking over the presented cages causes a details view to display. Selecting this option provides additional information about the frames and cages, as shown in the details.

Figure E-3 SE View Frame Layout panel

Displaying the Channel to PCHID Assignment panel

Follow these steps to display the Channel to PCHID Assignment panel on the SE: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266). 2. On the SE workplace, double-click Groups. 3. Double-click CPC. 4. Click the CPC icon and select Channel PCHID Assignment from the CPC Configuration task. The PCHID Assignment panel appears, as shown in Figure E-4 on page 270.

Appendix E. Using HMC and SE for problem determination information

269

Figure E-4 Channel PCHID Assignment

In addition to showing search and different view modes, this panel displays information about the channels in use such as: Channel Location Physical location of a channel including Cage, Card Slot and Jack (card port) Book and Fanout Provides information about what fanout is used to access the channel, its book location and port Channel State The current state of the channel, such as Online and Offline PCHID - the channels corresponding PCHID number CSS.CHPID - the CHPID and CSSs that this channel belongs to Card Type - the channels corresponding hardware

Displaying individual FICON channel information

Follow these steps to display detailed FICON channel information: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266). 2. On the SE workplace, double-click Groups. 3. Double-click CPC if you know the PCHID number associated. Otherwise, double-click Images and select the LPAR you are working from. 4. Right-click the CPC and select Channels. Because you selected CPC, this displays the PCHIDs. If you had selected Images, the panel displays the CHPIDS associated to the selected partition. 5. Select the PCHID or CHPID of interest and double-click over the icon to display the PCHID or CHPID details.

270

FICON Planning and Implementation Guide

Figure E-5 shows the PCHID Details panel for PCHID 05A2. Note that there is a panel called Status in the field. The status reflects the current state of the PCHID or CHPID.

Figure E-5 PCHID 05A2 Details panel

This list shows the possible state or status that a particular PCHID/CHPID might be in: Stopped Operating Stand-by Stand-by/Reserved IFCC Threshold Exceeded This state indicates that the channel is defined, but not in use. This is the normal indication of a working CHPID. This indicates that the CHPID has been taken offline. This means that the channel has been put in service mode for repair or test. This is a channel status condition that will appear when the IFCC events have exceeded a coded preset value. For FICON channels, the threshold established limit number is 4. For each IFCC detected, this value will be decremented by the code. When a value of 0 (zero) is reached, this will be indicated in the CHPID status/state. The threshold value is shown in the Analyze Channel Status SE PD panel. This condition is caused by the defined device not matching the attached device. The channel is a FICON channel but it is not compatible with the channel type defined in the IOCDS.

Definition Error

Note: When a definition error is detected, the CHPID icon will default to the icon that represents a FICON Converter CHPID, and it is not necessarily representative of the defined channel type. Sequence Time Out A FICON sequence is a FICON frame or a group of FICON frames transmitted between the Channel and the CU or between the CU and the channel. The number of frames that make up a

Appendix E. Using HMC and SE for problem determination information

271

sequence is determined by specific fields inside the frames themselves. Sequences are logically grouped to form an exchange. An exchange represents an I/O operation in the FICON implementation. A Sequence Time Out is detected by the channel or the CU when an ongoing exchange does not detect the arrival of a required sequence. Sequence Not Permitted This error is reported in the Analyze Channel information panel as Illegal Sequence. It reports that the FICON channel-attached control unit or FICON Director port does not support the presented FICON service parameters.

Displaying the Analyze Channel Information panels

Follow these steps to open the Analyze Channel Information (PD) panels: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266). 2. On the SE workplace, double-click Groups. 3. Double-click CPC if you know the PCHID number associated. Otherwise, double-click Images and select the LPAR you are working from. 4. Right-click the CPC and select Channels. Because you selected CPC, this displays the PCHIDs. If you had selected Image, the panel will display the CHPIDS associated to the selected partition. 5. Highlight the channel that you want to view. 6. Double-click Channel Problem Determination located in the CHPID Operations task list. Note: The SE will post a panel for an Image query if you select a PCHID instead of a CHPID. 7. Select Analyze Channel Information and click OK. The Analyze Channel Information panel is shown in Figure E-6 on page 273.

272

FICON Planning and Implementation Guide

Figure E-6 Analyze Channel Information PD panel for CHPID 2.7E

Note the blocks that are highlighted: On the top left side, in the block labeled Definition Information, notice that Partition ID 23 is the LPAR ID associated to the partition where this panel was selected from. MIF image ID 3 corresponds to the LPAR MIF ID on CSS 2. also note that CHPID 7E is reported as Spanned, meaning that it has been shared between LPARs in different CSSs and is associated to CHPID 5A2. It also reports that CHPID 7E is connected to Switch number 65. The block labeled Status Information indicates the status and the state information of this CHPID. This block also displays the Error code and the Bit Error Counter (Ber) accumulated values for the CHPID. The Error code, when not equal to 0 (zero), may be displayed using a panel button called Error details. The block on the top right side labeled Hardware Card type Cascade definition indicates that this channel was defined using 2 byte control unit link address. Normally, this is an indication that cascaded switch topology is being used. Note: HCD allows a dual-byte address definition even when a single FICON Director is used in the configuration. (In this case, the entry port and the destination port belong to the same switch number). The block labeled Error Status shows the link status, the connection port, and the SAP affinity associated to the CHPID. The IFCC threshold is has an initial value of 4 and decrements at every occurrence of an IFCC on this channel. When the value reaches 0 (zero), the indication of a IFCC Threshold Exceeded will be associated to the CHPID icon on the SE desktop. The Temporary error threshold works the same way as the IFCC threshold. The SAP affinity indicates which of the SAPs this CHPID is associated to.

Appendix E. Using HMC and SE for problem determination information

273

The block labeled Card Type Connection rate shows the FICON card description type associated with this CHPID. It has a direct relationship with the information provided about the Hardware type and Subtype presented in the top right corner of the panel.

Understanding Node (ND) information displayed on this panel

The data listed here is received from the request node identification (RNID) during the initialization of the link: Node type Self means the shown Node Descriptor (ND) information is from the channel itself. Attached indicates the shown Node Descriptor (ND) information is from the attached node. Node status Valid means there has been a successful port login. Invalid means there has not been any port login. VALID BUT NOT CURRENT means there has been a port login, but the link got dropped afterwards. Type/Model Plant Sequence number Tag This field indicates the machine type and model of the node. This field indicates the plant where the node was manufactured. This field represents the serial number of the node. The Tag field represents the CSS.CHPID for System z server. For the attached node, you need to know whether the tag is vendor-dependent, and you must also refer to the vendors documentation to understand how to decode this field. The System z server adds a new flavor to the Tag field, as explained here: - The first character of the Tag field represents the CSS. - The last two digits of the Tag field represent the CHPID (the second digit showing a zero (0) is reserved). Table E-1 on page 275 illustrates the CSS/Tag relationship. The tag value in the top line gives the corresponding CSS IDs that the node has been defined to. For example, with Tag = A07E, the A indicates that the node is defined to CSS0, and CSS2 represents CHPID 7ETable E-1 on page 275 lists the CSS-to-tag relationship.

274

FICON Planning and Implementation Guide

Table E-1 CSS-to-tag relationship CSS\Tag 0 1 2 3 X X X X X X X X X X X X X X X X X 1 2 3 4 5 6 7 8 X 9 X A X B X C X X D X X E X X X F X X X X

World Wide Node Name World Wide Port Name

This is a unique 64-bit address that is used to identify the node in an FC topology. This address is assigned by the manufacturer. This is a unique 64-bit address that is used to identify the port in an FC topology. This address is assigned by the manufacturer.

Displaying the Analyze Subchannel Information PD panel

Follow these steps to open the Analyze Subchannel Information panel: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266). 2. On the SE workplace, double-click Groups. 3. Double-click CPC if you know the PCHID number associated. Otherwise, double-click Images and select the LPAR you are working from. 4. Right-click the CPC or Images and select Channels. If you select CPC, the PCHIDs will display. If you select Images, the CHPIDs associated to a logical partition will display. 5. Highlight the channel you that want to view. 6. Double-click Channel Problem Determination, which is located in the CHPID Operations task list. Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. 7. Select the option Analyze Subchannel Information and click OK. 8. Select the option Search by Device Number (you can also search by a unit address or subchannel number, if appropriate) and click OK. 9. Enter the device number (or unit address or subchannel number, depending on what you selected in the previous step) and select OK. In our example, we used device number D000 (see Figure E-7 on page 276).

Appendix E. Using HMC and SE for problem determination information

275

Figure E-7 Analyze Subchannel Information PD panel

Note the fields that are highlighted: The field Irpt parm: 02362250 displays this subchannels associated UCB SCP absolute address. During IPL device mapping process, the subchannel is connected to its UCW and this field contains the respective pointer to the UCB in SCP memory. The Pathing information is displayed in the block in the center of the panel. The list of candidate CHPIDs is shown in the top right corner. This Pathing information has various fields that are bit-position-relative to the defined channel paths to this device. Each bit position represents one channel path. Bit 0, or the first position bit from the left, corresponds to CSS.CHPID0.PCHID. Bit 1, or the second position bit from the left, corresponds to CSS.CHPID1.PCHI, and so on. The following fields are shown: CHPID - Associated CHPID number. LPM - Logical Path Mask: the bits show which CHPIDS can be used to access the device. LPE - Logical Path Established: this field indicates the paths to this device that successfully went through the pathing process. PNOM - Path Non-Operational Mask: this field indicates which paths from the list of candidate CHPIDs are in the Not Operational state.

276

FICON Planning and Implementation Guide

LPUM - Last Path Used Mask: this field indicates the last path used to access the device when successful or any abnormal condition occurred for the I/O operation. It is updated when the Ending Status is received from the device. PIM - Path Installed Mask: this field indicates the channel paths that are defined in the IOCDS. POM - Path Operational Mask: this field contains the paths online to the device. Initially (at System Reset), this value is set to FF (all ones) and the corresponding bit will be turned OFF when the first attempt to access the device fails (Inoperative). PAM - Path Available Mask: this field indicates the Physically Available channel paths. The initial value will be the paths defined in the IOCDS for the LPAR. This bit will be turned off when: The corresponding CHPID is configured OFF from the HMC. The CF CHP(xx), OFFLINE z/OS command is issued from the operating system console.

