FOS_v8.2.3e_releasenotes_v7.0
FOS_v8.2.3e_releasenotes_v7.0
3e2
Fabric OS v8.2.3e Release Notes
Release Notes
Version 7.0
Broadcom FOS-823e-RN
December 17th, 2024
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Copyright © 2024 Broadcom. All Rights Reserved. The term “Broadcom” refers to Broadcom Inc. and/or its subsidiaries.
For more information, go to www.broadcom.com. All trademarks, trade names, service marks, and logos referenced herein
belong to their respective companies.
Broadcom reserves the right to make changes without further notice to any products or data herein to improve reliability,
function, or design. Information furnished by Broadcom is believed to be accurate and reliable. However, Broadcom does
not assume any liability arising out of the application or use of this information, nor the application or use of any product or
circuit described herein, neither does it convey any license under its patent rights nor the rights of others.
The product described by this document may contain open source software covered by the GNU General Public License
or other open source license agreements. To find out which open source software is included in Brocade products or to
view the licensing terms applicable to the open source software, please download the open source attribution disclosure
document in the Broadcom Support Portal. If you do not have a support account or are unable to log in, please contact
your support provider for this information.
Use of all versions of Brocade’s Fabric OS is subject to the terms and conditions of the Brocade Fabric Operating System
and Feature Licenses and License Keys End User License Agreement, effective October 1, 2019, as amended by
Brocade from time to time. It is the user’s responsibility to understand and comply with the terms of the EULA. By
downloading, installing, using, posting, distributing or otherwise making available FOS, you agree to be bound on an
ongoing basis by the EULA as updated by Brocade from time to time.
Broadcom FOS-823e-RN
2
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Table of Contents
Broadcom FOS-823e-RN
3
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Broadcom FOS-823e-RN
4
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Broadcom FOS-823e-RN
5
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Chapter 1: Preface
If you purchased Brocade product support from a Broadcom OEM/solution provider, contact your OEM/solution provider
for all your product support needs.
OEM/solution providers are trained and certified by Broadcom to support Brocade products.
Broadcom provides backline support for issues that cannot be resolved by the OEM/solution provider.
Brocade Supplemental Support augments your existing OEM support contract, providing direct access to Brocade
expertise. For more information on this option, contact Broadcom or your OEM.
For questions regarding service levels and response times, contact your OEM/solution provider.
To expedite your call, have the following information immediately available:
General Information:
Technical support contract number, if applicable.
Switch model.
Switch operating system version.
Error numbers and messages received.
supportSave command output and associated files.
For dual-CP platforms running Fabric OS 6.2 and above, the supportSave command gathers information from both
CPs and any AP blades installed in the chassis.
Detailed description of the problem, including the switch or fabric behavior immediately following the problem and any
specific questions.
Description of any troubleshooting steps already performed and the results.
Serial console and telnet session logs.
Syslog message logs.
Broadcom FOS-823e-RN
6
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
FT00X0054E9
Broadcom FOS-823e-RN
7
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
This section outlines how to locate and download Brocade product manuals and release notes from myBroadcom.
Although the illustrations show Fibre Channel and Fabric OS (FOS), they work for all Brocade products and operating
systems.
Complete the following steps to locate your product manuals on Broadcom.com.
1. Go to https://www.broadcom.com.
2. Enter the product name or the software version number in the Search box.
For example, the following search is for software and documentation files for software version 8.2.
Broadcom FOS-823e-RN
8
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Chapter 3: Overview
Broadcom FOS-823e-RN
9
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
NOTE Embedded switch platforms (6543, 6547, 6548, 6558) do not require a FOS upgrade certificate.
Broadcom FOS-823e-RN
10
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Broadcom FOS-823e-RN
11
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
6.3.1 MAPS
FOS v8.2.3c provides MAPS support for monitoring of the following SmartOptics:
8G DWDM 80 km [ALL_80Km_8GELWL_SFP]
16G DWDM 40 km [ALL_40Km_16GELWL_SFP]
32G DWDM 40 km [ALL_40Km_32GELWL_SFP]
Note: Downgrade from FOS v9.x that support this feature to FOS v8.2.3c (or later) is blocked.
Broadcom FOS-823e-RN
12
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
FOS-834912 SANnav reports error: "Registration for telemetry profile 'xxx' has failed and Switch panic after cald
termination.
FOS-836031 Switch panic after FDMI daemon terminated.
FOS-836265 During code upgrade from FOS v8.2.1x to FOS v8.2.3x, FOS cannot completely be brought up due to
cald core dumps. User observes the switch hanging if cal.esrs configuration keys are present.
For a full list of fixes, see Defects.
Broadcom FOS-823e-RN
13
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Broadcom FOS-823e-RN
14
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
aaaConfig
New option -tls_mode added to specify the mode of the connection with the LDAP Server.
Valid options include the following:
starttls
Initiates LDAP connection with StartTLS. The default port is 389.
ldaps
Initiates LDAPS connection. The default port is 636.
Synopsis
aaaconfig --add | --change server -conf radius | ldap | tacacs+
[-p port] [-d domain] [-t timeout] [-s secret]
[-a chap | pap | peap-mschapv2] [-e -encr_type none | aes256]
[-tls_mode starttls | ldaps]
NOTE In a chassis, both CPs must be loaded with FOS v8.2.3a before configuring the LDAPS protocol.
Before configuring LDAPS, verify that both CPs are running FOS v8.2.3a; otherwise, the command
results in a no-op operation if the standby CP does not support LDAPS.
When LDAP TLS Mode configuration is set to LDAPS in FOS v9.0.1x, it's not recommended to
downgrade to FOS v8.2.3a.
If downgrade to FOS v8.2.3a from FOS v9.0.1x is necessary, first configure LDAP TLS Mode to
STARTTLS before downgrading to FOS v8.2.3a to avoid potential user login issues.
portCfgLosstov
New options –dwdmloyncon and –dwdmlosyncoff added to enable or disable the configuration for DWDM lossSync
fixed speed port.
Synopsis
portcfglosstov port [-dwdmlosyncon | -dwdmlosyncoff]
Broadcom FOS-823e-RN
15
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Broadcom FOS-823e-RN
16
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Broadcom FOS-823e-RN
17
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
11.1 Hardware
The following sections list new hardware introduced with this release and hardware that is no longer supported with this
release.
Broadcom FOS-823e-RN
18
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
FOS 8.2.1 introduces counterfeit license protection (CLP) on the Brocade 6505, 6510, G610, G620, and 7810 to prevent
misuse of the licenseAdd command. CLP performs the following actions based on the number of times that the
licenseAdd command is invoked:
RASLOG WARNING message when 25 invalid licenseAdd operations are detected within a 24-hour period.
RASLOG CRITICAL message when 40 invalid licenseAdd operations are detected within a 24-hour period.
Shuts down a switch when 50 invalid licenseAdd operations are detected within a 24-hour period.
A switch that is shut down due to CLP detection will no longer be able to boot to a functional state. Customers may
request a replacement unit from their support provider.
FOS 8.2.1 expands REST API support from that supported in FOS 8.2.0a with the following new features and
corresponding RESTCONF modules:
Physical chassis, FRU, and optical transceiver: brocade-chassis, brocade-fru, and brocade-media
Switch configuration: brocade-fibrechannel-configuration
SupportSave: brocade-operation-supportsave and brocade-operation-showstatus
Trunking: brocade-fibrechannel-trunk
RASLOG and syslog configuration: brocade-logging
System security: brocade-security
System time zone and time server: brocade-time
Monitoring and Alerting Policy Suite (MAPS): brocade-maps
FOS 8.2.1 enhances the following existing RESTCONF modules:
brocade-fibrechannel
brocade-fibrechannel-switch
brocade-access-gateway
For a detailed description of the new modules and attributes, refer to the Brocade Fabric OS REST API Reference Manual
for FOS 8.2.1.
FOS 8.2.1 implements hardware-based ingress rate limiting on the management interface of X6 directors to prevent
denial of service (DOS) attacks through the Ethernet management interface. When this feature is enabled, hardware
performs ingress rate limiting when a DOS attack on the Eth0 or Eth3 management interface is detected. Normal
management application traffic from Web Tools and SNMP is not affected when there is no DOS attack. These
applications may appear sluggish when a DOS attack is detected.
19
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Miscellaneous
FOS 8.2.1 adds MAPS monitoring of Cisco UCS Fabric Interconnect (FI) connections to Brocade switches in NPV mode.
MAPS monitors the distribution of UCS servers over the uplinks between FI and Brocade switches to remain balanced
over time. When MAPS detects an imbalance, it alerts the SAN administrator of the condition or triggers an automatic
rebalance action.
11.2.2.1.2 Miscellaneous
FOS 8.2.1 replaces the default rule names in FOS 8.2.0x detailed in the following table with the new default rule names:
NOTE The default rules will be automatically converted to the new rule names during firmware upgrade from
FOS 8.2.0x to FOS 8.2.1 or later. However, user-defined rules of the IP_EXTN_FLOW monitoring
system must be changed manually to follow the new rule name format.
FOS 8.2.1 changes the alert severity from ERROR to CRITICAL for the following default rule:
defALL_E_PORTSC3TXTO_20 ALL_E_PORTS(C3TXTO/MIN>20) RASLOG
FOS 8.2.1 changes the alert severity from WARNING to ERROR for the following default rule:
defALL_FANFAN_STATE_FAULTY ALL_FANS(FAN_STATE/NONE==FAULTY) RASLOG
FOS 7.4 and later support Slow-Drain Device Quarantine (SDDQ) to mitigate congestion due to slow-drain devices
connected to switches in native mode. FOS 8.2.1 enhances SDDQ support on AG to be able to quarantine slow-drain
devices connected to Access Gateway.
FOS 8.2.1 adds support of the portCfgNpivPort command for a switch in AG mode. With this support, administrators
can enable or disable NPIV mode for a port when a switch is in Access Gateway mode.
Broadcom FOS-823e-RN
20
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
FOS 8.2.1 adds support of the Get Zone Set (GZS) and Get Active Zone Set (GAZS) commands under FC-GS-8 for zone
servers. GZS queries for Zone Set Database, whereas GAZS queries for Active Zone Set (or Effective Configuration Set).
FOS 8.2.1 enhances the impaired port feature introduced in FOS 8.2.0. With FOS 8.2.0, an impaired port can be used as
a principal link, which is used in fabric-related events, even though there are parallel nonimpaired links available. With
FOS 8.2.1, when a principal link is impaired and parallel links exist, an alternate link will be selected as the principal link.
FOS 8.2.1 adds the new -minDiff to passwdcfg CLI command to enable SAN administrators to configure the
password policy to require the minimum number of characters that must be different between a current password and a
new password.
FOS 8.2.1 displays a logout message for SSH or Telnet session logout, exit, or timeout on the standard session terminal
window.
FOS 8.2.1 adds the -force option to the rootAccess command to bypass the interactive prompt of the command
execution.
FOS 8.2.1 adds support of HTTP server connection KeepAlive. The server KeepAlive is supported only with secure
HTTPS connections to switches. The KeepAlive support may be enabled with the command mgmtapp --enable
keepalive in FOS 8.2.1 or later.
FOS 8.2.1 introduces options to the configure CLI command to allow a number of switch configuration parameters to
be set without going through the interactive CLI menu. For details, refer to the Brocade Fabric OS Command Reference
Manual for FOS 8.2.1.
FOS 8.2.1 supports dynamic LAG (LACP) on Gigabit Ethernet LAN ports on extension platforms.
Broadcom FOS-823e-RN
21
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Some power supply models for Brocade Gen 6 platforms support field upgrade of the firmware image used by their
microcontrollers. FOS 8.2.1 introduces the psutil CLI command for field upgrade of the power supply microcontroller
firmware. The power supply firmware is packaged as part of FOS 8.2.1 and later. Administrators can use the psutil
command to check the power supply firmware version and when necessary upgrade to a later version.
FOS 8.2.1 enhances the chassisName command to increase the name length from 15 characters to 31 characters on
the DCX 8510 and X6 directors.
FOS 8.2.1 supports the portcfgislmode command to configure ISL R_RDY mode on the ISLs in a base switch, that is,
the XISLs. With this enhancement, devices that support R_RDY mode can also be used on XISLs.
FOS 8.2.1 displays the alert thresholds for peer port optics through the sfpShow -link option. The alert thresholds are
displayed for voltage, temperature, Tx Bias, Tx Power, and Rx Power metrics. The alert thresholds are displayed for peer
port optics only.
Broadcom FOS-823e-RN
22
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Broadcom FOS-823e-RN
23
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
NOTE On Brocade G620, G630, Brocade X6-8, and Brocade X6-4 platforms, this license enables the use of
IO Insight capability. The license itself is identified as “Fabric Vision and IO Insight” on these
platforms.
FICON Management Server – Also known as CUP (Control Unit Port), this license enables host control of switches in
mainframe environments.
Integrated Routing – This license allows any Fibre Channel port in a DCX 8510-8, DCX 8510-4, Brocade 6510,
Brocade 6520, Brocade 7840, or Brocade G620 to be configured as an EX_Port supporting Fibre Channel Routing (FCR).
This eliminates the need to add an FR4-18i blade or use a Brocade 7500 for FCR purposes, and it also provides either
quadruple or octuple the bandwidth for each FCR connection (when connected to another 16Gb/s- or 32Gb/s-capable
port).
Integrated Routing Ports on Demand – This license allows any Fibre Channel port in a Brocade 7810, G630, X6-8, or
X6-4 to be configured as an EX_Port supporting Fibre Channel Routing. The maximum number of EX_Ports supported
per platform is provided in the license. This eliminates the need to add an FR4-18i blade or use a Brocade 7500 for FCR
purposes, and it also provides octuple the bandwidth for each FCR connection (when connected to another 32Gb/s-
capable port).