Displaying the Analyze Control Unit Header PD panel

Follow these steps to display the Control Unit Header Problem Determination panel: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266). 2. On the SE workplace, double-click Groups. 3. Double-click CPC if you know the PCHID number associated. Otherwise, double-click Images and select the LPAR you are working from. 4. Right-click the CPC or Image and select Channels. If you select CPC, the PCHIDs will display. If you select Images, the CHPIDs associated to a logical partition will display. 5. Highlight the channel you want to view. 6. Double-click Channel Problem Determination, which is located in the CHPID Operations task list. Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. 7. Select the option Analyze Control Unit Header and click OK. 8. Select the option Search by Device Number (you can also search by a unit address or subchannel number if appropriate) and click OK. 9. Enter the device number (or unit address or subchannel number, depending on what you have selected in the previous step) and select OK. The Control Unit Header PD panel is shown in Figure E-8 on page 278.

Appendix E. Using HMC and SE for problem determination information

277

Figure E-8 Analyze Control Unit Header PD panel

The starting point to obtain this data is provided by the device number entered when selecting this PD action. The output presented shows: This indicates all possible paths to get to the entered device. This indicates the channels associated to the displayed paths. This indicates the destination link address in use by each of the CHPIDs and respective paths. Node Type This indicates the type of attachment to the CHPIDs and Paths. Node Status This indicates the current status of the node. Type/Model, MFG, Plant, Seq. Number and Tag All together, these represent the Control Unit RNID (Remote Node Identification) information. Path CHPIDS Linkqq All the possible paths to get to the entered device are displayed, along with the CHPIDs associated to these paths and their respective FICON Director links, where applicable. This devices control unit characteristics are also displayed, along with the tags (which represent the attaching Host Bay Adapters) that the channels connect to. In our example we highlighted the FICON CPID 2.7E in the control unit header frame, which is connected to a 2107 using tag 0003. Refer to DS8000 port layout on page 262 for the logical port layout of a 2107. Using Table D-2 on page 263 to translate tag value 0003 to a physical port results in the following: Rack 1 (R1) I/O enclosure 1 (I1) Card 1 (C1) Port 3 (T3)

Displaying the Analyze Paths to a Device PD panel

Follow these steps to display the Analyze Paths to a Device panel: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266).

278

FICON Planning and Implementation Guide

2. On the SE workplace, double-click Groups. 3. Double-click CPC if you know the PCHID number associated. Otherwise, double-click Images and select the LPAR you are working from. 4. Right-click the CPC or Image and select Channels. If you select CPC, the PCHIDs will display. If you select Images, the CHPIDs associated to a logical partition will display. 5. Highlight the channel you want to view. 6. Double-click Channel Problem Determination, which is located in the CHPID Operations task list. Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. 7. Select the option Analyze Paths to a Device and click OK. 8. Select the option Search by Device Number (you can also search by a unit address or subchannel number if appropriate) and click OK. 9. Enter the device number (or unit address or subchannel number, depending on what you have selected in the previous step) and select OK. The Analyze Path to a Device panel is shown in Figure E-9.

Figure E-9 Analyze Paths to a Device PD panel

It can display important device path information, as follows: The devices associated subchannel. All paths by which the device can be accessed (this is relevant only for the partition the channel path has been selected from), with the following additional information: Avail CHPID Switch This shows if that path is available at that instance. This is the associated CHPID number. This shows if there is a FICON Director in the link.
Appendix E. Using HMC and SE for problem determination information

279

Switch Number Linkaddr Cuadd

This is the FICON Director number (if applicable). This is the destination link address (the port address the CU is connected to). This is the logical CU that the device is connected to.

Displaying the Analyze Device Status PD panel

Follow these steps to open the Analyze Paths to a Device panel: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266). 2. On the SE workplace, double-click Groups. 3. Double-click CPC if you know the PCHID number associated. Otherwise, double-click Images and select the LPAR you are working from. 4. Right-click the CPC or Image and select Channels. If you select CPC, the PCHIDs will display. If you select Images, the CHPIDs associated to a logical partition will display. 5. Highlight the channel you want to view. 6. Double-click Channel Problem Determination, which is located in the CHPID Operations task list. Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. 7. Select the option Analyze Device Status and click OK. 8. Select the option Search by Device Number (you can also search by a unit address or subchannel number if appropriate) and click OK. 9. Enter the device number (or unit address or subchannel number, depending on what you selected in the previous step) and select OK. The Analyze Device Status PD panel is shown in Figure E-10 on page 281.

280

FICON Planning and Implementation Guide

Figure E-10 Analyze Device Status panel

The Analyze Device Status panel reports the current state of all devices pertaining to a Logical Control Unit. Note that a Filter Status button is provided to find a specific status a particular device.

Displaying the Analyze Serial Link Status PD panel

Follow these steps to open the Analyze Paths to a Device panel: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266). 2. On the SE workplace, double-click Groups. 3. Double-click CPC if you know the PCHID number associated. Otherwise, double-click Images and select the LPAR you are working from. 4. Right-click the CPC or Images and select Channels. If you select CPC, the PCHIDs will display. If you select Images, the CHPIDs associated to a logical partition will display. 5. Highlight the channel you want to view. 6. Double-click Channel Problem Determination, which is located in the CHPID Operations task list Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. 7. Select the option Analyze Serial Link Status and click OK. The Analyze Serial Link Status panel is shown in Figure E-11 on page 282.

Appendix E. Using HMC and SE for problem determination information

281

Figure E-11 Analyze Serial Link Status PD panel

The following logical pathing information is displayed: MIF Image ID - This is the MIF ID of the image we selected the channel from. CHPID - This is the logical channel consisting of CSS and CHPID. Channel type - This is the channel type as defined in the IOCP. Switch number - This is the switch number of the FICON Director that the channel is connected to. Switch number valid - This is the total number of switches in the links. Channel link address - This is the link address (port address) that the channel is connected to. The status of all logical link addresses can be seen in this panel, as well. This provides information about the logical CU initialization from a channel point of view. The given CU address cannot be accessed if the status is other than INTIALIZATION COMPLETE. Channel 2.7E from MIF Image 3 connected to Switch number 65 port address 03. A series of logical Control Units (residing in the same physical CU) being accessed from Channel 2.7E show the status INITIALIZATION COMPLETE. Any status other than INITIALIZATION COMPLETE indicates a potential problem that may need to be investigated. Note, however, that a status of INITIALIZATION PENDING is not necessarily an error indication. It depends on whether the logical CU is expected to be available. For example, if a channel is trying to establish a FCTC connection to another server or image but the LPAR is not activated, this would show a status of INITIALIZATION PENDING. However, the status is expected to change after the target LPAR is activated.

282

FICON Planning and Implementation Guide

 Other typical error messages are: Initialization Complete - CU Busy Initialization Complete - Dynamic Switch Port Busy Initialization in Progress CU Reset in Progress Remove Logical Path in Progress Link Configuration Problem - Port Reject - Address not Valid Switch Port Not Defined - Port Reject - Undefined Destination Address Switch Port Malfunction - Port Reject - Port Destination Malfunction Switch Port Not Available - Port Reject - Port Intervention Required Link Configuration Problem - Link Reject - Unrecognized Device Level Link Configuration Problem - Link Reject - Uninstalled Link Control Function Link Level Error - Link Reject - Transmission Error Link Level Error - Link Reject - Destination Address not Valid Link Level Error - Link Reject - Acquire Address Error Link Configuration Problem - Initialization Failure - Channel/CU mismatch CU Resources Exceeded - Init Failure - No Resources Available Link Level Error - Channel Detected Error Internal CTC Initialization Complete Link Level Error - FICON FLOGI ELS Error Link Level Error - FICON PLOGI ELS Error Link Level Error - FICON RNID ELS Error Link Level Error - FICON SCR ELS Error Link Level Error - FICON LIR ELS Error Link Level Error - FICON Invalid Attachment CU resources exceeded - This indicates that the CU ran out of logical CUs. CU reset in progress - This indicates that there is a reset initiated by the CU itself in progress. Channel/CU mismatch - This indicates that the logical CU address (defined in the CU) does not match its definition in the IOC.P PLOGI error - This indicates that the link initialization did not succeed to the Destination Port.

Note: The latest implementation of the System z Server SE Problem Determination facility improves the PLOGI and FLOGI error messages substantially. With this new implementation the panel will show, in plain English, what went wrong during the CU link initialization. Figure E-12 shows one of the new messages issued when initialization failures occur.

Figure E-12 New PLOGI Initialization message

Appendix E. Using HMC and SE for problem determination information

283

Displaying the Analyze Link Error Statistics Block PD panel

Follow these steps to open the Analyze Paths to a Device panel: 1. Log on to the HMC and SE (see Logging on to the HMC and SE on page 266). 2. On the SE workplace, double-click Groups. 3. Double-click CPC if you know the PCHID number associated. Otherwise, double-click Images and select the LPAR you are working from. 4. Right-click the CPC or Images and select Channels. If you select CPC, the PCHIDs display. If you select Images, the CHPIDs associated to a logical partition display. 5. Highlight the channel you want to view. 6. Double-click Channel Problem Determination, which is located in the CHPID Operations task list. Note: The SE will post a panel for an image (LPAR) selection if you are working with a PCHID instead of a CHPID. 7. Select the option Analyze Link Error Statistics Block and click OK. The Analyze Link Error Statistics Block PD panel is shown in Figure E-13.

Figure E-13 Link Error Statistics Block PD panel

The Analyze Link Error Statistics Block (LESB) is a new implementation on the System z Servers Problem Determination SE panels. It displays the Channel N-Port LESB concurrently with channel operations.

284

FICON Planning and Implementation Guide

The various types of errors are accumulated to provide you with a view of how the channel is operating. A refresh provides a dynamic display of the counters. The information gathered from the Channel N-Port is also recorded in the EREP SLH records when a Purge Path Extended (PPE) operation is performed. For more information about the PPE process, refer to 10.5, FICON Purge Path Extended on page 222.

Appendix E. Using HMC and SE for problem determination information

285

286

FICON Planning and Implementation Guide

Appendix F.

Useful z/OS commands

This appendix discusses the information pertaining to the FICON environment that can be obtained from various z/OS commands.

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

287

Using z/OS commands for problem determination

Figure F-1 represents the FICON environment that we used to produce the output of the z/OS commands discussed in this appendix.

z10 EC
SCZP201
LPAR SC30 A23 z/OS V1R10
FICON Express8 LX CSS2 MIF 3 PCHID 5A2 CHPID 7E

System Storage SAN384B

Switch # 66 Switch @ 66
Port 8D Port C4 Port 88 Port CF

DS8000
CU# D000
LX 0003

D0xx

ISL
Port 03 Port 2D Port 34

LPAR SC31 A24 z/OS V1R10

FICON Express8 LX CSS2 MIF 4 PCHID 5D2 CHPID 82

Switch # 65 Switch @ 65

LX 0242

D1xx

CU# D100
Port 1B
* All cable connectors are LC Duplex type

SAN768B
Figure F-1 Our FICON configuration

Displaying system status using D M=CPU

Example F-1 shows the output of a z/OS D M=CPU command.The output shows the CPU type and serial number. It also defines the CPU characteristics associated to a specific partition (LPAR A23, in this case). The output shows the General CPs, zIIPs and zAAPs configured, as well as the CPU Node Descriptor (ND), the System Information taken from the Store System Information (STSI) instruction, the CPC name, the CSS ID where this partition is running, as well as its MIF and LPAR IDs.
Example: F-1 Displaying the CPU characteristics

D M=CPU IEE174I 11.38.13 DISPLAY M 944 PROCESSOR STATUS ID CPU SERIAL 00 + 23DE502097 01 + 23DE502097 02 03 04 -A 05 -A 06 +I 23DE502097 07 -I

288

FICON Planning and Implementation Guide

CPC ND = 002097.E26.IBM.02.00000001DE50 CPC SI = 2097.714.IBM.02.000000000001DE50 Model: E26 CPC ID = 00 CPC NAME = CZP201 LP NAME = A23 LP ID = 23 CSS ID = 2 MIF ID = 3 + ONLINE - OFFLINE . DOES NOT EXIST W WLM-MANAGED N NOT AVAILABLE A APPLICATION ASSIST PROCESSOR (zAAP) I INTEGRATED INFORMATION PROCESSOR (zIIP) CPC ND CENTRAL PROCESSING COMPLEX NODE DESCRIPTOR CPC SI SYSTEM INFORMATION FROM STSI INSTRUCTION CPC ID CENTRAL PROCESSING COMPLEX IDENTIFIER CPC NAME CENTRAL PROCESSING COMPLEX NAME LP NAME LOGICAL PARTITION NAME LP ID LOGICAL PARTITION IDENTIFIER CSS ID CHANNEL SUBSYSTEM IDENTIFIER MIF ID MULTIPLE IMAGE FACILITY IMAGE IDENTIFIER

The other fields in the display output have the following meanings: CPC name This reflects the CPC object name customized to the HMC at System z installation time. The CPC name is also defined in the HCD/HCM (a CPC can only be defined once). HCD/HCM also requires a Proc.ID to be specified for the CPC. LP name This is defined in the HCD/HCM logical partition definition or in the IOCP RESOURCE statement. The name has to be unique for each logical partition across all four Channel Subsystems (CSSs) on System z. Changing, adding, or deleting LPARs requires a new IOCDS and a Power-on Reset (POR) on System z9, and is fully concurrent on System z10. Note: System z10 (2097 and 2098) implemented a fixed Hardware System Area (HSA) storage approach that allows for concurrent adding logical partitions, channel subsystems, subchannel sets, logical processors and cryptographic co-processors. Adding logical processors to a LPAR requires z/OS 1.10 or later. CSS ID This indicates in which CSS this LPAR has been created. The number of CSSs that a System z supports depends on the machine model. The Business Class (BC) models for the z9 and z10 support two CSS. The Enterprise Class (EC) models support four CSSs in total. Increasing or decreasing the number of CSSs in use in a z9 requires a POR and is concurrent on z10. Each CSS supports two subchannel sets: SS0 and SS1. MIF ID This is a value, between 1 and 15, that is associated to a partition in one CSS. The MIF ID numbers are defined in the HCD/HCM in the IOCP RESOURCE statement. The MIF ID must be unique for each logical partition in a System z CSS. The MIF ID is used by the System z channel subsystem and channels to identify the source of an I/O

Appendix F. Useful z/OS commands

289

request. You need to know the MIF ID to resolve LP-related problems or to resolve failures in establishing a FICON logical path. LP ID The logical partition ID (LP ID) is specified in the Logical Partition Image Profile using the HMC. The LP ID number, from x00 to x3F, must be unique across all partitions on a System z.

Displaying additional z/OS information using D IPLINFO

Example F-2 displays the output of a z/OS D IPLINFO command.
Example: F-2 Displaying the z/OS IPL information

D IPLINFO IEE254I 09.58.13 IPLINFO DISPLAY 899 SYSTEM IPLED AT 09.16.10 ON 03/12/2009 RELEASE z/OS 01.10.00 LICENSE = z/OS USED LOAD01 IN SYS0.IPLPARM ON C730 ARCHLVL = 2 MTLSHARE = N IEASYM LIST = XX IEASYS LIST = (00) (OP) IODF DEVICE C730 IPL DEVICE DC0D VOLUME Z1ARC1 The output displays the following information: The z/OS release (RELEASE z/OS 01.10.00)that this partition is running The time stamp, which shows that this partition was IPLed (SYSTEM IPLED AT 09.16.10 ON 03/12/2009) The LOAD member selected from the PARMLIB (USED LOAD01 IN SYS0.IPLPARM ON C730) The system architecture level (ARCHLVL = 2), indicating a 64-bit operating system in this case The IODF device The IPL device and its VOLSER (IPL DEVICE DC0D VOLUME Z1ARC1) Note: IPLing the system from the wrong device, or selecting an unappropriated IODF device, may lead to I/O-related and configuration-related problems.

Displaying the I/O configuration using D IOS,CONFIG

Example F-3 on page 291 displays the following information about the I/O configuration in use by the System z: The IODF data set name (ACTIVE IODF DATA SET = SYS6.IODF75) The active CSS and subchannel sets in use (ACTIVE CSS:2 SUBCHANNEL SETS CONFIGURED:0, 1) The token that will allow dynamic changes to the configuration A match between the TOKEN IODF and the ACTIVE IODF DATA SET names will indicate whether the system allows dynamic hardware and software changes to the I/O configuration.

290

FICON Planning and Implementation Guide

Example: F-3 D IOS,CONFIG z/OS command output

D IOS,CONFIG IOS506I 10.00.19 I/O CONFIG DATA 901 ACTIVE IODF DATA SET = SYS6.IODF75 CONFIGURATION ID = TEST2094 EDT ID = 01 TOKEN: PROCESSOR DATE TIME DESCRIPTION SOURCE: SCZP201 09-05-04 09:50:17 SYS6 IODF75 ACTIVE CSS: 2 SUBCHANNEL SETS CONFIGURED: 0, 1 CHANNEL MEASUREMENT BLOCK FACILITY IS ACTIVE

Displaying HSA usage using D IOS,CONFIG(HSA)

Example F-4 displays information about the amount of physical control units and subchannels in use per subchannel set in each defined channel subsystem (CSS). Note: The maximum number of subchannels on subchannel set 0 (SS0) is 65280. The maximum number of subchannels on subchannel set 1 (SS1) is 65535. A subchannel corresponds to a device Unit Control Word (UCW), which contains hardware controls and definition information about a particular device. Note the following points: System z9 EC and z10 EC can have up to four Channel Subsystems (0,1, 2 and 3) with two subchannel sets each. System z9 BC and z10 BC can have up to two Channel Subsystems (0 and 1) with two subchannel sets each.
Example: F-4 D IOS,CONFIG(HSA) z/OS command output

D IOS,CONFIG(HSA) IOS506I 10.01.22 I/O CONFIG DATA 903 HARDWARE SYSTEM AREA AVAILABLE FOR CONFIGURATION CHANGES PHYSICAL CONTROL UNITS 8064 CSS 0 - LOGICAL CONTROL UNITS 3980 SS 0 SUBCHANNELS 51215 SS 1 SUBCHANNELS 65535 CSS 1 - LOGICAL CONTROL UNITS 3987 SS 0 SUBCHANNELS 51375 SS 1 SUBCHANNELS 65535 CSS 2 - LOGICAL CONTROL UNITS 4004 SS 0 SUBCHANNELS 52515 SS 1 SUBCHANNELS 65535 CSS 3 - LOGICAL CONTROL UNITS 4050 SS 0 SUBCHANNELS 56520 SS 1 SUBCHANNELS 65535 The output displays: The total number of physical control units defined in the configuration: 8064 The number of logical control units per Channel Subsystem The number of subchannels defined in each CSSs subchannel set (0 and 1)

Appendix F. Useful z/OS commands

291

Display units command D U

Use the z/OS Display Units command as the starting point for determining the status of a device defined to z/OS, and the z/OS status of the defined paths to the device. There are a few exceptions to this, where the z/OS Display Units command will not display the status of a defined device. (DASD ALIAS devices is one of these exceptions; for more information refer to Using a D M=DEV(dddd) command for an ALIAS device number on page 299 and to DEVSERV command - DS P,dddd,n on page 300.) The Display Units z/OS command can be used to: Determine if the device is defined to z/OS (exception devices are not displayed). If the z/OS status for the requested device number is displayed, it is defined to z/OS. Determine the defined device type: 3390, 3590, CTC, and so on. Identify whether the defined z/OS type is not equal to the device being addressed (when trying to bring the device online); if it is not equal, then the device will not be brought online to z/OS. Indicate the z/OS status of the device; the statuses include: O F-NRD Allocated A A-BSY Online. Offline. Online and in use by a job or task. Allocated and busy, meaning that there is an I/O operation in progress for the z/OS device in the channel subsystem image (in the channel subsystem for the logical partition that z/OS is running in). For devices that support volumes, it indicates which volume is currently mounted on that device.