Advanced Extension – This license enables two advanced extension features: FCIP Trunking and Adaptive Rate
Limiting. The FCIP Trunking feature allows multiple IP source and destination address pairs (defined as FCIP circuits) via
multiple 1GbE or 10GbE interfaces to provide a high-bandwidth FCIP tunnel and failover resiliency. In addition, each FCIP
circuit supports four QoS classes (Class-F, High, Medium, and Low Priority), each as a TCP connection. The Adaptive
Rate Limiting feature provides a minimum bandwidth guarantee for each tunnel with full utilization of the available network
bandwidth without impacting throughput performance under a high-traffic load. This license is available on the
DCX 8510-8/DCX 8510-4 for the FX8-24 on an individual slot basis. The upgrade license on Brocade 7810 includes this
license to enable 10GbE ports.
Broadcom FOS-823e-RN
24
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
10GbE FCIP/10G Fibre Channel – This license enables the two 10GbE ports on the FX8-24 and/or the 10G FC
capability on FC16-xx blade ports supported on DCX 8510 platforms except for the FC16-64 blade. On the Brocade 6510
and Brocade 6520, this license enables 10G FC ports. The upgrade license on Brocade 7810 includes this license to
enable six 10GbE ports. This license is not applicable to the Brocade 7840, Brocade G620, or Brocade X6 platforms.
On the FX8-24:
With this license installed and assigned to a slot with an FX8-24 blade, two additional operating modes (in addition to
10x1GbE ports mode) can be selected:
10x1GbE ports and 1x10GbE ports
or
2x10GbE ports
On the FC16-xx:
Enables 10G FC capability on an FC16-xx blade in a slot that has this license.
On the Brocade 6510 and Brocade 6520:
Enables 10G FC capability on Brocade 6510 and Brocade 6520 switches.
This license is available on the DCX 8510-8 and DCX 8510-4 on an individual slot basis.
Advanced FICON Acceleration – This licensed feature uses specialized data management techniques and automated
intelligence to accelerate FICON tape read and write and IBM Global Mirror data replication operations over distance,
while maintaining the integrity of command and acknowledgement sequences. This license is available on the
Brocade 7840 and the Brocade DCX 8510-8 and DCX 8510-4 for the FX8-24 on an individual slot basis.
ICL POD License – This license activates ICL ports on DCX 8510 or X6 platform core blades. An ICL license must be
installed on the director platforms at both ends of the ICL connection.
On the Brocade DCX 8510-8 and X6-8:
The first ICL POD license enables 16 (half of the total) UltraScale ICL QSFP ports on the DCX 8510-8 or X6-8
Directors, enabling 8 ICL ports on each core blade. These are QSFP port numbers 0, 1, 2, 3, 4, 5, 6, and 7 on the
DCX 8510-8; while on the X6-8, the QSFP port numbers are 0, 1, 2, 3, 8, 9, 10, and 11. The second ICL POD license
enables the remaining 16 UltraScale ICL QSFP ports on the directors. These are QSFP port numbers 8, 9, 10, 11, 12,
13, 14, and 15 on each core blade of the DCX 8510-8; while on the X6-8, these are QSFP port numbers 4, 5, 6, 7, 12,
13, 14, and 15 on each core blade.
Note that the trunk boundaries are different between CR32-8 core blades on the X6-8 and CR16-8 core blades on the
DCX 8510-8.
On the Brocade DCX 8510-4 and X6-4:
ICL POD licenses are different between X6-4 and DCX 8510-4 Directors. On the X6-4, the first ICL POD license
enables 8 (half of the total) UltraScale ICL QSFP ports on the director, enabling 4 ICL ports on each core blade, which
are QSFP port numbers 0, 1, 4, and 5. The second ICL POD license on the X6-4 enables the remaining 8 UltraScale
ICL QSFP ports on the director, which are QSFP port numbers 2, 3, 6, and 7 on each core blade. On the
DCX 8510-4, a single ICL POD license enables all 16 UltraScale ICL QSFP ports on the director.
Enterprise ICL (EICL) License – The EICL license is required on a Brocade DCX 8510 chassis when that chassis is
connected to four or more Brocade DCX 8510 chassis via ICLs. This license is not applicable to X6 Directors.
This license requirement does not depend upon the total number of DCX 8510 chassis that exist in a fabric, but only on
the number of other chassis connected to a DCX 8510 via ICLs.
NOTE The EICL license supports a maximum of nine (9) DCX 8510 chassis connected in a full-mesh
topology or up to twelve (12) DCX 8510 chassis connected in a core-edge topology. Refer to the
Brocade SAN Scalability Guidelines document for additional information.
WAN Rate Upgrade 1 License – The WAN Rate Upgrade 1 license provides additional WAN throughput up to 10Gb/s on
a Brocade 7840. The base configuration for a Brocade 7840 without this license provides WAN throughput up to 5Gb/s.
Broadcom FOS-823e-RN
25
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
WAN Rate Upgrade 2 License – The WAN Rate Upgrade 2 license provides unlimited WAN throughput (up to the
hardware limit) on a Brocade 7840. WAN Rate Upgrade 2 licenses also enable the use of two 40GbE ports on a
Brocade 7840. The 40GbE ports cannot be configured without the WAN Rate Upgrade 2 license. A WAN Rate Upgrade 1
license must be installed on a Brocade 7840 before a WAN Rate Upgrade 2 license is installed. A WAN Rate Upgrade 1
license cannot be removed before the WAN Rate Upgrade 2 license has been removed.
NOTE The WAN Rate Upgrade 1 and WAN Rate Upgrade 2 licenses apply only to Brocade 7840 platforms.
They control the aggregate bandwidth for all tunnels on that Brocade 7840. The entire capacity
controlled by the licenses can be assigned to a single tunnel, or a portion of the capacity can be
assigned to multiple tunnels. The total bandwidth aggregated for all tunnels should not exceed the
limits established by the licenses.
Extension Upgrade License – The Extension Upgrade license is available on the Brocade 7810, enabling
additional ports, capacity, and features that provide the following: 12 32Gb/s FC ports, 6 10Gb/s Ethernet ports,
4 tunnels, 6 circuits per tunnel, 2.5Gb/s WAN throughput, Fabric Vision, Extension Trunking, Brocade ISL
Trunking, Integrated Routing Ports on Demand, and Brocade Extended Fabrics. This license is shown as a
combination of existing FOS licenses that enable the above capabilities and features.
NOTE FOS v8.2.2 and later supports 6 10Gb/s Ethernet ports in base configuration of Brocade 7810.
10Gb/s Ethernet ports can be enabled without the Extension Upgrade license.
NOTE Temporary licenses for features available on a per-slot basis enable the feature for any and all slots
in the chassis.
Broadcom FOS-823e-RN
26
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Temporary and universal temporary licenses have durations and expiration dates established in the licenses themselves.
FOS will accept up to two temporary licenses and a single universal license on a unit. Universal temporary license keys
can be installed only once on a particular switch, but they can be applied to as many switches as desired. Temporary use
duration (the length of time for which the feature will be enabled on a switch) is provided with the license key. All universal
temporary license keys have an expiration date after which the license can no longer be installed on any unit.
Temporary and universal temporary licenses for Brocade 7810 Extension Upgrade are supported with FOS v8.2.2 or later.
Temporary and universal temporary license for Extension Upgrade do not enable additional ports on 7810.
Broadcom FOS-823e-RN
27
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Use of this Fabric OS release on a switch that has reached its end of support date will result in restricted use of some
support level functions. Firmwaredownload, SupportSave and other support commands will not be available. All other
basic operational capabilities will be unaffected.
Blades OS Support
FC16-32, FC16-48 16G FC blades Supported.
FC16-64 blade 1, 2
0F 1F Supported.
FC8-64 64-port 8-Gb port blade Not supported.
FC8-32E, FC8-48E Not supported.
FCIP/FC Router blade (FR4-18i) Not supported.
Virtualization/Application blade (FA4-18) Not supported.
Encryption blade (FS8-18) Not supported.
Extension blade (FX8-24) Supported.
Up to a maximum of four blades of this type.
FCoE/L2 CEE blade FCOE10-24 Not supported.
1
8510 core blade QSFPs, part numbers 57-1000267-01 and 57-0000090-01, are not supported in the FC16-64. The QSFPs supported in the FC16-64, part number
57-1000294-02, is supported on 8510 core blades.
2
E_Port connections on the FC16-64 blade have the following restriction: connecting a QSFP port between an FC16-64 blade and an ICL QSFP port on a core blade is not
supported.
Broadcom FOS-823e-RN
28
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
13.3.1.1 Typical Power Supply Requirements for Blades in DCX 8510-8 Backbones
(For a specific calculation of power draw with different blade combinations, see Appendix A, Power Specifications, in the
Brocade DCX 8510-8 Backbone Hardware Reference Manual.)
Configured Blades Type of DCX 8510-8 DCX 8510-8 Comments
Number of Blade @110 VAC @200–240 VAC
Ports (Redundant (Redundant
Configurations) Configurations)
Any FC16-32, Port Blade 4 Power 2 Power Supplies 200–240 VAC: 1+1 Power
combination of FC16-64 Supplies Supplies
8-Gb or 16-Gb 110 VAC: 2+2 3 Power Supplies
2F
ports with
QSFP ICLs
256 16-Gb FC16-32, Port Blade 4 Power 2 Power Supplies 200–240 VAC: 1+1 Power
ports + QSFP FC16-48 Supplies Supplies
ICLs (Maximum of 110 VAC: 2+23 Power Supplies
fully populated
FC16-32 Max 8 FC16-32 port blades
blades),
FC16-64
192 16-Gb FC16-32, Port / 4 Power 2 Power Supplies 200–240 VAC: 1+1 Power
ports & max 2 FC16-48, Intelligent Supplies Supplies
intelligent FC16-64, Blade 110 VAC: 2+23 Power Supplies
blades (FX8-24) FX8-24
with QSFP ICLs Max four FC16-48 port blades
and max 2 intelligent blades
336 16-Gb FC16-48 Port Blade 4 Power 2 Power Supplies 200–240 VAC: 1+1 Power
ports + QSFP (Maximum of Supplies Supplies
ICLs seven FC16-48 110 VAC: 2+23 Power Supplies
blades, with Max 7 FC16-48 port blades
one empty port
blade slot)
384 16-Gb FC16-48 Port Blade Not Supported 4 Power Supplies 200–240 VAC:
ports + QSFP For DCX 8510-8, four (2+2)3
ICLs 220 VAC Power Supplies are
required.
384 16-Gb FC16-64 Port Blade 4 Power 2 Power Supplies 200–240 VAC: 1+1 Power
ports + QSFP Supplies Supplies
ICLs 110 VAC: 2+21 Power Supplies
Any FC16-32, Intelligent Dependent on 2 or 4 Power For DCX 8510-8, four (2+2)3
combination of FC16-48, Blade / the configuration. Supplies, 220 VAC Power Supplies are
8-Gb or 16-Gb FX8-24 Combination Requires a depending on the required when any special-
ports and power configuration purpose blades are installed.
intelligent calculation for
blades with the specific
QSFP ICLs configuration.
3
When a 2+2 power supply combination is used, the users are advised to configure the MAPS setting for switch Marginal State to be one Bad Power Supply.
Broadcom FOS-823e-RN
29
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
13.3.2.1 Typical Power Supply Requirements for Blades in DCX 8510-4 Backbones
(For a specific calculation of power draw with different blade combinations, refer to Appendix A, Power Specifications, in
the Brocade DCX 8510-4 Backbone Hardware Reference Manual.)
Configured Blades Type of Blade DCX 8510-4 DCX 8510-4 Comments
Number of @110 VAC @200–240 VAC
Ports (Redundant (Redundant
Configurations) Configurations)
96 ports max FC16-32 Port Blade 2 Power Supplies 2 Power Supplies 1+1 redundancy
with QSFP ICLs with 110 or
200–240 VAC
power supplies
Any combination FC16-32, Intelligent Blade / Not Supported 2 Power Supplies 200–240 VAC:
of 8-Gb or 16-Gb FC16-48, Combination 1+1 Power
ports and FC16-64, Supplies
intelligent blades FX8-24
with QSFP ICLs
Broadcom FOS-823e-RN
30
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
14.1.2 Loading FOS PSDs via Web Tools or FOS Command Line
Installing a FOS PSD image on a switch is performed in the same manner as using a traditional full FOS image. If a FOS
PSD image is loaded on an incorrect switch model (for example, attempting to load a FOS PSD image for a Gen 6 entry
level switch on a Gen 6 Director), the following error message displays:
Cannot download the requested firmware because the firmware doesn't support this platform. Please enter
another firmware.
Brocade SANnav Management Portal version 2.1.1 or earlier does not support FOS PSD images. However, FOS PSD
images are supported with SANnav 2.1.1.3 and later releases. SANnav 2.1.1.3 and later can both host and install FOS
PSD images onto Brocade switches.
Broadcom FOS-823e-RN
31
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
The image files can be downloaded from https://support.broadcom.com/, with the exception of YANG files which are
available on https://www.broadcom.com.
NOTE Embedded switch platforms (6543, 6547, 6548, 6558) do not require a FOS upgrade certificate, and
FOS upgrade certificates are not available on the Broadcom Support Portal and OEM Assist sites for
these platforms.
Broadcom FOS-823e-RN
32
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Broadcom FOS-823e-RN
33
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
This chapter contains information that you should consider before you use this Fabric OS release.