VOLSER = NE800

Whenever a z/OS Display Units command (or any other general form of this command such as D U,DASD,,dddd,n or D U,,ALLOC,dddd,n) is entered, z/OS will display the status of the devices that the z/OS SCP has recorded at that time. If the devices are not defined to z/OS, or if they are exception devices, then z/OS will display one of the following pieces of information: The next device number or the next device number of that type, for a display units type request, or The next device of that status, for a display units by status, or If a combination of both type and status is used, then it will be the next device of that type and status (refer to Example F-6 on page 294) Because of this, always verify that the device (or devices) displayed in response to a z/OS Display Units command request is for your display units requested device number, or that it includes your requested device number.

D U,,,dddd,1
The Display Units command is useful for determining the z/OS status of a device. The output from this command is shown in Example F-5 on page 293. The example shows the normal result of displaying the z/OS status of a group of devices. Devices D000, D001, and D002 are online and not allocated; that is, the devices are not in use by a job or application. 292
FICON Planning and Implementation Guide

Device D003 is both online and allocated. Being allocated means the device is in use by a job or an application. If the device is online and allocated, you can use the z/OS display allocation command to determine who the device is allocated to. Enter the display allocation command DU,,ALLOC,8003,1 and it will show who the device is allocated to.
Example: F-5 z/OS display device status of a group of four devices defined and online

D U,,,D000,4 IEE457I 20.35.52 UNIT STATUS 727 UNIT TYPE STATUS VOLSER VOLSTATE D000 3390 O Z1BRZ1 PRIV/RSDNT D001 3390 O Z18DL2 PRIV/RSDNT D002 3390 O Z18RB1 PRIV/RSDNT D003 3390 A TOTD94 PRIV/RSDNT There are also a few other device statuses that may be returned for a D U command: Automatically switchable. Hardware failure. The device has been BOXED (refer to Boxed status notes: below). BSY Busy. C Console. F Offline. F indicates that more than one bit is set in the UCB. This is used when a combination of offline and some other status value needs to be displayed (for example, F-NRD). L The release on a device is pending and reserve may or may not have occurred. M A device managed by a device manager, such as JES3 or a non-IBM tape management subsystem. MTP - Mount Pending NRD - Not Ready O - ONLINE OFFLINE This is used when the only status value that needs to be displayed is OFFLINE. P Reserve Pending PND Offline pending PO Offline pending, and also not ready. This status value is displayed only if NRD is also displayed on the status line. PUL Unload pending. R Reserved, shared DASD or exclusively-assigned device. RAL Restricted to Allocation. S - SYSRES SPD Suspended. (a paging volume). The channel program is temporarily suspended while the system is using the device. SYS Allocated to system. UNAVL The device has been marked as unavailable for allocation by the VARY xxxx,UNAVAIL operator command. AS BOX

Appendix F. Useful z/OS commands

293

Boxed status notes: If the reported status of a device is O-BOX (Online and Boxed), that status can be cleared by using a vary online with unconditional parameter: V dddd,ONLINE,UNCOND. If the reported status is F-BOX (Offline and Boxed), the device can be brought back online with the VARY dddd,ONLINE command. This will enable the UCW and perform the online processing to the device. Assuming that the error condition has been resolved, the device will come online. If the error condition still exists, however, the device may remain in the boxed state.

Displaying a device not in current configuration

If the z/OS display units request is for a device that is not defined in the current OSCONFIG member (in use by z/OS), or if the device is an exception device, then z/OS will not be able to display the status for that device. z/OS will display the next displayable device number; or the next device number of the requested device by type (when using the display units type command); or the requested device by status (when using the display unit status command). For this reason, always check that the devices displayed include your requested device; see in Example F-6.
Example: F-6 D U request of a non-existing device

D U,,,8045,1 IEE457I 08.57.21 UNIT STATUS 822 UNIT TYPE STATUS VOLSER VOLSTATE 8100 3390 O NW8100 PRIV/RSDNT The D U request shown in this example was for device 8045; however, notice that the actual display reports the status for device 8100.

D U,,,ALLOC,dddd,n
The D U,,,ALLOC,ddd,n command is one of the various formats for the Display Units z/OS device commands. Its main purpose is to display the allocations of a device when its status has been determined to be Allocated. Example F-7 shows the device number D003 as online, allocated to two jobs, and currently in use.
Example: F-7 z/OS display device allocation

D U,,,ALLOC,D003,1 IEE106I 08.07.58 UNITS ALLOCATED 722 UNIT JOBNAME ASID JOBNAME ASID JOBNAME ASID JOBNAME ASID D003 DUMPSRV 0005 MVSRECVA 0028

The information displayed shows that the jobs using the device are: Jobname - DUMPSRV Jobname - MVSRECVA

V dddd,offline
In Example F-8 on page 295, the V dddd,OFFLINE command is issued against device D003. However, as shown in Example F-7, device D003 is allocated and in use. Therefore, the action of varying this device offline will be in a pending status.

294

FICON Planning and Implementation Guide

Example: F-8 Varying an allocated device offline

V D003,OFFLINE IEF524I 8003, VOLUME TOTD94 PENDING OFFLINE .... D U,,,D000,4 IEE457I 08.16.21 UNIT STATUS 756 UNIT TYPE STATUS VOLSER VOLSTATE D000 3390 O Z1BRZ1 PRIV/RSDNT D001 3390 O Z18DL2 PRIV/RSDNT D002 3390 O Z18RB1 PRIV/RSDNT D003 3390 A-PND TOTD94 PRIV/RSDNT

Device in OFFLINE status

Example F-9 shows a display reporting that device D003 is in the OFFLINE state.
Example: F-9 Device is now offline after the allocation ended

D U,,,D003,1 IEE457I 18.27.12 UNIT STATUS 674 UNIT TYPE STATUS VOLSER D003 3390 OFFLINE

VOLSTATE /RSDNT

A device can be offline for a number of reasons: It has been varied offline by the operator (this is the case shown in Example F-90. It was defined to the operating system as being offline at system IPL time. The device is defined to the operating system but not to the logical partition that the z/OS is running in; that is, there is no subchannel. The device is defined to the operating system but not to the channel subsystem; that is, there is no subchannel. There is no operational path to the control-unit or device. All the operational paths were logically varied offline. The definition to access the control unit device is not correct. CHPID. Link destination link address (destination port address). CUADD - the CU logical image address. For an IBM 2107, the CUADD is unique for an LSS. For a CTC, the CUADD is the MIF ID of the target LPAR. The device addressing definition is not correct. The device is not ready. Link destination link address (destination port address). The device is in an error state. The defined device type and the physical device type (the actual device) do not match. An error condition occurred during the use of the device (that is, a reserve or assign was lost) and the device was forced offline (it may have been boxed during the recovery process, as well).

V dddd,ONLINE
Use the z/OS command shown in Example F-10 on page 296 to bring a single device online. There are variations to the V dddd,online command that allow more than one device to be brought online simultaneously. A group of devices can be varied online using dddd,dddd or dddd-dddd as variables for the vary command.

Appendix F. Useful z/OS commands

295

Example: F-10 V dddd, ONLINE results

V D003,ONLINE IEE302I D003

ONLINE

A failure to bring a device online can be caused by: There is no UCB for the target device. There is no UCW (subchannel) for the target device. All paths to the device are offline. None of the defined paths are operational. The CU and or the device addressing is not correct. There is a mismatch of device type between the definition and the actual device.

D M=DEV(dddd)
The D M=DEV(dddd) command displays the device status for device dddd. Example F-11 shows the output of this display device matrix command. It shows the path status of all paths that have been defined to this single device address (device D003, in the example). The destination link address as well as the entry link address are individually provided for each channel accessing the device.
Example: F-11 z/OS display output for D M=DEV(dddd)

D M=DEV(D003) IEE174I 09.50.12 DISPLAY M 767 DEVICE D003 STATUS=ONLINE CHP 50 54 58 5C 7E 82 ENTRY LINK ADDRESS 0F 0F 1B 1B 6503 651B DEST LINK ADDRESS 2C 2C 0A 0A 6688 66CF PATH ONLINE N N Y Y Y Y CHP PHYSICALLY ONLINE N N Y Y Y Y PATH OPERATIONAL N N Y Y Y Y MANAGED N N N N N N CU NUMBER D000 D000 D000 D000 D000 D000 MAXIMUM MANAGED CHPID(S) ALLOWED: 0 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107.900.IBM.75.0000000BALB1.0300 SCP TOKEN NED = 002107.900.IBM.75.0000000BALB1.0000 SCP DEVICE NED = 002107.900.IBM.75.0000000BALB1.0003 HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW, ZHPF Example F-12 is a variation of the D M=DEV(dddd) command where a unique CHPID that accesses the device dddd is selected for display.
Example: F-12 Variation of D M=DEV(dddd,(cc))

D M=DEV(D003,(7E)) IEE174I 18.00.58 DISPLAY M 907 DEVICE D003 STATUS=ONLINE CHP 7E ENTRY LINK ADDRESS 6503 DEST LINK ADDRESS 6688 PATH ONLINE Y CHP PHYSICALLY ONLINE Y

296

FICON Planning and Implementation Guide

PATH OPERATIONAL Y MANAGED N CU NUMBER D000 DESTINATION CU LOGICAL ADDRESS = 00 SCP CU ND = 002107.900.IBM.75.0000000BALB1.0300 SCP TOKEN NED = 002107.900.IBM.75.0000000BALB1.0000 SCP DEVICE NED = 002107.900.IBM.75.0000000BALB1.0003 HYPERPAV ALIASES CONFIGURED = 143 FUNCTIONS ENABLED = MIDAW, ZHPF For both D M=DEV examples, the following important information is displayed: The device current status: DEVICE D003 STATUS=ONLINE. The FICON Director entry port: 6503 - FICON Director @ 65, Port 03. The FICON Director exit port: 6688 - FICON Director @ 66, Port 88. Note that the FICON Director addresses are different for the entry and exit (Dest) Ports. This is an indication that a FICON Director cascading topology is being used. PATH ONLINE = Y. The PATH ONLINE field indicates whether the path is logically online to z/OS and will change as a result of the VARY PATH command. Information for this field is obtained from the UCBLPM (UCB Logical Path Mask). CHP PHYSICALLY ONLINE = Y. The CHP PHYSICALLY ONLINE field indicates whether the path is physically available. This field will change as a result of the z/OS CF CHP command. Information for this field is obtained from the Path Available Mask (PAM) in the Unit Control Word (UCW) or subchannel. Note: The CHP PHYSICALLY ONLINE field might not be accurate when the CHPID has been stolen from z/OS using the HMC/SE CHPID Config OFF/ON facility instead of the z/OS CF CHP(cc),OFFLINE command. PATH OPERATIONAL = Y This means that the PATH OPERATIONAL field status is changed as a result of the channel subsystem attempting an I/O operation on a path and the path responding with a not operational I/O interface sequence. This response can be due to: Disabling a channel interface Host Bay Adapter (HBA) at the control unit Disabling a port on a FICON Director Powering off a control unit or cluster Powering off a DASD controller Note: The PATH OPERATIONAL status can be easily misinterpreted. The information displayed is a direct reflection of a path mask called the Path Operational Mask (POM) in the UCW. PATH NOT VALIDATED = Y The PATH NOT VALIDATED field may be shown if the device was not online at IPL time. If a device is not online when the IPL device pathing process occurs, the UCBVALPH bit is set to On to indicate that the paths to the device were not validated. Validation will occur when the device is varied online. MANAGED = N This field indicates whether the channel is using the Dynamic Channel Management (DCM) facility of z/OS.

Appendix F. Useful z/OS commands

297

CU NUMBER = D000 This field presents the Control Unit Number associated to the physically attached defined CU. DESTINATION CU LOGICAL ADDRESS= 00 The DESTINATION CU LOGICAL ADDRESS field shows the logical link address of a control unit that is used to access the I/O device associated with the specified channel paths. SCP CU ND = 002107.900.IBM.75.0000000BALB1.0300 This shows the Node Descriptor (ND) LAST obtained by the System Control Program (SCP). ACTUAL CU ND = xxxxxx.xxx.xxx.xx.xxxxxxxxx.xxxx (not shown in the example) This shows the Node Descriptor of the attached subsystem read from the device in response to this command. SCP TOKEN NED = 002107.900.IBM.75.0000000BALB1.0000 This shows the Node-Element Descriptor last obtained by the SCP. ACTUAL TOKEN NED = xxxxx.xxx.xxx.xx.xxxxxxxxxx.xxxx (not shown in the example) This shows the Node-Element Descriptor read from the device in response to this command. SCP DEVICE NED = 002107.900.IBM.75.0000000BALB1.0003 This shows the Node-Element Descriptor last obtained by the SCP. ACTUAL DEVICE NED = xxxxx.xxx.xxx.xx.xxxxxxxxxx.xxxx (not shown in the example) This shows the Node-Element Descriptor read from the device in response to this command. RNID = xxxxx.xxx.xxx.xx.xxxxxxxxxx.xxxx (not shown in the example) This is the Control Unit Remote Node IDentifier obtained by the Channel Subsystem during Channel - CU initialization using Store Subsystem Channel Information (CHSC). This RNID information is kept in the Hardware System Area (HSA). HIPERPAV ALIASES CONFIGURED = 143 This field shows the number of (ALIAS) HIPERPAV ALIAS devices configured in this LCU. See Example F-14 on page 299 for a display of an alias device via D M=DEV command. FUNCTIONS ENABLED = MIDAW, ZHPF This field shows the additional facilities that this device is capable of using.

Using a D M=DEV command without specifying a device

Example F-13 shows the output of a display device (D M=DEV) command without specifying a device number. The information for each defined device will be provided by the display, including device-specific information about its current state.
Example: F-13 Using D M=DEV without specifying a device

D M=DEV IEE174I 13.46.46 DISPLAY M 238 DEVICE STATUS: NUMBER OF ONLINE 0 1 2 3 4 5 6 7 8 0006 . 1 1 1 1 2 2 . . 00B0 # # # # # # # # # 0180 4 4 4 4 4 4 4 4 4 298
FICON Planning and Implementation Guide

CHANNEL PATHS 9 A B C D . . . . . # # # # # 4 4 4 4 4

E . # 4

F . # 4

0181 4 4 4 4 4 4 4 4 4 0182 4 4 4 4 4 4 4 4 4 0183 4 4 4 4 4 4 4 4 4 ..... 0404 DN DN DN DN 1 1 1 1 DN 04B0 1 1 1 1 . . . . 1 0504 DN DN DN DN 1 1 1 1 DN ..... 0FB0 # # # # 1 1 1 1 1 ************************ SYMBOL @ ONLINE, PHYSICALLY ONLINE, AND + ONLINE # DEVICE OFFLINE BX DEVICE IS BOXED SN DN DEVICE NOT AVAILABLE PE AL DEVICE IS AN ALIAS UL HA DEVICE IS A HYPERPAV ALIAS HU

4 4 4

4 4 4

4 4 4

4 4 4

4 4 4 . . .

4 4 4 . . .

4 4 4 . . .

DN DN DN . 1 1 1 . DN DN DN .

1 1 1 1 1 1 1 EXPLANATIONS ************************ OPERATIONAL INDICATORS ARE NOT EQUAL . DOES NOT EXIST SUBCHANNEL NOT AVAILABLE SUBCHANNEL IN PERMANENT ERROR DEVICE IS AN UNBOUND ALIAS HYPERPAV ALIAS UNUSABLE

D M=DEV, which is shown in Example F-13 on page 298, provides symbols identifying the state of existing devices. Significant symbolic meanings are explained here: DN The DN symbol means that the device has been defined to the z/OS IOGEN (OSCONFIG) but has not been defined in the IOCP; that is, there is no subchannel for this device known to the channel subsystem.This normally occurs when the wrong IOCDS or OSCONFIG is loaded. The UCB for a DN device has the Not Connected bit left on during the device mapping time during IPL. The dot (.) indicates that the device has not been defined to the z/OS in the config. The numeric digit (1) indicates the number of paths defined for the device. Its possible to have a combination of a digit and a symbol. For example, 4@ indicates an out-of-line path condition; online, physically online, and operational indicators are not equal with at least one of these paths, where: Online - means the path is online. Physically online - means the CHPID is physically online. Operational - means the path is operational (POM).

. 1

Using a D M=DEV(dddd) command for an ALIAS device number

Example F-14 shows an output of a D M=DEV(dddd) issued against an ALIAS device address.
Example: F-14 Displaying an ALIAS device using D M=DEV(dddd)

D M=DEV(D080) IEE174I 18.22.19 DISPLAY M 660 DEVICE D080 STATUS=HYPERPAV ALIAS

Using a D M=DEV(dddd) command to a non-existing device

Example F-15 on page 299 shows the output of D M=DEV(dddd) issued against a non-existing device. There is no UCB available for this device in the OSCONFIG being used.
Example: F-15 Displaying a non-existing device via D M=DEV(dddd)

D M=DEV(4000)

Appendix F. Useful z/OS commands

299

IEE174I 18.23.52 DISPLAY M 666 DEVICE 4000 STATUS=NOT IN SYSTEM

DEVSERV command - DS P,dddd,n

The DEVSERV command requests a display of the current status of DASD and TAPE devices. The response is a display of basic status information about a device, a group of devices, or storage control units, and optionally can include a broad range of additional information. With the DEVSERV command you can display: Device Number Device type Logical mode of the device Number of data sets allocated on the volume Volume serial label Channel path ID Status of the path Status of an SMS-managed device Volume Status Storage Group name Storage Group status Control unit type and model Control unit serial number Device capacity, in cylinders Device extended function status Unit Control Block (UCB) device type information Help text, when you request it The following information can be displayed, if the device belongs to the DASD storage subsystem: Real device type (if what is shown is an emulated device type) Control unit type and model (or emulated control unit type and model, if the real and emulated control units are not the same) Subsystem ID for this storage subsystem Cache fast write state Track caching state DASD fast write state State of pinned data State of dual copy, PPRC, or SPARing (if there is any) Address of the other device in a dual copy pair Channel subsystem device address Subsystem internal logical device address

300

FICON Planning and Implementation Guide

 An indication if the device extended function status information is inconsistent between z/OS control blocks and the storage subsystem An indication if the defined (UCB) device type is inconsistent with the real device type Optionally, the total number of cylinders for each unique track format (3380, 3390, and 9345) for all of the devices within the scope of the request The following information, if the device belongs to a tape library: Device type equivalent to DTYPE from the DS P command Device status indicating online/offline and ready/not ready Device type and model Device serial number Library identification number An indication if the defined (UCB) device type is inconsistent with the real device type Example F-16 shows the output of a DEVSERV command. The response is a display of basic status information about a device, a group of devices, or storage control units, and optionally can include a broad range of additional information. The path status for all defined channel paths is included, as well. There are different ways of using the DEVSERV command: DS P,1C00 - Test device 1C00 only through all paths. DS P,2000,16,ON - Test only ONLINE devices in 2000-200F. DS QD,3200,1,UCB,DCE,SSSCB,DPCT - Display MVS control blocks (1 device). DS QD,4800,1,RDC,RCD,SNSS - Get hardware information and update control blocks.
Example: F-16 Devserv QDASD with options RDC and DCE

DS QD,9000,RDC,DCE IEE459I 09.53.04 DEVSERV QDASD 445 UNIT VOLSER SCUTYPE DEVTYPE CYL SSID SCU-SERIAL DEV-SERIAL EF-CHK 9000 IN9000 2107921 2107000 32760 1000 0113-00511 0113-00511 **OK** READ DEVICE CHARACTERISTIC 2107E833900A5E80 FFF720247FF8000F E000E5A205940222 1309067400000000 0000000000000000 24241F02DFEE0001 0677080F007F4A00 003C000000000000 DCE AT V020A2A20 3878807100C445C0 0000000001F0FED8 D8007FF87FF72424 1FF7080000410000 00FC24DC9400F0FE 001F3C1E00078000 0000000000000000 **** 1 DEVICE(S) MET THE SELECTION CRITERIA **** 0 DEVICE(S) FAILED EXTENDED FUNCTION CHECKIN DS QT,TYPE=3490<,DEFINED> - Display TAPE with 3490 information. You can use the QTAPE (QT) parameter to display tape information. Using the MED (medium) option DS QT,xxxx,MED,nnn allows you to display information for the device type, media type, and cartridge volume serial number. DS QPAVS,dddd,n - Display status of Parallel Access Volume devices. To learn about other uses of the DEVSERV command, refer to z/OS MVS System Commands, SA22-7627, or to DFSMS Storage Administration Reference, SC26-7402.