15.1 Scalability
All scalability limits are subject to change. Limits may be increased once further testing has been completed, even after
the release of this version of the Fabric OS software. For current scalability limits for Fabric OS software, refer to the
Brocade SAN Scalability Guidelines for Brocade Fabric OS 8.X document.
15.2 Compatibility/Interoperability
NOTE Microsoft Edge version 79 and later (Chromium-based) is not supported with Web Tools.
Broadcom FOS-823e-RN
34
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
4
Support merged from embedded FOS releases.
Broadcom FOS-823e-RN
35
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
You can download the MIB files required for this release from the Downloads area of the Broadcom support portal. To
download the Brocade-specific MIBs, you must have a user name and password.
Perform the following steps.
1. Go to https://support.broadcom.com/, click Login, and enter your username and password.
If you do not have an account, click Register to set up your account.
2. Select Hardware > Brocade Storage Networking > My Downloads.
3. Navigate to the FOS PSD image for your Brocade platform.
4. Navigate to the link for the MIBs package and download the file to your drive.
Distribution of standard MIBs has been stopped. Download the required standard MIBs from the http://www.oidview.com/
or http://www.mibdepot.com/ or https://www.simpleweb.org/ietf/mibs/.
YANG is a standard data modelling language that defines the data sent over the FOS REST API. Each FOS REST API
module is defined in a YANG module file with a .yang name extension.
To download the Brocade FOS-specific YANG files go to https://www.broadcom.com/products/fibre-channel-
networking/software/fabric-operating-system.
Alternatively, download the YANG files for a specific FOS version from https://github.com/brocade/yang
Broadcom FOS-823e-RN
36
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
15.3.2 FCoE
The following topologies for FCoE on the FC32-64 are not supported with FOS 8.2.2 or later:
Cisco UCS server directly connected to the FC32-64 without a Fabric Interconnect module.
Cisco UCS server with a Fabric Interconnect module connected to the FC32-64 via a Nexus 5000 series switch in
between. Neither running FCoE NPV mode nor L2 switching mode on the Nexus 5000 is supported.
15.3.3 FC-NVMe
FOS 8.1.0 or later is required to support FC-NVMe devices.
FOS 8.2.1 or earlier does not support FC-NVMe over FCR configurations. FOS 8.2.1b supports FC-NVMe in edge
fabric to edge fabric over FCR configuration.
15.3.5 VM Insight
VM Insight is supported on the Brocade G610, G620, G630, and X6 running FOS 8.1.0 and later. Brocade Gen 5
Fibre Channel platforms support frames with the optional FC Application Header for VM Insight to pass through. The
Brocade 7840 and SX6 running FOS 8.1.0 support the Application Header in the FCP emulating tunnel. The FCP
emulating tunnel or FICON emulation is not supported in other extension platforms or earlier firmware. Nonemulating
tunnel on extension platforms support pass through of the Application Header.
VM Insight is not supported across FCR, but frames with the Application Header may traverse through FCR.
FOS 8.2.0 supports VM performance metrics in flows on the ingress F_Port only. The Brocade G610 and G620
support FC metrics. Brocade X6 Directors support both FC metrics and SCSI IO metrics.
Legacy static flow does not monitor SCSI IOPS statistics for frames with the Application Header.
Duplicate subflow entries are displayed after a switch HA failover or multiple restarts of VM traffic without clearing the
status. The workaround is to use the flow --reset sys_mon_all_vms command.
FOS 8.2.0 does not support VM Insight for FC-NVMe traffic.
Broadcom FOS-823e-RN
37
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
The D_Port long-duration test can be run only on one port at a time.
Long-duration electrical loopback tests are not supported.
D_Port tests on 4x32GFC breakout QSFP optics (P/N 57-1000351-01) and 128GFC non-breakout QSFP optics (P/N
57-1000331-01) have the following restrictions:
D_Port for these modules in X6 ICL ports is supported without electric or optical loopback tests.
D_Port on any user port connected by a QSFP require all four user ports within the same QSFP to be in D_Port
mode.
D_Port tests require all user ports in a QSFP to be in the same logical switch.
Dynamic or On-Demand D_Ports are not supported on the user ports in these modules.
If a D_Port test between a Brocade switch and an Emulex adapter is stopped shortly after the test has started, the
adapter firmware may display “No FC Cables connecting the port to switch.” The workaround is to
restart the D_Port test until completion.
If a D_Port test is run through optical media for long distance but the ports are not configured for long distance,
D_Port test can fail without a pre-check error message about the configuration mismatch.
5
Adapter firmware or driver versions that are later than the ones listed in the table may not work.
Broadcom FOS-823e-RN
38
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
15.3.11 Extension
IP extension (IPEXT) between a Brocade 7840 and an SX6 blade is supported only if the 7840 is running FOS 8.0.2
or later. FCIP extension between a Brocade 7840 with FOS 8.0.2 or later. An SX6 blade with FOS 8.0.2 or later is
supported. Extension between a Brocade 7840 or SX6 and a Brocade 7810 is supported only if the 7840 or SX6 is
running FOS 8.2.0 or later. The following table documents the combinations.
Site1 Switch/Blade Site1 Firmware Site2 Switch/Blade Site2 Firmware Supported
7840 8.0.2 or later 7840 8.0.2 or later Both FCIP and IPEXT traffic
SX6 8.0.2 or later 7840 8.0.2 FCIP traffic but not IPEXT traffic
SX6 8.0.2 or later 7840 8.0.2 or later Both FCIP and IPEXT traffic
SX6 8.0.2 or later SX6 8.0.2 or later Both FCIP and IPEXT traffic
7840 8.2.0 or later 7810 8.2.1 or later Both FCIP and IPEXT traffic
SX6 8.2.0 or later 7810 8.2.1 or later Both FCIP and IPEXT traffic
Do not to configure the HA VE pair (VE16, VE26), (VE17, VE27), (VE18, VE28), and so on, where each VE in the pair
is in a different LS with a different traffic policy (port-based routing and exchange-based routing). The workaround is
to configure different HA VE pairs such as (VE16, VE27), (VE17, VE26), and so on when putting each VE pair in a
different LS with a different traffic policy.
When Non-Terminate TCP (NT-TCP) is enabled on traffic control lists (TCLs) and a firmware downgrade to
FOS 7.4.1d is attempted on the Brocade 7840, the downgrade will be blocked. Users must remove NT-TCP from the
TCLs with NT-TCP enabled in order to downgrade the firmware. After the firmware is downgraded to FOS 7.4.1d,
users can re-enable the NT-TCP flag.
Nondisruptive firmware downloads for IP extension support nondisruption to IP traffic for all terminate TCP
connections. UDP and non-terminate TCP traffic may be disrupted during HCL.
After firmware downgrade completion, the Brocade 7840 needs a switch reboot and the SX6 needs a blade power-
cycle. HCL is not supported on firmware downgrades.
If a Brocade 7840 or Brocade X6 Director with an SX6 blade in a non-VF is assigned a fabric ID other than 128, then
during a heavy traffic load, the back-end ports on the 7840 and SX6 may encounter credit loss, which can result in
traffic disruption over the VE ports. This is tracked as Defect 660208.
Broadcom FOS-823e-RN
39
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
If a 10 Gb/s port is configured on a base configuration of Brocade 7810 with FOS v8.2.2 or later without an Extension
Upgrade license, users may see configuration replay errors after the firmware on the 7810 is downgraded to FOS
v8.2.1x.
Some 10G SFP's can run at the 1G speed even though FOS 8.2.3 does not support the 1G setting. The unsupported
setting causes issues. FOS 8.2.0 added a check to verify that the configured port speed matches the installed SFP;
and displays a RASlog message if it does not match. Starting in FOS 8.2.1, when a 10G SFP is placed in a port
configured to 1G, FOS displays the RASlog message and does not bring the port online.
To resolve this issue, you must replace the SFP with a 1G SFP; or configure the port speed to 10G to match the SFP
speed.
Some 10G SFP's are able to run at 1G speed even though the setting is not supported. This was causing issues. In
FOS 8.2.0 a check was added to verify the configured port speed matches the installed SFP and display a raslog
message if it does not.
Starting from FOS 8.2.1 the messages will be displayed, and the port is not brought online when a 10G SFP is placed
in a port configured to 1G.
The resolution is to replace the SFP with a 1G SFP or configure the port speed to 10G to match the SFP speed.
Broadcom FOS-823e-RN
40
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
vTap and in-flight encryption or compression compatibility are supported only on the following platforms:
Brocade DCX 8510, X6, G620, G630, and 6520. On DCX 8510 and 6520 platforms, the chassis configuration “vTap
and Encryption/Compression Coexistence Mode” must be enabled when ports with vTap enabled and ports with in-
flight encryption or compression enabled belong to the same ASICs. Refer to the hardware installation guides for
these platforms for the port-to-ASIC mapping boundary.
When “vTap and Encryption/Compression Coexistence Mode” is enabled, the total IOPS on the same ASIC is limited
to 250,000. If the IOPS exceeds the limit, the vTap flow will be deactivated.
When “vTap and Encryption/Compression Coexistence Mode” is enabled, the effective default zone access mode
must not be “All Access.”
Running the flow --show sys_analytics_vtap command when vTap and QoS High compatibility mode is
enabled but the vTap flow is not active may display the following message incorrectly: “Enable vTap and QoS
High Priority Zone Compatibility Mode to active vTap flow. Please use the
configurechassis command to enable this compatibility mode.” This is tracked as Defect 604429.
After a configuredownload followed by switchenable or a flow statistics reset on the Brocade X6 and G620,
MAPS may incorrectly report a VTAP IOPS > 250,000 violation.
AMPOS 2.2.0 or later is required to support FC32-64 blades. Users should upgrade the Brocade Analytics Monitoring
Platform to AMPOS 2.2.0 or later before adding FC32-64 blades to the X6.
Broadcom FOS-823e-RN
41
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
15.3.14 FICON
For FICON-qualified releases, refer to the Additional Considerations for FICON Environments section of the Appendix for
details and notes on deployment in FICON environments. (This appendix is included only for releases that have
completed FICON qualification.)
15.3.15 MAPS
MAPS monitoring of UCS server login does not support the FENCE action even though the mapsrule command does
not block the configuration.
15.3.16 Miscellaneous
If the ambient temperature is above the recommended operational limit, the power supply units may shut down, in
particular when the ambient temperature is above 62°C for Brocade X6 Directors. This will result in the switch being
shut down without any warning. Refer to the Brocade G620 Hardware Installation Guide and the Brocade X6-8/X6-4
Hardware Installation Guide for the recommended ambient temperature limits for the switches.
After a power supply unit is removed from a Brocade G620, the historyshow command may miss the entries for
this FRU removal or insertion event. In addition, the RASLog error message EM-1028 may be logged when the power
supply is removed. This condition can be corrected by power-cycling the switch.
After running offline diagnostics mode 1 on QSFP ports, a Brocade G620 must be rebooted before operational use.
All links in an ICL QSFP connection on a Brocade X6 Director must be configured to the same speed using the
portcfgspeed command from one of the following supported speeds: 16Gb/s, 32Gb/s, or ASN. To connect an ICL
from an X6 with a 4x32GFC breakout optic (P/N 57-1000351-01) or a 4x16G FC optic to a 4x16G FC optic in a
DCX 8510, the X6 port’s speed must be set to 16Gb/s.
ASN is not supported with 4x32GFC breakout optics (P/N 57-1000351-01).
When connecting 4x32G FC breakout optics (P/N 57-1000351-01) to 32Gb/s SFP peer ports on Gen 6 platforms,
ports may auto-negotiate to 16Gb/s after switchdisable and switchenable on the Gen 6 platforms with 32Gb/s
SFP ports. To avoid this issue, the Gen 6 platforms with 32Gb/s SFPs should be upgraded to FOS 8.2.0 or later or to
FOS 8.1.2b or later. This issue is more likely to occur when the Gen 6 platforms with 32Gb/s SFPs are X6 directors.
Brocade G630 LEDs illuminate amber and green during power-up.
When launching Web Tools over an HTTPS connection and a security warning message for an untrusted certificate
pops up, the pop-up message should be responded to within 20 seconds.
When replacing a FC32-64 blade with a FC32-48 blade, flexport and FCoE configurations should be removed before
the FC32-64 blade is removed.
In ISL between G720 or FC64-48 (64G optic) to a G630 (32G optic) may take longer to converge at 32G speed. This
is applicable when the G630 switch is running pre-FOS 9.0.1. Upgrade the G630 to FOS 9.0.1 or later when
connecting a G630 to a G720 or FC64-48.
The output of CLI command sfpShow or any other interfaces to retrieve information from 32Gb/s 100m QSFP (Part
Number 57-1000490) does not match the Part Numbers on the media sticker labels. The output shows Part Number
57-1000351 instead. This does not affect operation of the optics.
Broadcom FOS-823e-RN
42
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
In addition to defect fixes, software releases may also contain updates to address Common Vulnerabilities and Exposures
(CVEs). The latest security vulnerability disclosures and descriptions of each CVE can be found by visiting the Brocade
Security Advisories web page:
www.broadcom.com/support/fibre-channel-networking/security-advisories
Specific CVEs addressed within any given software release will be publicly released a short period after the initial posting
of the software. This is done to provide enough time for OEMs to qualify security updates prior to public disclosure.
The exact CVEs addressed within the Fabric OS v8.2.x software releases are provided in the following security
announcement:
https://support.broadcom.com/external/content/SecurityAdvisories/0/24997
Broadcom FOS-823e-RN
43
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: SAN switch monitoring tools via SNMP do not function properly
Condition: Performing an SNMP GET NEXT query with multiple varbinds belonging to the same index on
FOS v8.2.3e release
Symptom: Supportsave, firmware upgrade and other support commands stop working
Condition: Gen 5 Directors or Fixed-port switches may prematurely enter into End Of Support (EOS).