Appendix F. Useful z/OS commands

301

Note: The DEVSERV command is one of the very few commands that actually performs an I/O operation against the device or devices specified. Most display commands do not actually perform a real I/O operation and report back the state of specific fields in the UCB and UCW (subchannel). Example F-17 shows the DS P,dddd,nnn command.
Example: F-17 Output of one type of DEVSERV command: DS P,dddd,nnn

DS P,D000,4 IEE459I 18.33.40 DEVSERV PATHS 919 UNIT DTYPE M CNT VOLSER CHPID=PATH STATUS RTYPE SSID CFW TC DFW PIN DC-STATE CCA DDC CYL CU-TYPE D000,33909 ,O,000,Z1BRZ1,50=- 54=- 58=+ 5C=+ 7E=+ 82=+ 2107 89E0 Y YY. YY. N SIMPLEX 00 00 10017 2107 D001,33909 ,O,000,Z18DL2,50=- 54=- 58=+ 5C=+ 7E=+ 82=+ 2107 89E0 Y YY. YY. N SIMPLEX 01 01 10017 2107 D002,33909 ,O,000,Z18RB1,50=- 54=- 58=+ 5C=+ 7E=+ 82=+ 2107 89E0 Y YY. YY. N SIMPLEX 02 02 10017 2107 D003,33909 ,A,002,TOTD94,50=- 54=- 58=+ 5C=+ 7E=+ 82=+ 2107 89E0 Y YY. YY. N SIMPLEX 03 03 10017 2107 ************************ SYMBOL DEFINITIONS ************************ O = ONLINE + = PATH AVAILABLE - = LOGICALLY OFF, PHYSICALLY OFF The fields reported by the DASD DEVSERV command are listed here. DTYPE = 33909 This is the device type of the device reported by DEVSERV command. In this case, device D000 reported back as a 3390 model 9. RTYPE = 2107 This is the real (true) device type of the device reported by the DEVSERV command. 2107 is the true device type where the DEVSERV command has been executed. M=O This field represents the Device UCB status: A - Allocated F - Offline M - Mount Pending O - Online P - Offline Pending N - Not Allocated CNT This field indicates the number of data sets allocated on this volume. SSID This field indicates the subsystem ID of the device CU. VOLSER This field indicates the VOLSER of the device pointed to by the DEVSERV command.

302

FICON Planning and Implementation Guide

CFW Cache Fast Write, which indicates the status of the CFW: Y - CFW is Active N - CFW is Inactive S - CFW is Suspended with Pinned Data TC (c,d) C=Device Cache; D=Subsystem Cache These two characters indicate the Device Cache and Subsystem Cache status: Y - Active N - Inactive A - Pending Active F - Pending Inactive M - Disabled for maintenance (this will override other statuses) P - Pending Inactive - Destage in Progress S - CU is running in Single Cluster mode (half of the cache in one cluster is not in use) T - Terminated due to a Subsystem Cache Error DFW (e,f) E=DASD Fast Write; F=NVS Status These two characters indicate the DASD Fast Write and the NVS status: Y - Active/Available N - Inactive/Unavailable F - NVS Pending Unavailable/Disable - Destage failed I - DFW - Deactivate Pending - Destage in Progress I - NVS - Battery Defective P - NVS Pending Unavailable - Destage in Progress or has failed S - DFW is temporarily Suspended with Pinned Data U - DFW Deactivate Pending with Destage failed U - NVS Terminated due to error PIN (p) This single character indicates whether pinned data exists: N - No Pinned Data Y - Pinned Data in Cache or NVS - CFW/DFW allowed S - Retriable Pinned Data in Cache/NVS - CFW/DFW temporarily suspended DC-STATE This field reports the Dual Copy current status: SIMPLEX (Simplex: Not Duplex-Pair) PRIMARY (Primary Device of Active Duplex-Pair) SECONDARY (Secondary Device of Active Duplex-Pair) PRI-PNDG (Establish/Copy in Progress as Primary) SEC-PNDG (Establish/Copy in Progress as Secondary) PRI-SDPL (Suspended, current and original Primary) - DDC same) SEC-SDPL (Suspended, current and original Secondary) - DDC same) PRI-SSEC (Suspended, was Secondary and now Primary - DDC Swapped) SEC-SPRI (Suspended, was Primary and now Secondary - DDC Swapped) SPARE (RAMAC: RDC byte 57=33/34 and SNSS byte 36 bit 4,5=01/10) SPAR-PNDG (Spare Device being copied or has been copied and Copy-Back is pending) SPAR-BRKN (Broken Device copied to Spare device) PPRIMARY (PPRC Primary) PPRI-PNDG (PPRC Primary Pending) PPRI-FAIL (PPRC Primary Fail) PPRI-SUSP (PPRC Primary Suspended) PSECONDRY (PPRC Secondary) PSEC-PNDG (PPRC Secondary Pending) PSEC-FAIL (PPRC Secondary Fail) PSEC-SUSP (PPRC Secondary Suspend)

Appendix F. Useful z/OS commands

303

MIRR-OPER|PEND|FAIL (mirroring status per SNSS byte 26 bit 6, 7) CCA-xx xx - This indicates the Channel Connection address: The xx value should match UNITADD-XX in the IOCP. DDC-yy yy - This indicates the Device-to-Director Connection address. CYL This field reports the number of cylinders of the subject device, when applicable. CU-TYPE This field reports the control unit type of the subject device.

304

FICON Planning and Implementation Guide

Appendix G.

Adding FICON CTC connections

This appendix contains recommendations for adding CTC connections to a FICON environment. You can choose one of the following FICON channel CTC configurations: One FICON channel in a point-to-point, switched point-to-point, or cascaded FICON Directors configuration Two FICON channels in a point-to-point, switched point-to-point, or cascaded FICON Directors configuration The channel operates in full duplex mode, sending and receiving data concurrently. The FICON channel CTC communication does not require a pair of channels because it can communicate with any FICON channel that has a corresponding FCTC control unit defined. Although a single FICON native channel per server can provide CTC connections across multiple LPARs, for large FICON configurations we recommend using at least one pair of FICON native channels. Using a pair of FICON native channels allows the installation to maintain the same CTC device definition methodology for FICON as was previously used for ESCON. But the FICON channels can support the definition of FCTC control units as well as other I/O control units, such as disk and tape, at the same time. The FICON channel at each end of the FICON CTC connection, supporting the FCTC control units, can also communicate with other FICON native control units, such as disk and tape connected via a FICON Director. Because the FICON native channel supports a larger number of devices, installations with a high number of logical partitions in an FCTC complex are easier to design and define. With the System z, a single FICON CHPID can be used simultaneously to talk to another FICON CHPID and any other native FICON controller. In a mixed CTC/controller configuration, performance levels can be kept at a premium if the channel utilization is kept at recommended levels (under 50%).

Copyright IBM Corp. 2005, 2006, 2009, 2006. All rights reserved.

305

FCTC device numbering scheme

There are a number of schemes that can be used for developing the device numbers for the FICON CTC devices. Defining a FICON CTC configuration requires an understanding of CTC operations and the different methods available for defining the connections. It is also necessary to take into consideration the subsystem requirements for using CTC connections. As the number of images in a complex grows, the CTC definition process becomes increasingly complex. Using the FICON sender/receiver device numbering scheme is the recommended approach. Apart from ease of migration, operations and systems programmers can easily identify the use (send or receive) and target system for any given CTC device. It also simplifies the CTC definition process in that the same operating system CTC device definitions can be used by all z/OS operating systems, independent of the image and processor where that operating system runs. Using two FICON (FC) channels per server, define all send FCTC control units and devices on one FICON (FC) channel, and define all receive FCTC control units and devices on a second FC channel. The method makes use of the 4-digit z/Architecture device number where: First digit An even hexadecimal number is used for the send CTC control unit and device. An odd hexadecimal number is used for the receive CTC control unit and device. Second and third digits These digits represent an assigned CTC image-ID (that is, assigned on paper) for the LPAR image, and can be any unique value within the CTC complex. This CTC image-ID is used as a target identifier (for the image) that you use when defining how all other images access this image. Fourth digit This digit is used to indicate whether the CTC connection is the primary or alternate (backup for availability purposes). A value of 0 to 7 is used for the primary CTC connection devices, and a value of 8 to F is used for the alternate CTC connection devices. To fully exploit the full duplex data flow capabilities of the FICON channels, a further performance enhancement can be made. However, this requires a large amount of customization and may not be practical with very large configurations. Using two FICON channels per server, evenly spread the send CU/device definitions across the two FC channels, and evenly spread the receive CU/device definitions across the two FC channels. We do not recommend using the matching device number scheme because it does not enable the operator or systems programmer to identify the use or target system for a given CTC device number. It also makes it difficult to relocate operating systems between logical partitions on the same server or a different server. For more details, refer to: FICON CTC Implementation, REDP-0158

306

FICON Planning and Implementation Guide

Related publications
The publications listed in this section are considered particularly suitable for a more detailed discussion of the topics covered in this book.