Broadcom FOS-823e-RN
44
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: Fabric OS v8.2.3e contains a regression that can lead to a memory leak when performing
SNMP Bulk Requests with all the objects(OIDs) belonging to ConnUnitPortStatTable and the
same port index.
Workaround: Disable Port Performance Monitoring in any management applications, for example SANnav.
Symptom: High CPU load observed on switch once a connection via WebTools is made with https.
Broadcom FOS-823e-RN
45
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Audit log is not generated when LDAP authorization fails in non-VF mode.
Condition: Observed on switch having virtual fabrics disabled, LDAP configured as authentication service,
and user authorization fails.
Symptom: With a custom UDP port, syslog is not happening with IPv4 FQDN.
Condition: The switch is configured with hostname/fqdn and fos messages are not going out with a
custom UDP port.
Condition: With AUTHPriv and NTP server configured and time drifts into the past.
Broadcom FOS-823e-RN
46
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Frame drops are seen on EX-ports after the edge fabric switch reboot and devices are stuck in
init state without being imported.
Condition: There is a 10 minute polling timer which will get triggered when there is no change in ex-port
condition, followed by some ex-port action. During build fabric if the 10 minute polling timer
got triggered which will result in creating partial device pairs and leads to device import issues.
Condition: When multiple supportShows are executed, MAPS reports CPU High usage alert.
Condition: High CPU usage when fetching connunitportstat quries using SNMP on Gen5 platforms.
Broadcom FOS-823e-RN
47
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: In rare occasions, snmp terminated while processing GET or GETBULK operations with FCIP
related OIDs during firmwareupgrade.
Symptom: ISL ports become disabled when connecting FOS 8.x with FOS 7.x switch after updating
Enhanced Object Zoning. switchshow as below: Index Port Address Media Speed State
Proto ================================================== 28 28 011c00 id
8G No_Sync FC Disabled (ESC Enhanced Zone Object Name Conflict)
Condition: When Enhanced Object Zoning was defined and added in the effective cfg, and there were
also other non-Enhanced Object zoning following it, removing all Enhanced Object zoning
cannot clear the counter to zero. This cause switch wrongly segmented with other switches
that do not support the Enhanced Object Zoning figure, such as FOS 7.x.
Broadcom FOS-823e-RN
48
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Slow Drain devices on E or EX Port connected switches can cause excessive XTUN-1006
FCIP TX Frame Drop RASLOGs per second.
Condition: Slow Drain Device(s) and other traffic running over one highly utilized FCIP VE port can result
in many XTUN-1006 RASLOGs per second
Condition: This problem is encountered in any environment, such as a setup with ESRS
configured, where many CRITICAL alerts (raslogs) are logged, and trigger multiple
supportsaves in a short span of time.
Workaround: Closely monitor switches for Critical alerts and address the conditions causing critical alerts
promptly, or disable ESRS until the critical alerts are being addressed.
Symptom: After a non-critical daemon failed, it did not restart successfully and the switch persistently lost
HA sync.
Condition: It was a double fault scenario: a daemon had to panic first, then the daemon took longer to
change the process state during exit causing FOS to fail to restart the daemon.
Broadcom FOS-823e-RN
49
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Restartable daemons such as mdd, cald, snmp etc terminate and could not be restarted
properly, causing HA out of sync and daemons are left in a defunct state.
Condition: Daemons take too long to dump core files, typically after a failure caused by memory leak or
memory corruption, causing the HA module to timeout.
Symptom: Duplicate member are allowed for a peer zone via GUI and CLI by different alias; However,
switch loss HA sync after user created alias with same member.
Condition: When user create alias peer zone having duplicate members across principal and peer
members.
Recovery: After removing duplicate member entries via aliremove and cfgsave, sync can be attained
through hadisable and haenable or standby CP reboot.
Broadcom FOS-823e-RN
50
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: When an invalid port mappings in the VF configuration file is synced to the standby CP, buffer
overrun would happen and HASM would terminate. This applies to pre-FOS9.0.0 releases.
Condition: The "From address" in the SMTP header was not properly intialized.
Condition: WAN TCP timer reset failure. This could cause a failure and closure of a WAN TCP circuit.
FCIP Tunnel will attempt to close and then re-open a new TCP connection.
Recovery: FCIP Tunnel will auto attempt to close and then re-open a new TCP connection.
Broadcom FOS-823e-RN
51
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Path unstable when zone miss counter increases but the cam entries are verified to be good.
Plogi ack , abts frames are dropped.
Condition: Happens when ports are moved out of a LS and then switchdisable is performed on the LS.
Recovery: Power cycle the slot with the affected ports or cold boot the switch.
Symptom: FCR (iSwitchd) uses RSCNs to convey incremental updates to the device directory. When an
RSCN fails to be delivered, the Name Server database can become out of sync across routed
fabrics.
Condition: When iswitchd does not receive both RSCN ACCept and ACK_1, the re-tries made could go
out-of-order.
Broadcom FOS-823e-RN
52
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: User may see repeated logging of an internal firmwaredownload raslog messages.
Condition: This could happen when there is an unlikely file system error and corrupted
firwaredownloadstatus file size goes to zero. Once the file size is zero, the frequency of the
internal firmwaredownload raslog messages correspond to the frequency of the invocation of
the CLI command firmwaredownloadstatus or the corresponding REST API to fetch
firwaredownloadstatus.
Workaround: To stop the flooding of the interal raslog, user may run the following command via root access.
rm /etc/fabos/upgrade_status1
Symptom: Firmwareupgrade failed due to time-out from a busy standby CP; HA lost sync after cald panic
with a large sized core file and high compact flash usage.
Condition: This is seen on an ESRS configured director, where standby CP continuously attempts to get
the port capabilities causing firmwaredownload to timeout during ESRS report generating.
Following messages repeatedly logged on console or via dmesg: "kernel: : IOCTL not
allowed: cmd 0x2000455e, pid 2277 (cald) -temporary" This left large /var/log/messages*
files on the switch. If there is a hafailover while at this busy standby CP moment, CPs could
lost HA sync and cause cold recovery.
Broadcom FOS-823e-RN
53
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: User can't perform firmware download on the switch from SANNav.
Condition: This issue will be observed when host IP (which is used for performing firmware download) is
already present in the SSH known host with custom port number.
Recovery: Need to manually login into switch using SSH interface and execute following command
"sshutil delknownhost" to delete the host entry from SSH known host.
Condition: When user used incorrect credential to perform mib walk and left a large /tmp/snmpd.log file
filled with" Authentication failed for snmpadmin1"
Workaround: Use correct credential or as root, cat /dev/null > /tmp/snmpd.log before running out of memory.
Symptom: Host reboot will trigger MAPs alert Condition=ALL_SFP(RXP<=200), Current Value:[RXP, 2
uW]
Condition: After a port state change, such as port offline & online, the SFP values are not yet updated
during the polling cycle, but used by MAPs.
Broadcom FOS-823e-RN
54
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: User may encounter VPD write error to WWN card after the system transitions to CF_ACTV
state for WWN Data Protection (WDP).
Condition: This is applicable only for the DCX-8 and 8510-8 platforms.
Symptom: MAPs alarms continue to be generated following the addition of WWN Data Protection license.
Condition: This is applicable to DCX-8 and 8510-8 only, which are the only platforms covered by WWN
Data Protection.
Symptom: UCS Server FLOGI is being held. Issue doesn't disappear when server is shutdown. Server
cannot see paths through this adapter, can only see storage through other Fabric.
Condition: After configuring FCoE enodes, doing hafailover, and then removing all FCoE enodes, NPIV
device logged in and got the 64th or higher AL_PA on one of the first 16 ports on a 48-port
blade on a chassis. Then the same device tried to login again without logging out.
Recovery: Disable and then enable the switch port where the device experienced the issue. This is only a
temporary recovery because NPIV devices that re-login frequently will eventually hit the issue
again.
Broadcom FOS-823e-RN
55
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: When an old certificate expires, MAPS module continues to maintain the old certificate on the
switch and generates a false alarm.
Symptom: MAPS traps being sent to other monitoring software even though maps rule are configured with
action 'NONE'.
Condition: MAPS ignores action set to NONE when BNA or SANnav has registered as a trap receiver.
The setting for BNA/SANnav monitoring ( Second field is 3 in snmp.trapFilter:0,3;) is causing
the NONE for SNMP in the MAPS rule to be overridden, resulting in traps being sent
Condition: During a 3rd party SNMP application sending SNMP GET requests with invalid port index
(swFCPortTable) causing the NULL pointer access.
Broadcom FOS-823e-RN
56
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: It impacts BR6547 embedded switch only when user implements the Factory default on
chassis management module.
Condition: This may occur if the user invokes the CLI command "switchviolation --dump -dcc" or executes
scripts invoking this CLI command.
Broadcom FOS-823e-RN
57
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Workaround: User can enable https keepalive using "mgmtapp --enable keepalive".
Symptom: Switch panic is observed when a timer is added before the same exact timer expires.
Condition: This condition occurred due to a small timing window where an end device went offline at
about the same time as doing a FLOGI or PLOGI.
Symptom: G610 switch state is set to faulty after switchdisable/switchenable. G610:admin> switchshow
... switchType: 170.2 switchState: Faulty switchMode: Native switchRole: Faulty ...
Condition: This impacts G610 switch only and it could happen in a timing race condition when the port is
bounced while traffic is flowing.
Broadcom FOS-823e-RN
58
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Brocade 7840 may encounter kernel panic due to OOM, with tunnels failing to come online
after reboot
Symptom: WWN is missing in the oid when run snmpgetnext command to connUnitPortTable.
Broadcom FOS-823e-RN
59
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: CLI "nsshow -t" does not list the correct device type (host or target); it shows as "Physical
Unknown(initiator/target)".
Condition: Run CLI "nsshow -t" on switch with a specific 3rd party device connected.
Condition: A critical race condition during warm code upgrade, when name server daemon starts sending
rscns to the application like msd, ficud, appsrv etc. and these applications are not yet ready to
receive. It's more likely to happen in a large scale fabric. This only impacts pre-FOS 9.x
releases.
Broadcom FOS-823e-RN
60
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: This issue could be hit on a system configured with multiple Virtual Fabrics when there is an
HTTP/HTTPS based client (SANnav, WebEM, REST clients) monitoring the switch at the same
time as certificates being installed on the switch.
Workaround: Stop the HTTP/HTTPS based clients prior to the installation of the certificates from CLI. Once
the certificates are installed, the HTTP/HTTPS based management applications can start
monitoring the switches.
Symptom: Edge to edge routing was not working and frames dropped on E-port or EX-port.
Condition: A port could be left in a state with a zoning flag set, when it was initially an F-Port and the
device had sent a LOGO or quickly went offline. When such a port is moved to a logical
switch, which does not have any zoning such as a base switch, and the port becomes an
Eport/Export, the issue will be seen on the port.
Workaround: Use 'filterportshow' to confirm zoning is disabled before using a port as an E-Port or EX-port.
Recovery: Move the port to a logical switch with zoning enabled, enable the port, and then move it back to
the original logical switch.
Broadcom FOS-823e-RN
61
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: Observed during a race condition where cald failed to perform internal common transport
registration.
Symptom: Switch panic after cald failed to allocate memory for maintaining keep alive with ESRS.
Symptom: Devices in default allaccess zone cannot communicate to each other across LISLs in FICON
environment on all platform.
Condition: FICON logical switch with allaccess default zone, no configured zones, and FMS mode
enabled. Traffic across LISL is dropped. This should happen only for FOS release 8.2.2c and
later, FOS8.2.3, and FOS9.0.0 and later.
Broadcom FOS-823e-RN
62
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: All hosts on two AGs lost access to storage after ports were added to F-Port trunk group
Condition: This is encountered when N-port trunk master changes to a port in the same trunk that has
never been an individual N-port or an N-port trunk master since the last cold boot or power
cycle.
Recovery: Disable all ports in the N-port trunk at the same time, and then enable them again, all at the
same time.
Condition: Happens when the BI link get a LR and during the online state after LR, the FICON CUP DST
and the BYpass RAM bit is set.
Condition: This issue occurs on platforms where there are no GigE ports or when using a 3rd party
monitor application, which performs snmpget via multiple varbinds to FC ports and interface
tables, this impacts 32G Gen5 platforms only.
Broadcom FOS-823e-RN
63
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: The user will observe that Tunnel with Preshared key will not come up after config download.
Condition: A functioning IKE session between switches running different IKE revisions is restarted at the
higher revision IKE, after the IKE session has been rekeyed at least once.
Symptom: Frames to the remote fabric across the IFL are dropped after D-port testing.
Condition: The port was first configured as D-port and D-port test was interrupted by disabling the port.
The port is then configuraed as IFL port.
Condition: This is encountered when the switchtype CLI command is invoked in Admin account
Broadcom FOS-823e-RN
64
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Periodic XTUN-1997 triggers when running FICON and FCP/SCSI flows over an FCIP Tunnel
Port Based or Device Based Routing configuration. The XTUN-1997 triggers are for Keepalive
timeouts on the medium priority circuits.
Condition: FCIP Tunnel with FCP/SCSI flows in a non-EBR logical switch. There are times when an
FCP/SCSI flow uses the control VTN between the switches. This causes a KATO on the
H/M/L active connections.
Condition: This may be encountered on a very busy Chassis System. MDD watchdog timeout occurs
when it takes too long for mdd to get slot count information on a busy chassis platform.
Symptom: 'snmpconfig' CLI returns error 'Failed to get snmp config info' due to SNMP service not
restarting after getting disrupted.
Condition: When 3rd party application polls FC data using SNMP with negative value as port index.