IBM Redbooks
For information about ordering these publications, see How to get Redbooks on page 309. Note that some of the documents referenced here may be available in softcopy only. IBM System z Connectivity Handbook, SG24-5444 Getting Started with the IBM 2109 M12 FICON Director, SG24-6089 Implementing an IBM/Brocade SAN with 8 Gbps Directors and Switches, SG24-6116 Implementing the Cisco MDS9000 in an Intermix FCP, FCIP, and FICON Environment, SG24-6397 FICON Implementation Guide, SG24-6497 IBM System Storage DS8000 Architecture and Implementation, SG24-6786 Getting Started with the McDATA Interpid FICON Director, SG24-6857 Getting Started with the INRANGE FC/9000 FICON Director, SG24-6858 IBM Tivoli System Automation for z/OS Enterprise Automation, SG24-7308 IBM System z10 Enterprise Class Technical Introduction, SG24-7515 IBM System z10 Enterprise Class technical Guide, SG24-7516 IBM/Cisco Multiprotocol Routing: An Introduction and Implementation, SG24-7543 IBM System Storage/Brocade Multiprotocol Routing: An Introduction and Implementation, SG24-7544 Implementing an IBM/Cisco SAN, SG24-7545 IBM System z10 Enterprise Class Configuration Setup, SG24-7571 FICON CTC Implementation, REDP-0158 Disk Storage Access with DB2 for z/OS, REDP-4187 How Does the MIDAW Facility Improve the Performance of FICON Channels Using DB2 and Other Workloads?, REDP-4201 Multiple Subchannel Sets: An Implementation View, REDP-4387 Cisco FICON Basic Implementation, REDP-4392

Other publications
These publications are also relevant as further information sources: z/OS MVS Diagnosis: Reference, GA22-7588 z/OS MVS Diagnosis: Tools and Service Aids, GA22-7589

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

307

System z Planning for Fiber Optic Links (ESCON, FICON, InfiniBand, Coupling Links, and Open System Adapters), GA23-0367 IBM System Storage SAN768B Installation, Service, and User's Guide, GA32-0574 IBM System Storage SAN384B Installation, Service, and User's Guide, GA52-1333 IBM System Storage DS8000 Introduction and Planning Guide, GC35-0515 FICON Express2 Channel Performance Version 1.0, GM13-0702 System z10 Enterprise Class System Overview, SA22-1084 z/OS MVS System Commands, SA22-7627 z/OS MVS System Messages SA22-7637 z/Architecture, Principles of Operation, SA22-7832 System z Input/Output Configuration Program Users Guide for ICP IOCP, SB10-7037 System z10 Processor resource/Systems Manager Planning Guide, SB10-7153 DFSMS Storage Administration Reference, SC26-7402 HMC Operations Guide, SC28-6830 HMC Operations Guide, SC28-6873 System z10 Support Element Operations Guide, SC28-6879 HCD Users Guide, SC33-7988 z/OS RMF Users Guide, SC33-7990 z/OS RMF Report Analysis, SC33-7991 z/OS RMF Performance Management Guide, SC33-7992 z/OS RMF Programmers Guide, SC33-7994 System z Maintenance Information for Fiber Optic Links (ESCON, FICON, Coupling Links, and Open System Adapters), SY27-2597 IBM Fiber Optic Cleaning Procedure, SY27-2604 z/OS RMF Reference Summary, SX33-9033 Brocade Fabric OS Administrators Guide, 53-1001185 High Performance FICON for System z Technical Summary for Customer Planning, ZSW03058USEN Performance Considerations for a Cascaded FICON Director Environment Version 0.2x, Richard Basener and Catherine Cronin
http://www-03.ibm.com/servers/eserver/zseries/library/techpapers/pdf/gm130237.pdf

IBM System z9 I/O and FICON Express4 Channel Performance, ZSW03005USEN IBM System z10 I/O and High Performance FICON for System z Channel Performance, ZSW03059USEN IBM System Storage SAN768B, TSD03037USEN IBM System Storage SAN768B Fiber Backbone Interoperability Matrix Cisco MDS 9506 for IBM System Storage, TSD00069USEN Cisco MDS 9509 for IBM System Storage, TSD00070USEN Cisco MDS 9513 for IBM System Storage, TSD01754USEN Cisco MDS 9506, 9509, 9513 for IBM System Storage Directors Interoperability Matrix

308

FICON Planning and Implementation Guide

Online resources
These Web sites are also relevant as further information sources: Fibre Channel standard Web site http://www.t11.org Brocade Communications Systems, Inc. Web site http://www.brocade.com Cisco Systems, Inc. Web site http://www.cisco.com

How to get Redbooks

You can search for, view, or download Redbooks, Redpapers, Technotes, draft publications and Additional materials, as well as order hardcopy Redbooks publications, at this Web site: ibm.com/redbooks

Help from IBM

IBM Support and downloads ibm.com/support IBM Global Services ibm.com/services

Related publications

309

310

FICON Planning and Implementation Guide

Index
A
active CSS 108, 125, 143, 290 Active Zone Config 188 Advanced Performance Tuning Policy 176 American National Standard Institute (ANSI) 4 attached FICON Director error condition 44 CHPID Mapping Tool (CMT) 103, 120, 138, 255 CHPID number 113, 131, 149, 209, 257 CHPID statement 38, 103, 120, 138, 257 CHPID type FC 10, 26 FCP 26 FCV 9, 29 CHPIDS 12, 26, 103, 119, 138, 268 Cisco MDS 9506 44 Cisco MDS 9509 44 Cisco MDS 9513 44 CNTLUNIT CUNUMBR 104, 120, 139 CNTLUNIT statement 104, 120, 139, 257 CNTUNIT statement 122 command mode 14 CCW operation 20 channel program 21 command response (CMR) 204 CONFIG command 108, 125, 143 connecting FICON Director Domain ID 42 control processor (CP) 45, 153 Control unit logical link address 298 logical-path establishment 14 login 14 node-identifier acquisition 14 serial number 300 state-change registration 14 type 300 control unit 271 N_Port address 39 node descriptor 24 control unit (CU) 1, 8, 14, 34, 99, 115, 133, 185, 206, 218, 255, 272, 295 Control Unit Port (CUP) 39, 152 control units as specified 138 D000 120, 139 CPC icon 268 CRC errors 91 CSS 2 103, 120, 138, 273 CSS.CHPI D 2.7E 129, 147 2.7F 111 CU 4044 105, 123, 140 Cyclical Redundancy Checking (CRC) 41

B
buffer credit 23, 39, 152, 211 build IOCDS 107, 124, 142 Business Class (BC) 28, 289

C
Cache Fast Write (CFW) 302 cascaded FICON Director Additional configuration steps 164 CCW 17 other parameters 17 CCW chaining 204 CCW execution 20 control unit 20 Central Processor Complex (CPC) 107, 124, 202, 255, 266, 288 CF CHP 109, 126, 144, 297 changing passwords on Director 162 Channel feature 26 Channel Information PD panel 272 window 112, 130, 148 channel path functional details 127 shared definition 122 channel path (CP) 11, 14, 99, 115, 119, 133, 220, 235, 255, 266 Channel Path Identifier 12 channel paths, CUs (CPC) 138, 219 Channel Problem Determination 272, 275, 277, 279281, 284 Channel Subsystem channel path 257 logical control units 291 logical partition 289 not-operational reason 237 Channel Subsystem (CSS) 288 Channel-to-Channel (CTC) 11, 26 CHPID 100, 116, 135, 220, 270, 295 CHPID 7E 103, 116, 135, 273 current status 126 direct connection 105 CHPID detail 111, 129, 147

D
DASD Fast Write (DFW) 302 Data Center Fabric Manager (DCFM) 47, 152155, 226, 246 DCFM server 153, 229, 246 IP address 162

Copyright IBM Corp. 2005, 2006, 2009. All rights reserved.

311

ODBC password 162 welcome window 162 defined channel path path status 301 Device Number 109, 127, 145, 220, 275, 292, 306 option Search 277 Device Status (DS) 109, 127, 145, 280, 293 device type 110, 128, 146, 224, 257, 292 cam display information 301 DEVSERV command 292 other uses 301 Devserv command 300 Director port (DP) 208 Domain ID 37, 135, 152, 257 DS P 225, 300 DS8000 storage control unit 108, 125, 143 controller 102, 118, 137 device 100, 116, 134 system 100, 116, 134 Dynamic Load Sharing (DLS) 49, 152 Dynamic Path Selection (DPS) 49, 233 Dynamic Pathing Switch (DPS) 231

E
E_Port 6, 179, 183 Enterprise Class (EC) 244, 289 Enterprise Fabric Connectivity Manager (EFCM) 155 Environmental Record, Editing, and Printing (EREP) 222 ESCON Director 39 ESCON solution 9 Establish Logical Path (ELP) 15 Extended Link Service (ELS) 19 Extended Subchannel Logout Data (ESLD) 223, 233

F
F_Port 6 fabric 6 Fabric Manager (FM) 54 fabric port (F_PORT) 6, 54, 179 FC link 6, 16, 35 FC-SB-2, SB-3, SB-4 Web site 5 FCTC control unit 11, 105, 140, 305 Feature Check 100, 116, 135 Feature code 26, 46, 136, 249 fiber optic cable 6, 38, 101102, 134135 CHPID 82 105 Fibre Channel adapter 17 Arbitrated Loop 6 architecture 1, 39 bit error 223 data traffic 55 fabric 39 link 18 logical switch 197 Physical 4 Physical and Signaling Standard 8 physical framing 5

physical interface 4 port 222 Protocol 10, 19, 43, 222 Protocol traffic 26 Routing 48 Security Protocol 54 Single Byte Command Sets-3 13 Single Byte Command Sets-4 13 Single-Byte-3 8 Single-Byte-4 8 standard 4 Switch Fabric and Switch Control Requirements 8 Fibre Channel (FC) 3, 13, 34, 197, 202 FICON 1, 3, 13, 3334, 115, 151, 201, 255, 265, 305 FICON advantage 9 FICON channel 5, 14, 34, 99, 115, 133, 185, 220221, 247, 257, 268, 271 CHPID type 257 different flavors 203 fiber links 108 full duplex data flow capabilities 306 IOS051I message 232 PEND time ends 204 physical connectivity 37 physical transmission path 36 topologies 34 FICON channel (FC) 210 FICON Channel-to-Channel (FCTC) 121, 140 FICON Director 1, 10, 15, 34, 104, 116, 133, 150152, 203, 218, 244, 251, 257, 272, 297, 305 arbitrary number 257 Basic functions 37 control unit 35 cooling capability 37 CP cards 154 CUP port 121 default buffer credit assignment 248 detailed configuration 210 Domain ID range 38 fabric binding database 42 FC links 35 Fiber optic cables 144 IBM qualification testing 44 in-band management 39 ISL port 190 Other settings 125 physical port 11 port address 185 switch module 37 Switch number 282 FICON environment 4, 33, 57, 97, 99, 119, 133, 175, 199, 217, 225, 244, 287, 305 FICON Express 30 FICON Express2 3, 16, 101, 118, 136 FICON Express4 10KM LX 26 10KM LX feature 27 4 KM LX feature 28 4KM LX 26 CHPIDs 27

312

FICON Planning and Implementation Guide

feature 9, 26 SX 26 FICON Express4 10KM LX 27 FICON Express4 4KM LX 28 FICON Express4 SX 28 FICON Express4-2C 4 KM LX 28 FICON Express4-2C SX 28 FICON Express8 3, 16, 100, 116, 134, 244 FICON FC-SB3 and SB-4 Web site 19 FICON feature 8, 13, 136, 247 FICON Purge Path Extended 222 FL_Port 6

G
G_Port 6 Gbps 16, 37, 181, 244 graphical user interface (GUI) 54, 154

Input/Output architecture 11 Input/Output Supervisor (IOS) 16, 126 Insistent Domain (ID) 152 Inter-Chassis Link (ICL) 48 Inter-Switch Link (ISL) 35, 38, 135, 152, 175 Invalid Transmission Words 91 IOCP User 256 IODEVICE Address 104, 121, 139 IODEVICE statement 103, 119, 138, 258 PARTITION keyword 105 IODF data (ID) 107, 123, 141, 290 IOS 220 IOS command 108, 125, 143 IP address 55, 152, 154, 246, 260 ISL Trunking 54 ISLs 35, 135, 152, 245

K
km transceivers 28

H
Hardware Configuration Definition (HCD) 8, 38, 255, 289 Hardware Configuration Manager (HCM) 255, 266 Hardware Management Console 110, 128, 146, 219, 265, 289 Hardware System Area (HSA) 14, 103, 119, 138, 289 HCD User 103, 119, 138, 256 Host Bay Adapter (HBA) 221, 297 HYPERPAV ALIAS (HA) 299

L
L_Port 6 LC Duplex 27 LC Duplex connector 27, 101, 117, 135 Link-Incident-Record Registration (LIRR) 14 logical control unit (LCU) 100, 117, 135, 207, 247, 281, 291 logical partition channel subsystem 292 high number 305 logical partition (LP) 12, 51, 105106, 122, 141, 220, 257, 275, 289, 305 Logical Path 15, 35, 221, 276 Logical switch scalability.Partition Brocade SAN768B 50 logical switch 50, 152, 170, 229, 247 management isolation 50 port numbering 173 logical switch types 50 Long Wave (LW) 46 Longitudinal Redundancy Checking (LRC) 41 longwave laser 101, 117, 134135 Loss of Sync 91 lossless DLS 175, 248 LPAR A23 105, 122, 140, 288 CU 5034 105 LPARs 104, 116, 134

I
I/O Configuration Data 255 Data Set 99, 115, 133, 255, 267, 289 Program 255, 282, 289 I/O definition 103, 119, 137 file 106107, 124, 142, 206, 256 I/O definition file (IODF) 103, 119, 138 I/O device 11, 17, 255, 298 unit addresses 258 I/O information 206 I/O operation 12, 14, 34, 206, 231, 272, 292 exchange pair 19 FICON architecture 16 MIH value 22 I/O performance measurement points 203 I/O priority queuing (IOQ) 213 I/O processor (IOP) 209 I/O request 16, 42, 209, 235236, 289 FC-FS FC-2 frame 18 I/O specification 103, 119, 138 IBM RMF Web site 203 IBM SAN384B 44 IBM SAN768B 44 IBM System Storage Web site 93 in-order delivery (IOD) 49, 152 Input Output Configuration Program (IOCP) 8, 38, 99, 115, 133 Input/output (I/O) 3, 13, 203, 256

M
MIF ID 257, 273, 289 Missing Interrupt Handler (MIH) 21, 233 MM 62.5 30 mode conditioning patch (MCP) 2728 MULTIPLE IMAGE FACILITY (MIF) 288 Multiple Image Facility (MIF) 12

N
N_Port 4, 14 Index

313

Native FICON 26, 103, 120, 138, 210 average number 210 NL_Port 6 node 6 Node Descriptor (ND) 16, 163, 222, 274, 288 Node-Element Descriptor (NED) 296

RMF 204

S
SCP CU Neodymium 110, 127, 146, 237, 296 Secure File Transfer Protocol (SFTP) 54 selected central processor complex data sets 256 Setting the display to hex 163 Short Wave (SW) 46 Simple Network Management Protocol (SNMP) 37 single mode (SM) 47, 101, 117, 134 Single Object Operation (SOO) 110, 128, 146, 266 Small Form Factor Pluggable (SFP) 37 State-Change Registration (SCR) 14 Storage Area Network (SAN) 43 storage device 34, 100, 116, 134, 222 Store System Information (STSI) 288 subchannel 276, 279 Subchannel Logout Handler (SLH) 223 subchannel number 275 subchannel set 12, 108, 125, 143, 289 Support Element (SE) 240, 255, 265 switch address 11, 37, 117, 135, 174, 257 Switch Connection Control (SCC) 153, 192 switched fabric 11 System Activity Display (SAD) 201 System Control Program (SCP) 11, 298 System Information (SI) 288 System Management Facility (SMF) 203 System z 3, 13, 33, 100, 116, 244, 255, 266, 305 High Performance FICON 26 System z environment 8 System z server 134 System z10 8, 13, 244, 289 Systems Automation (SA) 39

O
operating system 17, 39, 100, 116, 134, 277, 295, 306 FICON Express support 137 FICON Express8 support 102 Operation Request Block additional parameters 17 Operation Request Block (ORB) 17

P
Parallel Access Volume (PAV) 8, 212 Path Available Mask (PAM) 277 PATH keyword 103, 120, 138, 257 Path Not Operational Mask (PNOM) 231 Path Operational Mask (POM) 231, 277, 297 PCHID 5A2 116, 135, 273 Essential information 129 owning image 129 PCHID 5A3 100 Essential information 111 owning image 111 PCHID detail 110, 128, 146 PCHID number 270 PCHIDs 268 performance FICON 9, 13 performance monitoring 203 point-to-point configuration 11, 100, 116, 257 point-to-point topology 35, 99, 115, 255 FICON environment 99 port 6 port address 8, 38, 139, 177, 185, 247, 280 port fencing 91 port number 38, 100, 117, 120, 135, 149, 161, 197, 261 port type 5, 16, 152, 179, 214 Port-Based Routing (PBR) 152 Power-on Reset (POR) 103, 119, 138 PPRC Secondary Suspend (PSEC-SUSP) 303 problem determination (PD) 108, 126, 144, 217, 265 Processor Card (PC) 160 Purge Path Extended (PPE) 222, 285

T
Tag 274 TE port 54 terminology 4 TI zone 190 Traffic Isolation (TI) zone 48 Traffic Isolation Routing feature 48 transceiver 126, 144 Transport Control Word (TCW) 16 Transport Indirect Data Address Word (TIDAW) 22 transport mode 14 I/O operation 14 Trivial file transfer protocol (TFTP) 43 trunk or ISL (TI) 152

Q
QoS SID/DID pairs 49 Quality of Service (QoS) 48, 152

U
UCB Logical Path Mask (UCBLPM) 297 unit address (UA) 9, 18, 258, 275 Unit Control Block (UCB) 17, 300 Unit Control Word (UCW) 291 Unrepeated distance 9, 27 Upper Level Protocol (ULP) 5

R
Redbooks Web site 309 Contact us xiii Registered State Change Notification (RSCN) 177 Remote Node Identification (RNID) 274, 298 Resource Measurement Facility (RMF) 202204

314

FICON Planning and Implementation Guide

User Interface (UI) 266

V
Virtual Fabric 51, 170 Virtual Fabrics feature 50 Virtual ISL 49 Virtual Storage Area Network (VSAN) 54

W
Wave Division Multiplexing (WDM) 196 working CHPID normal indication 271 World Wide Name 7 World Wide Node Name 16, 113, 131, 149, 193 World Wide Node (WWN) 113, 131, 149, 169 World Wide Node Name (WWNN) 42 World Wide Node_Name (WWNN) 7 World Wide Port Name 16, 113, 131, 149, 222 World Wide Port (WWP) 113, 131, 149 World Wide Port_Name (WWPN) 7

Z
z10 BC 12, 28, 291 z10 EC 12, 30, 247, 291 FICON connectivity 247 z10 server 100, 116, 134 configuration tasks 110 corresponding LPAR 140 designated ports 128 fiber optic cable link 113, 149 FICON Express8 feature 102 PCHID 5A2 128 PCHID 5A3 110 standard feature 100 zHPF feature 101, 117, 135, 248 zHPF on System (Z/OS) 100, 116, 134, 218 Zone Config 187

Index

315

316

FICON Planning and Implementation Guide

FICON Planning and Implementation Guide

FICON Planning and Implementation Guide

FICON Planning and Implementation Guide

(0.5 spine) 0.475<->0.873 250 <-> 459 pages

FICON Planning and Implementation Guide

FICON Planning and Implementation Guide

FICON Planning and Implementation Guide

Back cover

FICON Planning and Implementation Guide

Topologies, concepts, and terminology Planning, implementation, and migration guidance Realistic examples and scenarios

This IBM Redbooks publication covers the planning, implementation, and management of IBM System z FICON environments. It discusses the FICON and Fibre Channel architectures, terminology, and supported topologies. The book focuses on the hardware installation and the software definitions that are needed to provide connectivity to FICON environments.You will find configuration examples required to support FICON control units, FICON Channel-to-Channel (FCTC), and FICON Directors. It also discusses utilities and commands that are useful for monitoring and managing the FICON environment. The target audience for this document includes IT Architects, data center planners, SAN administrators, and system programmers who plan for and configure FICON environments. The reader is expected to have a basic understanding of IBM System z10 and IBM System z9 hardware, HCD or IOCP, as well as a broad understanding of the Fibre Channel and FICON architectures.

INTERNATIONAL TECHNICAL SUPPORT ORGANIZATION

BUILDING TECHNICAL INFORMATION BASED ON PRACTICAL EXPERIENCE

IBM Redbooks are developed by the IBM International Technical Support Organization. Experts from IBM, Customers and Partners from around the world create timely technical information based on realistic scenarios. Specific recommendations are provided to help you implement IT solutions more effectively in your environment.

For more information: ibm.com/redbooks

SG24-6497-02 ISBN 0738433284