Broadcom FOS-823e-RN
65
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: FICN_1062 and FICN_1063 RASLOGs every 1.5 seconds on FICON Emulation enabled FCIP
Tunnel
Condition: This is seen on XRC FICON Emulation enabled tunnel. If a device presents Attn status that is
Idle Status Accepted at the remote site, but then fails to be presented at the host side, and the
device re-presents the Attn status when the host side is still in progress, then the second and
subsequent Attn status presentations from the device side are aborted by the device.
Symptom: FOS accepts REST request with an empty audit class list.
Condition: Run a REST request to set an empty audit class which should be handled as an invalid input.
Symptom: TX rules for ISL ports do not get triggered for MAPS custom policy that includes both RX and
TX rules for the ISL ports.
Condition: This is specific to Gen 5 stats collection where the stats on the tx utilization is NOT reported
back to maps correctly.
Broadcom FOS-823e-RN
66
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Boot over SAN device does not work after upgrading firmware on 32G FC switches.
Condition: This is encountered on 32G switches soon after upgrading to FOS8.2.3a, FOS8.2.3b,
FOS9.0.1c, FOS9.1.0 and FOS9.1.0a.
Symptom: Different (lower) tunnel throughput on 7800 or FX8-24 configuration after reboot or slot power
cycle.
Condition: After reboot of a 7800 switch or FX8-24 blade, the FCIP data complex could incorrectly setup
egress internal FC ports leading to reduced FC throughput.
Symptom: SNMP daemon leaks memory and causes switch to hafailover/hareboot/panic when switch
runs out of memory.
Condition: Memory leak is observed from SNMP get request for connUnitSNStable objects.
Broadcom FOS-823e-RN
67
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Stale CAM entries are present on the ports, which were disabled.
Condition: When a port is disabled, the CAM entries on the port are not deleted except for the Loopback
CAM pair.
Recovery: Poweroff the slot or reboot the switch to recover from this incorrect state.
Symptom: Performance stats for VE ports are not present in connunitportstat table
Symptom: During the FOS upgrade process, initiated from SANnav, directors can experience
unexpected reboots during the upgrade process. In each director where this occurred the
FOS upgrade had completed on the Standby CP and then an unexpected reboot
occurred. Both CR blades reset and started POST diagnostics.
Condition: FICON logical switch with 256 ports that are bound to all available areas. ICLs then moved to
the FICON logical switch cause VERIFYs and ASSERT upon hafailover.
Workaround: Remove two ports bound to areas from the logical switch. Then remove all ICLs from the
logical switch and add them back.
Broadcom FOS-823e-RN
68
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: 7840, 7810 or SX6 blade encounters DP Linux out of Memory causing IO disruption
Condition: FCIP Tunnel down events after many (>> 25,000) FTRACE triggers as indicated by XTUN-
1997 RASLOGs.
Condition: Class 2 TX credit is lost when the end device delays PLOGI during Class 2 communication,
and then hafailover happens twice. This is a new device behavior that exposed a day-1 FOS
issue.
Broadcom FOS-823e-RN
69
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Weblinkercfg continues crashing while processing Enumerate telemetry profile query.
Condition: This occurs when telemetry profile index no longer starts with 0. For example when a user
creates 4 profiles and deletes the first 2 profiles, then the remaining two profile index no
longer starts at 0. This condition can be simulated by using two Sannav servers
simultaneously streaming to kafka brokers.
Symptom: Frame drops affecting entire fabric after creating smaller trunks from larger trunks.
Condition: This is encountered when a small trunk is created from ports in a larger trunk followed by
hafailover.
Workaround: Do not use port of larger trunk for the new trunk. Or alternatively, Disable all ports in the
larger trunk first and then create the smaller trunk.
Recovery: hafailover of the switch showing the drops will recover from this condition.
Broadcom FOS-823e-RN
70
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Code upgrade turned into cold recovery when weblinker cannot restart in time, or on an
normal operation switch, user may encounter failures in config change operations (e.g.
portcfg or lscfg)
Condition: This occurs during code upgrade or if config operations are performed within a minute of any
operation (e.g. install / remove / modify HTTP certificates etc) that forces a restart of the
weblinker.
Workaround: Upgrade one CP at a time. Upgrade the standby CP 1st and wait for 2-3 mins after HA sync is
achieved before failing over to the new firmware. Or Do not make any config changes (e.g
portcfg ftrace 3/24 del, etc) for one minute after any operation (e.g. install / remove / modify
HTTP certificates etc) that entails a restart of the weblinker
Condition: The Path loss occurs when F-port trunk slave on AG is disabled and followed by an hareboot.
Symptom: Switch fault after Portledtest causing EM-1334 & BL-1020 raslog.
Broadcom FOS-823e-RN
71
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Brocade 8510-8 director class switches with a single faulty WWN card, running FOS v8.2.3a
or FOS v8.2.3b may encounter a failure reading from the WWN cards.
Condition: Brocade 8510-8 director class switches and Brocade DCX-8 Director class switches upgraded
to be an 8510-8 running FOS releases v8.3.2a and v8.3.2b.
Symptom: LSAN traffic stopped after adding a new switch to the edge fabric.
Condition: When there is no device entry to send to edge fabric, edge fabric switch misinterprets the
small compressed GE_PT response payload.
Symptom: The following similar raslog may be see without actual excessive power consumption: EM-
1230 1 of 3 Excessive Power usage detected PS1(564)+PS2(144)=708W. System will
shutdown on consecutive readings above 700W.
Broadcom FOS-823e-RN
72
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: When application query ConnUnit Port table via snmp interface on Gen5 platforms.
Condition: Switch panic when trace module has memory corruptions, and module ID is out of range .
Symptom: FCPH-1003 reports duplicate port WWN with a port that does not have the same port WWN.
Condition: This is encountered when many devices are requesting 8-bit area and two such devices
come online at the same time vying for the last available 8-bit area.
Broadcom FOS-823e-RN
73
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: MAPS test email using FROM address of "[email protected]" which is not
always supported.
Condition: When sending a test email from switch, enhanced to user "postmaster" instead of "root".
Symptom: Duplicate PWWN detection resulted in disruption to the existing FICON CHPID.
Condition: Device logs into a FICON switch with a PWWN that matches the PWWN of another device
on the same switch.
Recovery: Remove devices with duplicate PWWNs and bounce the original devices.
Symptom: After upgrading to FOS 8.2.3b, the BSL report no longer included GE_PORT type in the
report.
Broadcom FOS-823e-RN
74
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Enable/disable root account using an AD account fails with error message: Cannot manage
the target account due to conflicting AD permissions.
Symptom: One or more XTUN-1008 RASLOGs followed by tunnel down notifications at the peer switch.
Condition: Server with NPIV connections to fabric is frequently logging in and out of the fabric and
issuing PLOGIs/PRLIs to devices over an FCIP Tunnel. Each PLOGI/PRLI has a new SID as
the NPIV PID normally changes at each fabric login.
Symptom: Switch shows the wrong ip addresses in the login information log
Condition: SEC-1203 messages may show wrong IP during user login. No other functional impact.
Broadcom FOS-823e-RN
75
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: After reboot or switchenable, F-ports and N-ports come online at the same time.
Workaround: Manually persistently disable all F-ports before reboot or switchenable. After reboot or
switchenable, manually persistently enable the F-ports.
Symptom: Switch ports experience immediate hard fault or busy buffer stuck status which leads to a hard
fault and possible switch fault of all ports.
Condition: Frequent plug/unplug of user port cables or a bad link causes physical errors that lead to port
hard faults and/or credit loss.
Recovery: Clean the link errors and bounce the ports with portdisable/portenable.
Condition: This is a very rare occurrence that happens during diagnostic image bootup in DP upon
execution of fcippathtest
Broadcom FOS-823e-RN
76
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Port initialization interop issues with a certain SFP, when connecting to a specific 3rd party
device. This may result in some ports showing errors and some going into no_sync, port_Flt
state.
Condition: This is seen with specific vendor SFP and is resolved via a new CLI "sfpprogram"
enhancement to adjust port optics setting for interop. The new CLI syntax is as follows:
sfpprogram port|port-range
Symptom: Firmwaredownload fails with a generic error-message: "Firmwaredownload sanity check failed.
Please contact the service provider if the issue persists."
Condition: This is seen in an environment with DNS configured, after the backend IP address between
CPs on director are removed during firmwaredownload; Consequently ipaddress cannot be
resolved and firmwaredownload timed out.
Broadcom FOS-823e-RN
77
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: This is seen as a result of switch Boot time not being consistent with queries from SNMP on
FOS8.2.3 and later and 9.0.1 and later. Applications relying on boot time to do calculations
may get unexpected results.
Symptom: Various operations fail after the switch runs out of disk space.
Condition: This only happens with extension platforms. Netstat process on DP coredumps, filling DP's
flash and in turn filling CP's flash during SupportSave operation.
Broadcom FOS-823e-RN
78
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: SNMP application may lose connection to switch momentarily during snmp walk of IPV6
address table and occasionally the user may observe CP lost HA SYNC if the walk is
performed during hafailover/hareboot.
Condition: This may occur when 3rd party applications or scripts perform mib walk on IPv6 address
present in IP Address table (ipAddressTable) with partial indexes, such as "snmpwalk -v 3 -u
snmpadmin1 10.155.22.131 1.3.6.1.2.1.4.34.1.2".
Workaround: Perform complete mib walk instead of querying for IPV6 entries alone.
Recovery: Normally snmpd will restart itself and connection will be restored shortly thereafter. However,
if HA goes out of sync due to back to back snmpd restart failure during HA, a reboot is needed
to recover.
Condition: This happens on extension platforms when the LACP/LLDP control frame has a length greater
than 2112.
Broadcom FOS-823e-RN
79
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: After upgrading Brocade 6548 to FOS 8.2.3a, the user is unable to login to the switch, with the
following error: login: admin "Inconsistency de" - Note: traffic may be impacted too.
Symptom: Disruption on Access Gateways when F-port trunks change trunk master.
Condition: With a F-port trunking group and 1 F -port with 127 NPIV logins, the trunking master port was
disabled, and the new master port got disabled too.
Symptom: When doing firmwarecleaninstall, the following will be displayed on the console: INSTALL26:
ERROR - Cannot get file - /sbin/create_release_file. Please verify the network is good and
retry firmwarecleaninstall after reboot INSTALL26: install FAILED with 0 warnings.
FirmwareCleaninstall failed, you can reboot to retry.
Broadcom FOS-823e-RN
80
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: CP may encounter kernel panic when the FCIP DP generates a large number of XTUN-1006
events per second.
Condition: During FC Egress timeouts from the FCIP DP on an FX8-24 blade, the current implementation
does not limit the number of XTUN-1006 RASLOGs that can be generated per second. When a
severe event occurs where many FCIP FC Tx Timeouts occur, the FCIP DP will generate
enough XTUIN-1006 RASLOGs to consume all CP memory resulting in the kernel panic.
Symptom: Encountered Weblinker posting raslog RAS-1004 with VERIFY after run out of file descriptor.
Condition: This issue is seen on switches running FOS8.2.3a with LDAP configuration during LDAP login
from the management application.
Condition: After using a 3rd party monitoring tool which sends a SNMPGET request with an invalid Fibre
Channel port index. This could happen with FOS8.2.3 or later, FOS9.0.1 or later, FOS7.4.2h or
later and the following CCEs: v8.2.2d4, v8.2.2c6, v8.2.1e1, v8.2.1c5.
Broadcom FOS-823e-RN
81
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Extension tunnel on 7810, 7840 or SX6 blade will not come online in a NAT (Network Address
Translation) enabled WAN environment.
Condition: When an extension circuit is configured to run over a WAN path that utilizes NAT, the circuit
will not establish.
Symptom: eHCL sequence encounters fatal FICON FD error, which results in VE recovery.
Condition: This occurs when performing eHCL sequence over FCIP extended paths with FICON
Emulation enabled on the tunnel(s) when there is active FICON IO.
Condition: If LAN connections being extended to site A are in the process of getting established, and at
the same time VE to another site B goes down, there is a small timing window in the LAN
connection state machine handling, which can bring the LAN connection to a stuck state and it
may never get cleaned up. In such cases, deletion of VE (under which LAN connections are
not cleaned up) will not complete until the switch is rebooted.
Broadcom FOS-823e-RN
82
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: firmwareactivation CLI fails with SULB-1043, after staging a firmware upgrade/downgrade:
2021/09/22-14:58:50:367354, [SULB-1043], 76469/35239, SLOT 1 | CHASSIS, INFO, ,
Firmwareactivate command failed.
Condition: This results from a race condition that may be seen intermittently on an X6 chassis, if/when the
firmwareactivate CLI context process falls behind the HA context process. The fix is only
effective for future upgrades. This issue can still occur when upgrading "from FOS versions"
that do not have this fix i.e. regardless of fix being in the target FOS version.
Recovery:
Symptom: SANnav reports error: "Registration for telemetry profile 'xxx' has failed and Switch panic after
cald termination.
Symptom: sfpshow CLI displays all zeros for txp, rxp etc smart data, and it triggers MAPs alert event.
Broadcom FOS-823e-RN
83
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Dual CP reboot following some zoning changes performed via HTTP.
Condition: Panic occurs when multiple switches are performing configuration changes by webtools at the
same time. This issue is more likely when one of the configuration changes fails.
Condition: The outage happened due to DP reset recovery for a Datapath FPGA error. Enhanced code to
gather more debug info for this situation.
Condition: This is caused as a result of stale LISL ports carried over from FOS v7.4.x / FOS v8.0.x
release
Broadcom FOS-823e-RN
84
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: This issue occurs on platform where there are no GigE ports.
Symptom: Individual ports, that were previously disabled, are enabled while running CLI command
switchcfgpersistentenable
Condition: This is seen with CLI command switchcfgpersistentenable when the switch persistent state is
already enabled and the switch is already online.
Condition: FDMI terminated after memory overrun during GPAT processing. This release addressed
some port attribute length issues and also added enhancements to address and fix corruption
around port attribute queries.
Broadcom FOS-823e-RN
85
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: DCX8510 returning chassis S/N when being queried for brocade-chassis info via REST, when
WWN 1 S/N was previously returned and used for entitlement.
Workaround: Use CLI chassisshow to get WWN 1 serial number, or upgrade to a FOS version with fix and
use the new leaf "entitlement-serial-number" via REST.
Symptom: CLI "sfpshow -all" did not display complete output and the polling of smart SFP data stopped. It
reported an very old "Last poll time:"
Condition: When SFP is plugged into a disabled port or SFP on disabled port is reseated.
Broadcom FOS-823e-RN
86
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: During code upgrade from FOS v8.2.1x to FOS v8.2.3x, FOS cannot completely be brought up
due to cald core dumps. User observes the switch hanging.
Condition: Issue is seen when ESRS is configured and call home alert is triggered before cald is initialized
completely.
Workaround: Issue won't be seen without ESRS configuration. Users can bring up the switch by removing
these "cal.esrs" entries from the configuration file.
Symptom: Port initialization interop issues with a certain SFP, when connecting to a specific 3rd device.
This may result in some ports showing errors and some going into no_sync, port_Flt state.
Condition: This is seen with specific vendor SFP and resolved via a new CLI "sfpprogram" enhancement
to adjust port optics setting for interop. The new CLI syntax is as follows:
sfpprogram port|port-range
Broadcom FOS-823e-RN
87
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: SNMP application may lose connection to switch momentarily during snmp walk of IPV6
address table and occasionally the user may observe CP lost HA SYNC if the walk is
performed during hafailover/hareboot.
Condition: This may occur when 3rd party applications or scripts perform mib walk on IPv6 address
present in IP Address table (ipAddressTable) with partial indexes, such as "snmpwalk -v 3 -u
snmpadmin1 10.155.22.131 1.3.6.1.2.1.4.34.1.2".
Workaround: Perform complete mib walk instead of querying for IPV6 entries alone.
Recovery: Normally snmpd will restart itself and connection will be restored shortly thereafter. However, if
HA goes out of sync due to back to back snmpd restart failure during HA, a reboot is needed to
recover.
Symptom: After upgrading Brocade 6548 to FOS 8.2.3a, the user is unable to login to the switch. Error
shown as below: login: admin Inconsistency de Note: traffic may be impacted too.
Recovery: Swap partition from boot-prom to boot up from the alternate partition.
Broadcom FOS-823e-RN
88
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Performing vulnerability scans (Qualys) leads to high CPU load and Multiple config and
secnotify processes are showing as consuming all memory.
Symptom: Switch is flooded with messages, "attempts to send message type (1) to invalid dest
(NSMIPC:0/0)", without any functional impact to switch.
Condition: This benign message occurs when performing SNMP IF.mib walk on platforms without Gige
ports.
Broadcom FOS-823e-RN
89
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: Conditional race condition when user disabled dhcpv6, but dhcpv6 timer is still running which
performs connection clean up by killing existing connection process. A code error passes an
invalid parameter to the kill command when the dhcpv6 connection is non-existent. This
resulting usage message from the failed kill command keeps growing the log file
/etc/ipadmd_log.txt and depletes the available CF space.
Condition: This occurs following a transient Power failure on any platform that has PSU FRUs with
integrated / embedded FAN.
Workaround: Hareboot
Broadcom FOS-823e-RN
90
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Periodic Cx-1014 "frame loss" raslog messages may be logged without any traffic impact.
Additionally, a link reset follows each frame loss detected occurrence.
Condition: This is encountered under following conditions: 1. A long distance Eport is configured with a
distance over 255Km, and remote side is a 32G/64G platform, or 2. An F-port is configured
with 8 or more credits on a 32G/64G platform.
Workaround: Disable the credit recovery or configure the HBA CR (bb_sc_n) number to less than 8.
Symptom: After replacing MAC value in seccryptocfg, SSH and nmap show the value is not actually
applied to one of the two CPs.
Condition: On executing “seccryptocfg --replace -type ssh -mac hmac-sha1 –force” on active CP, SSH
MAC of Standby CP was still running on older SSH CP.
Symptom: On a 32G DWDM port, D_Port diagnostics fails on the spinfab throughput test and DWDM line
flips fail.
Broadcom FOS-823e-RN
91
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: In a timing window during HCL or during error recovery sequences when a tcp connection is
going away on WAN interfaces, invalid pointer could be accessed.
Symptom: Switch paniced on long spinfab run - "ASSERT - Failed expression: VALID_PTR(ctl)"
Symptom: The REST peak counters for fibrechannel-statistics seen as zero despite traffic flowing.
Condition: This is encountered when collecting port stats with REST API.
Broadcom FOS-823e-RN
92
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: On a NPIV port, and attached device is RDP capable, CLI " sfpshow -f link" will not give the
correct peer port gbic information.
Symptom: Add fence and decommission actions for the monitoring items TXP and RXP in the MAPS
policy.
Condition: This enhancement will eliminate link instability due to SFP's Power fluctuations.
Symptom: 'filterportshow' output is incomplete and switch panic with console output, "BUG: Bad page
map in process supportsave".
Condition: Mirror flows are configured, either by the customer or due to AE-port, when there are high
number of cam entries. Then run the 'portfiltershow' command either manually or in
supportsave to trigger this.
Broadcom FOS-823e-RN
93
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: In an EX-port trunk. frames routed to the slave port are dropped, result in LUN discovery
issues after an edge fabric split and merge.
Condition: This is a corner case scenario occurring in specific topologies having trunked EX-Ports when
they are in the process of getting disabled due to Fabric ID (FID) conflict, then re-enabled. If
the EX-port trunk comes back onlline and uses a new PID for the master EX-port, this issue is
observed.
Condition: Trace module memory is somehow corrupted. Any subsequent event that triggers a FFDC or a
Supportsave on switch, could cuase the switch to panic.
Recovery: After panic, the switch is recovered without further action. With FOS8.2.3a, the panic is
avoided and a raslog "KTRC-1006: Trace corruption found." can identify the event has
occured.
Broadcom FOS-823e-RN
94
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: CORE and FFDC are generated on newly active CP after a dual CP firmwaredownload.
Condition: This is a very rare occurrence triggered by a hard to reproduce timing issue hit when EM
module cannot present the blade list information to snmp daemon in a timely manner when the
newly downloaded CP takes over the active CP role.
Condition: When one changes portname of a port after the Fabric Vision license has expired or without a
Fabric Vision license.
Broadcom FOS-823e-RN
95
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: cold boot after ONMd crash when standby becomes active during hafailover process.
Recovery: The switch should be recovered when the switch comes back up after the crash.
Symptom: Webtools hardware view does not show the Led status correctly for ports 50, 51 and 117 .
Condition: This issue is observed only for Brocade G630 switch and isolated to webtools only. CLI
correctly shows the ports as online.
Condition: The issue is seen while generating the certs using secertmngt via pyfos and REST.
Broadcom FOS-823e-RN
96
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Port Optics is not visible in Sannav with error: Fail to get SFP details.
Symptom: DP Panic after upgrading to FOS8.2.1 through FOS8.2.2d while running FICON XRC traffic
over XRC Emulation enabled FCIP Tunnel.
Condition: When processing back to back Command Retry Status and Device End status frames for an
XRC Read Record Set chain.
Condition: When recovering FICON control units over a FICON emulation enabled FCIP Tunnel with
FICON Dynamic Routing (EBR) mode enabled.
Broadcom FOS-823e-RN
97
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: FCIP DP Panic and DP Reset after switching an FCIP Circuit from wantool testing state back
to active FCIP Circuit
Condition: Run WAN Tool and return to normal state over an FCIP Circuit
Symptom: "XTUN-1001 : memory allocation error" messages in RASLOG after using FCIP SLA or
manual WTOOL functionality.
Condition: DRAM2 memory allocation failures after WTOOL DRAM2 leak gets too severe. Leak occurs
during WTOOL session deletion.
Workaround: reboot.
Condition: This is seen when the "runs" parameter of the "systemverification" CLI command is set to
value 4 or higher.
Broadcom FOS-823e-RN
98
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Following an hareboot, the agautomapbalance configuration is not the same as it was before
the hareboot.
Symptom: Issuing a switchenable when the switch is already enabled may lead to I/O disruption
Condition: Running switchenable when the switch is already enabled and has LISLs.
Broadcom FOS-823e-RN
99
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: SANNAV fails to generate switch supportsave on switches running FOS v8.2.2 and above.
Condition: When SANnav initially sets up to use non-default port, and then later changed to use the
default port of 22 for scp/sftp data transfer, switch continues to send data through non-default
port and is rejected by SCP/SFTP server. This defect does not apply to FOS v9.x.
Workaround: Configure SCP/SFTP server to listen on non-default port. Alternatively please gather
supportsave through CLI until FOS code is upgraded.
Symptom: WebTool "Switch Administration -> Extended Fabric" tab are missing slots for X6 directors.
Condition: Using Web Toool, slots 6 and 7 are not displayed for X6-8 and slots 4 and 5 are not disapled
for X6-4 on Extended Fabric tab.
Broadcom FOS-823e-RN
100
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: After running offline diag tests on G630, route may fail and frames may drop.
Workaround: Reboot the G630 after running offline Diag test (Plb, turboramtest , portledtest and
systemverification ).
Condition: This occurs when the user is mapped to a custom role with RBAC class of admin.
Symptom: X6 eth0 reverting to 10/half when ethif is used to configure speed/duplex settings while network
connection is removed
Broadcom FOS-823e-RN
101
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: The failure shows when the auditlog storage file is corrupted for unknown reason.
Symptom: Switches showing MAPS-1003 raslog with 0mvolt messages on port with SFP and QSFP.
Condition: QSFP values(volt,temp,rx power,current) is read as zero in regular polling. Added retry logic to
improve data sampling.
Broadcom FOS-823e-RN
102
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: D-Port test fails with Protocol Error on switch end and on HBA, EL failed despite the overall
result states Passed
Condition: D-port test gets failed/ stuck in IN_PROGRESS when the test is executed on the QSFP links
connected to GEN5 HBAs.
Condition: This occurred because the BE link DID check bit was set in the past, which led to frame drops
over that particular BE link.
Symptom: Following HA failover some ports appear to encounter ASIC stats errors which causes these
ports to be fenced if MAPS is configured to fence ports for ITW.
Condition: The low level asic stats errors are not caused by the HA failover action and appear to be from
a prior uninitialized condition.
Broadcom FOS-823e-RN
103
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: The ISL link between two G620 switches fails to recover and goes to "HRD_FLT" state.
Condition: This is seen after repeated cable-pull tests. Issue is also reported on F-port.
Condition: During hafailover, domain name is not getting synced to sendmail file, email will not be send.
Workaround: Avoid reboot switch if gateway address is set with 12 full digits until upgrade. Or use less than
12 digits gateway address
Recovery: After each reboot customer has to login via serial cable to make any changes.
Broadcom FOS-823e-RN
104
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: CLI "mapsdb --congestion" shows an event but "mapsdb --congestion -freq" doesn't show it.
Condition: MAPS had wrong port stats about congestion details leading to this mapsdb show
congestion/frequency inconsistencies.
Condition: Issue was observed during spinfab or portloopback traffic test. Especially when the test was
abort with a control-c.
Broadcom FOS-823e-RN
105
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: After a host behind an NPIV device reboots, it is unable to discover any of its LSAN devices.
Condition: Under a rare condition, intercommunication packets between name server were dropped.
Symptom: When a longer FQHN is used for TACACS+ authentication, local accounts as well as AAA
accounts are not accessible.
Symptom: Switch panics after various daemons, such as fcpd, tsd, webd, terminate while accessing
configuration database resource key.
Condition: This is mostly observed on Gen6 fix-port switches such as G620; However, it has been seen
on BR7840 and directors. Configuration database resource key is cached in shared memory
and frequently accessed by daemons. Sometimes, the shared memory used by daemon is no
longer valid while configDB on permanent storage is still intact.
Broadcom FOS-823e-RN
106
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: When a crypto template is applied, the existing SSHd session is not terminated.
Condition: After applying a SSH crypto template , the existing session remains open.
Symptom: Missing N-port info from "ag --show" after toggle N-ports.
Condition: After toggling N-ports that connect to Gen5 and Gen4p , "ag --show" has missing N-port info
"Attached_Switch", "Switch F-port" and "IP_Addr". This is seen on FOS v8.x and later.
Condition: This is seen when FDMI attempts to deallocate the buffers used to get the HBA/Port entries.
The dynamically allocated block attempted to be freed appears to be invalid due to memory
overrun.
Broadcom FOS-823e-RN
107
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: When there is a transaction already open in a switch (ex: via CLI) and another transaction
starts (distribution via ACA, SFC, UFC, RCA through WT/BNA) in a remote switch in the same
fabric simultaneously. This is a timing issue when a customer creates multiple zoning
transactions on different switches at the same time.
Workaround: Use the cfgtransshow cli command to confirm there are no outstanding zoning transactions
before initiating a new one.
Symptom: System may encounter emd termination resulting in failover to Standby CP.
Condition: This is seen during supportsave collection, possibly due to memory corruption
Broadcom FOS-823e-RN
108
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: Issue happens when perform HA failover / HA reboot in a logical switch without port index of
zero.
Workaround: Move the device to a logical switch where port index zero exists.
Symptom: Experience silent reboot (reboot reason: reset) or switch hung for hours/days.
Condition: Running script with nested or concatenated CLI commands over 512 bytes long.
Workaround: Allow 5 seconds of idle time between commands and keep CLI length under 512 bytes.
Broadcom FOS-823e-RN
109
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: A good replacement blade cannot come up with raslog [EM-1134] rc=20074: 2019/11/19-
11:28:33:895421, [EM-1134], 95966/2420, SLOT 2 | FFDC | CHASSIS, ERROR,, Slot 7 set to
faulty, rc=20074., OID:0x43700000, em_board_lib.c, line: 1944, comp:emd, ltime:2019/11/19-
11:28:33:895302 And after CLI "slotpoweron" bring back the blade, it is powered off again
after active CP reboot,
Condition: After a blade is intentionally powered down due to PCI errors by FOS, or by a user issued
slotpersistentpoweroff, while running FOS8. 1.0 or earlier, a replacement blade does not
power up, or if it is powered back up via "slotpoweron" CLI command then it is powered down
again after CP reboot.
Recovery: 1. Make sure that a good replacement blade is inserted and power it on via the
"slotpoweron" CLI command. 2. Perform an hafailover; wait for hasync 3. Perform another
hafailover Note: please use slotpoweron to recover, and do not use slotcfgpersistent --
poweron to recover the blade.
Symptom: SNMP community strings with basic SNMPv1 queries are not recognized.
Condition: A timing condition in code caused snmpconfig commands fail leading to database
inconsistency and snmp queries become unresponsive
Broadcom FOS-823e-RN
110
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Weblinker cannot restart with continuous messages; "system is not ready for HA/LS CLI cmd".
Condition: After a configuration change such as "seccertmgmt generate -cert https -years 20 -f".
Symptom: Customer may encounter frame drops corresponding to specific SCSI task management
commands.
Condition: It happens when multiple end devices connected to the same 16G ASIC going through an
error recovery which floods the ASIC with PLOGI/ADISC.
Symptom: SX6 blade is not powered-on, when initializing the blade, after EM_INCONSISTENT failure
during HA failover.
Broadcom FOS-823e-RN
111
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: Trace module is corrupted and subsequently when user daemon or kernel attempts to log a
trace, switch may experience a panic.
Recovery: Switch recovers automatically after the panic without requiring further action.
Condition: Ports with USER-PORT greater than the number of physical ports present were missed during
data collection. ICL ports can have a user port number greater than max_phy_port. However
the total physical ports will never exceed max_phy_port.
Symptom: After running statsclear on default switch, 'Current' counts of '4 History Data' section of maps
database for Base Switch is cleared.
Condition: This issue is reproducible on Port-3/0 only after running statsclear CLI.
Broadcom FOS-823e-RN
112
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: BR7840 or SX6 FCIP Tunnel failure or disable can result in DP reset
Condition: Small timing window when FC frames can be received on the FCIP tunnel after the tunnel is
reported as going offline. It is occasionally observed on FOS v8.x.
Symptom: Fibre channel port statistics counters report large values such as "4294967295" after statistics
reset.
Condition: Counters become invalid after stats reset or during simultaneous data gathering through CLI
and REST API.
Symptom: Lot of I/O failures and traffic disruption reported after performing HA failover on Gen6 chassis-
based switches.
Condition: Port blade and/or core blade slot power cycles, followed by HA failover.
Recovery: Core blade power off and power on recover from this state.
Broadcom FOS-823e-RN
113
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: SNMP OIDs swConnUnitPortTxRate and swConnUnitPortRxRate always show zero values
when querying the switch in AG-mode.
Condition: This is seen with switch running FOS 8.x and above and is in AG mode.
Symptom: The user will observe the Tunnel bounce when changing circuit bandwidth.
Condition: This stems from a bug during PMTU busy retry and applies to FOS 8.x and above.
Condition: This occurs due to a timing issue when CLI debug command is run early in the switch
initialization phase. This applies to FOS 8.x and above.
Broadcom FOS-823e-RN
114
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Customer enabled framelog to capture unroutables , but not all frame drops are captured in the
framelog.
Condition: This issue is seen when the frame drop is less than 20 per second on the port where the drop
is being observed
Condition: This applies to Brocade Switch BR7810 running FOS 8.x and above.
Symptom: Creating a user account with an uppercase password causes the login to fail.
Condition: Create user account in uppercase as below: >userconfig --add USERID -r admin -d "Default
Admin" -p PASSW0RD Account USERID has been successfully added. login: USERID
password: (PASSW0RD) LOGIN INCORRECT
Broadcom FOS-823e-RN
115
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: There is no functional impact to majority of customers with default switch configurations.
Switch may be forced into a reboot loop if *all* the following conditions are met: 1. User should
have automatic ‘firmwarecheck’ enabled on the switch (“firmwarecheck --enable") 2. User
should be using ASC-G 3. User should have imported enterprise CA certificate for the ASC-G.
In the above state, if and when the switch is rebooted it will result in rolling reboot due to
‘firmwarecheck’ failure. This applies to FOS8.2.1c and FOS8.2.2.
Symptom: SX6 AP blade fault occurs after internal backend port is disabled or a disruptive failover even
though it reports HA-IN-SYNC.
Condition: When the active CP has a non-default blade mode (either VE-MODE set to 20VE, or APP-
MODE set to hybrid) and the standby CP is installed for the first time, or netinstalled, the
standby CP will have the wrong app-mode.
Recovery: Reboot the standby CP before any failover attempts are made.
Broadcom FOS-823e-RN
116
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: The BR6520 switch gets into a rolling reboot loop and needs a manual intervention to recover.
Condition: When a bad asic access caused the PCI bus to hang. This impacts BR6520 switches only.
Condition: This is seen when enabling AG mode, the F-port-to-N-port mapping is incorrect.
Recovery: Manually unmap the wrong ports from N-port 0 and then map the correct ports to N-port 0.
Also, the unmapped ports may need to be converted to N-ports (portcfgnport) and have
mapping applied to them.
Condition: Perform CLI firmwaredownload on a switch with ethernet management port configured in half
duplex mode.
Workaround: Configure switch management port to Full Duplex mode. Half duplex mode is not supported.
Broadcom FOS-823e-RN
117
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: After aborts have been processed on a FICON Emulation enabled tunnel when end device re-
used an exchange.
Condition: It happens after a cold boot or core blade slotpoweron. Routes for ICL ports are incorrect such
that fabric services on different switches cannot communicate. This applies to FOS8.2.1 and
above.
Broadcom FOS-823e-RN
118
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Switch panic observed after essd daemon access freed memory.
Condition: During high CPU time (for example, during lots of SNMP requests), inter-switch
communication can be delayed and cause a daemon and/or switch crash.
Symptom: IPEX LAN TCP connection will report failed in the application.
Condition: When there is packet loss on the LAN with Jumbo frames enabled, the LAN TCP receive
process may stall causing LAN TCP connections to timeout.
Symptom: When changing templates and performing HA failover, the SSH config is not replicated to the
standby CP.
Broadcom FOS-823e-RN
119
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: These non-intrusive misbehaviors observed on device enforcement: 1. Move a port from
default switch to logical switch, original PID in the default switch still exists in the new logical
switch. 2. Prohibit two local ports by PDCM in FICON environment. Routing is still allowed
between the two ports.
Symptom: When the Access Gateway switch has a single connection to the director and porttrunkarea is
set on the director, the Access Gateway switch does not report the Switch F-port number.
Condition: Access Gateway is connected to a Gen6 chassis in non-VF mode, N_Port information is
missing the port index of the connecting fabric switch. This impacts FOS8.2.0 and later.
Symptom: Logical group used in rule can be deleted due to inconsistent handling of letter case.
Condition: While deleting the map rule under active policy. Also while deleting rules with -force option.
Workaround: Use proper group names or actual group names that are not differentiated by case alone.
Broadcom FOS-823e-RN
120
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: This is seen when system reset occurs with active IO in FX8-24 FCIP configuration with FICON
Emulation enabled for XRC traffic.
Condition: Devices logging in and out of F-ports multiple times per second.
Condition: When there is trunk and the trunk's Rx is divided by single port's sample time instead of trunk's
average sample time on all FOS8.x.
Broadcom FOS-823e-RN
121
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: It happens when connecting G610 AG switch to a future platform and more than 63 NPIV
devices are connected.
Symptom: When viewing the devicelogin --show output, fabric interconnect devices may not join the
existing members of a group.
Condition: Specific third party devices connected to ports that have area 0xff in the Port Address.
Symptom: SNMP-Get ipNetToMediaIfIndex returns "lo" on X6 and "chassis" on DCX8510. Both should
return "eth0".
Broadcom FOS-823e-RN
122
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: This may occur if the user previously performed a switchdisable / switchenable, in quick
succession without a time gap of minimum 5 minutes, on a base switch running pre-FOS8.1.
This may result in stale LISL left on the switch, which triggers the switch panic during
subsequent upgrade to FOS8.2.x .
Workaround: Provide a minimum time gap of 5 minutes between the last disruptive operations (e.g.
switchdisable/switchenable).
Condition: This occurs due to stale lisl ports in code being carried over from FOS8.0.x FOS8.2.x
Condition: OpenSSH version number of Brocade switches reported during nmap scan of SSH port
Broadcom FOS-823e-RN
123
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Historical graph values for TX/RX are all zeros. Real-time graph values are accurately
displayed.
Condition: When displaying historical graph values for TX/RX, the historical display may show all zeros
when monitoring on GEN 6 platforms. Happens only when monitoring for historical graph
values (Real-Time graph values for TX/RX have no issue). Issue is only seen when
monitoring from BNA 14.4.4 or BNA 14.4.5. Issue is not seen with older versions of BNA or
with SANnav.
Recovery: Rarely seen on GEN 5 platforms, but will recover on next historical display request.
Symptom: FCoE device does not login causes random storage disconnects.
Condition: When connected to a 3rd party switch, all LLDP TLV's are discarded on the port after a cold
reboot of the active CP.
Broadcom FOS-823e-RN
124
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Kernel Panics/cold recoveries encountered in core switches, following exchanges getting timed
out.
Condition: This is seen when an end device sends out-of-order mal-formed multi-frame Common
Transport sequence destined for the Name Server that appears to be a single-frame sequence
and a multi-frame sequence with the same sequence ID. Other Name Server destined frames
are being received at the same time.
Condition: credit zero stats were not correct when port joins or leave trunk.
Symptom: Switch reboot after execute "lscfg --delete" command to deleting a logical switch.
Condition: A race condition happened while executing “lscfg –delete ” and “agshow” command at the
same time. CLI agshow accessed memory just being freed by lscfg.
Workaround: Run “lscfg –delete ” and “agshow” commands two minutes apart.
Broadcom FOS-823e-RN
125
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Need to perform switch excess power check and shutdown for BR7840 and 6520.
Condition: This may happen under rare situation that may cause the board to draw excessive power
Workaround: A CRITICAL raslog EM-1229 is logged when the power consumption is detected to be within
20% limit of reaching the threshold for excessive power. The user is advised to power down
the system and contact Brocade support to have the system replaced
Recovery: In the event of that the EM-1229 raslog goes unheeded and the power drawn crosses the
excessive power threshold then the system will log raslog EM-1230 and power off the
system. The user must contact Brocade support to have the system replaced
Symptom: "Unable to handle kernel paging request" issue seen during SupportSave collection resulting
in HA failover.
Condition: Under heavy cpu load, running supportsave or chipregdmp caused ASIC thread deadlock.
Workaround: Do not run parallel supportsave or asic related dump when system is under heavy load.
Broadcom FOS-823e-RN
126
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: After making zone changes in the fabric, customer may encounter one or both of the
following: 1. Defined zone configuration DB is removed, HA Sync is lost, and verify error seen
on standby CP. 2. Fabric is segmented with reason "segment zone conflict" If the zone data
base is not repaired before issuing a cfgEnable to activate the new zone definition, it can
result in an impact to traffic.
Condition: This issue can be seen with FOS8.2.1, 8.2.1a, 8.2.1b, 8.2.1c,8.2.1d, 8.2.2, 8.2.2a or 8.2.2b.
When a very specific 3rd party device sends the Get Active Zone Size (GAZA) / Get Zone
Size (GZS) CT command with limited response size. Zone database is corrupted after
processing these commands, and an extra semicolon is left at the end of the zone member
list.
Workaround: Avoid using "network fcp zone show" from the device, which triggers GAZS/GZS request to
fabric.
Recovery: If the defined zoning database was truncated but no zone commit operations were performed:
Option 1: From another switch in the fabric that has not been corrupted with the extra
semicolon, create a "dummy zone" and then perform a "cfgEnable" command on the effective
cfgname to restore the zoning database across the fabric. Option 2: On the switch that has
the added semicolon in the zone data base, perform an "haFailover" if it is a director or an
"haReboot" if it is a fixed port switch. Option 3: Coldboot the switch with the bad zonset. If
the bad zonset was pushed to the rest of the fabric, and the zoning data base is erased:
Options presented in order of least disruptive to more disruptive: Option a: On a remote
switch that has an empty defined zoning database issue an "haFailover" if it is a director or an
"haReboot" if it is a fixed port switch. When complete, the zone database should be restored.
Next, create a "dummy zone" and perform a "cfgEnable" command on the effective cfgname
to restore the zoning database across the fabric. Option b: Issue an "haFailover" or
"haReboot" on all directors and switches in the fabric respectively. Option c: Cold boot all
switches in the fabric. Once zoning has been restored to all switches in the fabric, do not
issue the "Network fcp zone show" command from the third party storage array or the failure
could re-appear.
Broadcom FOS-823e-RN
127
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: For the configupload command, the -force flag when saving to the local file system does not
work. It still prompts the user if they would like to overwrite the previous file.
Condition: This is encountered when attempting to force save the config to local file: configupload -force -
local
Symptom: Brocade Gen6 Director switches may encounter one or both of the following occurrences: 1.
Standby CP reports hasmd panic and active CP reports pdmd panic, leading to cold
recovery. 2. User performs lscfg operation, which fails with raslog [PMGR-1006],
204690/5858, SLOT 1 | CHASSIS, WARNING, , Attempt to move port(s) -1 on slot -1 to
switch failed.
Condition: This can only occur on X6-8 or X6-4 director switches after upgrading to FOS 8.2.0 or later
versions from an earlier FOS version, and the PORTMAP table grows in size >= 1024.
PORTMAP table size can only grow beyond 1024 characters after using the "lscfg" CLI
command repeatedly moving ports between logical switches. The current PORTMAP table
size can be calculated by using CLI command "configupload -vf | grep PORTMAP" and count
the characters starting from "F" or "G" to the ending "]" character. X6-8 or X6-4 director
switches shipped from the factory with 8.2.0 or higher FOS already installed will not be
exposed to this issue unless they were downgraded to a version of FOS prior to v8.2.0 and
then upgraded again to FOS 8.2.0 or later.
Workaround: Upgrading from a FOS version prior to v8.2.0 directly to FOS 8.2.2b or later will avoid this
issue. If the switch was previously upgraded to a version of FOS 8.2.0 or higher prior to
upgrading to FOS 8.2.2b, then the upgrade to FOS 8.2.2b alone will not repair the PORTMAP
table.
Recovery: A non-disruptive process to repair the PORTMAP table can be performed, but requires ROOT
access to issue the required command(s) to repair the PORTMAP table. Contact Brocade
Support if the PORTMAP table is growing (GE or FC) near the 1024 size limit. The set of
commands to repair the PORTMAP table can then be provided after examining a
SupportSave from the switch. Alternatively, a disruptive recovery process can be followed
Broadcom FOS-823e-RN
128
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
for users that do not have ROOT access to their switch: 1. configupload -vf 2. Manually
modify the uploaded VF config file to NOT contain ranges that have port numbers between
1800 and 3399 3. configdownload -vf (using the modified VF config file) This will result in an
automatic cold reboot of the switch as the new repaired PORTMAP is installed.
Symptom: Switch panic when doing fosexec --fid all -cmd "portflagsshow" .
Condition: This occurs when running portflagsshow and it encounters an invalid data pointer on a port for
unknown reasons.
Symptom: Windows hosts in cascaded AG G610 do not see the target disk. It can also occur in a non-
cascaded AG environment but less likely
Workaround: Avoid bouncing ports after hareboot. Instead of port bounce, a cold boot can be done, after
which port bounces are fine.
Broadcom FOS-823e-RN
129
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Domain ID, SCC_POLICY, port names, and port configurations reverted to the previous
settings in a different environment.
Symptom: CLI command devicelogin --show output may reflect some ports missing from the CLI output.
Condition: This may occur with UCS device ports with a base address ending in 0x40 0x80 or 0xc0.
Workaround: Change addressing around to have these devices on ports with AL_PA=0x00.
Symptom: Extension device may hit out of memory condition signified by the following RAS Log: XTUN-
1001 FTNL Tunnel 24 Memory allocation failed tracker 3/831. The extension device may also
encounter DP panics once memory is depleted.
Condition: Varying a FICON chipid offline when running through an extension device with FICON
emulation enabled. In addition, this can be hit when hitting exception cases while running an
HCL FOS upgrade. This issue is more pronounced in a large device configuration.
Broadcom FOS-823e-RN
130
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: "Professional Management tool" message reference displayed in Web Tools is no longer
relevant.
Condition: User may see the irrelevant message when using Web Tools,.
Symptom: From BNA, when the user adds the rule HA_SYNC with operator ge, g, for the active policy,
the system does not generate an error as would be expected.
Condition: This occurs with BNA, when adding the rule HA_SYNC with operator ge, g.
Symptom: Encountered FCIP tunnels bounce between two Brocade-7810 switches during HA reboot.
Condition: This may occur during HA reboot in an FCIP or IP Extension routed setup if a burst of unknown
packets comes in on a 7810 GE port.
Broadcom FOS-823e-RN
131
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: The frequency of the stats collection has to be aligned to the wall clock's 00mts or 30mts. For
example: 11:00, 11:30, 12:00, 12:30. But due to an issue in the handling of drift, the Brocade
Support Link (BSL) data collection may not happen in the wall-clock aligned time, but could be
random. This can impact the FA's data co-relation logic.
Condition: This may occur with any Time zone change or system date change (using the "date" CLI
command) or any NTP server sync that results in drift greater than 30 seconds.
Workaround: Any Time zone change or system date change ( using date command ) must be
accompanied by system reboot for the BSL data to be properly interpreted.
Symptom: CLI command portledtest shows the test as passed but LEDs still continue to glow in any
color.
Condition: This is seen when the action parameter is used example: portledtest -action 5, or -uports 3,
or -ports 48
Broadcom FOS-823e-RN
132
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: After a hafailover, blades are no longer seen by CP with following raslog: [PLAT-1000], 5841,
SLOT 6 | FFDC | CHASSIS, CRITICAL, Brocade_DCX, fabPciFindBus No PCI bus found...
Condition: When a FPGA blade status register bit is delayed during a hafailover window, CP interprets
that as a blade is not present. This is a very rare incident.
Symptom: fcprlsprobe timing out on shared area ports with PLOGI being sent to invalid PID ending in FF.
Condition: Running 'fcprlsprobe' on a port with a PID with a base AL_PA of 0x40, 0x80, or 0xc0.
Workaround: Run fcprlsprobe only on ports with a PID with base AL_PA of 0x00.
Condition: Missing information was added to man page, for example dnsconfig.
Broadcom FOS-823e-RN
133
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: When automating commands through plink (part of putty), and the return code coming from the
switch for the command licenseshow is "1". This started appearing in v8.x code. Prior to that,
the return code was "0".
Workaround: Pipe the output to "more" when running the command to get a return code of "0".
Symptom: CLI configdownload failed for filter "snmp" with "Bad Syntax" and "Failed to Parse" errors.
Condition: This is seen when using a configfile uploaded from FOS9.x with keys not supported in FOS8.x
Symptom: After timezone / date change, the next day's Brocade Support Link (BSL) data collected will be
interpreted incorrectly.
Condition: This occurs if the switch is not rebooted following timezone / date change.
Broadcom FOS-823e-RN
134
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Customer may encounter CP resets triggered by SysRq, along with a display of numerous
"Help" messages seen on the console.
Condition: This results from the SysRq request that is triggered from spurious BREAK signal initiated by
the Terminal Server.
Workaround: Run the following command on the Active and/or Standby CP. echo "0" >
/proc/sys/kernel/sysrq To make this persist across CP reboots and hafailover, change the
kernel.sysrq setting in /etc/sysctl.conf file from 1 to 0. (i.e This setting in /etc/sysctl.conf file
should read "kernel.sysrq = 0")
Symptom: CLI "userconfig --change root -e yes" cannot manage the target account with error message of
"conflicting LF permission".
Condition: This occurs when the default account parameters from user defined accounts are changed.
Symptom: On G610, ISL trunk cannot form or port stays in Gport after moving in/out of F-port trunk group.
Condition: When using 4G and 8G links on G610 platform to perform ISL or Fport-trunking.
Broadcom FOS-823e-RN
135
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: This is triggered by the termination of httpd which dumps large number of core files into
memory mapped /tmp directory.
Condition: Special character in the BNA password, such as "$", lead to this supportsave failure.
Symptom: After code upgrade, the user defined flows are not showing up in the admin level command of
"flow --show";
Workaround: Use the "flow --show all" as admin to see the user defined flows.
Broadcom FOS-823e-RN
136
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: Updated man page to removed the statement that says fabric vision license required
Symptom: CLI "flow --show" no longer displays any of the customized flows and flow monitor deactivate
fails with " ioctl failed" message.
Condition: This is encountered when attempting to deactivate a flow with CLI command "flow --deact" or
to display a flow via "flow --show".
Symptom: High CPU load observed on switch once a connection via WebTools is made with https.
Broadcom FOS-823e-RN
137
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: BR6547 is not using credit value configured on peer switch using portcfgeportcredits command
Condition: portcfgeportcredits CLI run on a BR6547 switch may not immediately apply the specified credit.
All other platforms work without issue.
Symptom: Management applications (BNA, SANnav, Webtools) cannot communicate with the switch, and
BSL data collection will also stop. Traffic is not impacted and operation of the switch via CLI is
not impacted.
Condition: Issue will only be observed on GEN 5 platforms that have ASC enabled for BSL. The switch
needs to also be monitored by one or more management applications. A race condition
between the BSL CLI collection, data compression operation and external management
applications can cause the weblinker (http demon) to hang. The likelihood of this occurring will
increase with the amount of data being collected and compressed by BSL.
Recovery: Perform HA Reboot or HA fail-over to restart the weblinker process. This is not disruptive to
traffic.
Broadcom FOS-823e-RN
138
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Customers may encounter: 1. Specific 3rd party devices flood the switch with GID_PN
requests, leading to delays and impact monitoring applications. 2. Maps publishes an alert
about Tx Power of SFP being 0.
Condition: 1. The flood of GID_PN only happens on switches that have QSFP, QSFPP, QSFP28(YTA)
and running FOS v8.1.2g/later, v8.2.1c/later and v8.2.2/later. 2. Maps alert can happen on
any SFP type and any release.
Symptom: Switch experience kernel panic shortly after a SupportSave command is invoked.
Condition: Under an unknown rare condition, the trace module is corrupted triggering switch panic.
Workaround: Avoid using supportsave as monitoring tool, but only use it for debug data gathering.
Broadcom FOS-823e-RN
139
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Observed a weblinker process termination when generating certificate for HTTPS and flash
space eventually went to over 90%.
Condition: This is encountered when Installing/deleting the https certificate while the switch is being
monitored by management applications, such as SANNav.
Workaround: Stop the management applications, such as SANnav, that are monitoring the switch, when
installing and/or deleting https certificates. It's also recommended to allow 3-4 minutes gap
between generation and deletion of HTTPS certificates.
Recovery: Use coreshow to remove all the corefiles. If the issue is seen again, please unmonitor the
switch from management application and use hareboot /hafailover to recovery.
Symptom: Following an abrupt power failure on port blade, the switch continues to access the blade via
PCI causing subsequent system level failures such as CP reboot, hung CP, or a blade that
cannot be powered on.
Condition: Abrupt power loss on port blade is mostly seen with SX6 port blade. Enhancements have
been added in FOS v8.2.2d and later to reduce the timing window for the undesired system
level side effects caused by SX6 sudden power loss.
Broadcom FOS-823e-RN
140
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: This is encountered with an NPIV device on the first 16 ports of a FC32-48 blade in a Brocade
X6-8, and when the devices on that port are bouncing frequently.
Symptom: Message of the day (MOTD) is not displayed upon connecting to the switch.
Condition: On a 32G Logical Switch (not on Default Switch or non-VF Switch), 16G HBA with D-port
feature enabled or 32G HBA with D-port and Fabric Assigned WWN feature enabled together.
Workaround: Disable D-port on the 16G HBA. Disable D-port or Fabric Assigned WWN feature on 32G
HBA.
Broadcom FOS-823e-RN
141
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Edge to edge routing was not working and frames dropped on E-port or EX-port.
Condition: When a port was first a F-Port and device had sent a LOGO or quickly went offline, which
could leave the port in a state with a zoning flag set. When such a port is moved to a logical
switch which does not have any zoning such as base switch, and the port becomes an
Eport/Export, the issue will be seen on the port.
Workaround: Use 'filterportshow' to confirm zoning is disabled before using a port as an E-Port or EX-port.
Recovery: Move the port to a logical switch with zoning enabled, enable the port, and then move it back to
the original logical switch.
Broadcom FOS-823e-RN
142
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: User may experience FCIP ingress traffic stopped due to no more internal credits available and
associated C3-1014/C3-1015 RASLOG messages on a 7840 platform or C4-1014/C4-1015
RASLOG messages in a chassis with an SX6 blade. In the case of an SX6 blade, the blade will
be faulted after multiple attempts to recover internal credits.
Condition: This stems from a timing issue either during tunnel disable or tunnel down processing where
credits from FCIP back to FTNL aren't sent. This leads to permanent flow control on the peer
side and the credit loss on the peer side FC ingress side.
Symptom: Performance degradation for FICON flows after FOS upgrade to FOS 8.2.1 or higher
Condition: When running IO over 7840 or SX6 FCIP Tunnels with FICON emulation enabled.
Symptom: BNA Dashboard and historical performance graphs may report inaccurate data.
Condition: There is a timing window problem in polling between software modules resulting Tx and Rx
rates returned with huge junk values.
Broadcom FOS-823e-RN
143
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Interactive firmwaredownload presents the EULA but does not allow the user to accept it.
Workaround: Reverse the option order via firmwaredownload "-sn", "-sb" etc., such that the "s" option is not
at the end.
Condition: It only happens with a specific 32G LW optics port connected to a specific server.
Broadcom FOS-823e-RN
144
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: After a finite number of outstanding Fastwrite (FW) sequences FW processing will not generate
a transfer ready sequence, but rather will shuttle the exchange (pass through mode). This will
result in the pass through IOs experiencing higher IO response time when compared to the FW
sequences. The FW sequence IOs will complete in as little as 1 Round Trip Time (RTT),
whereas the pass through IOs will complete no sooner than 2 RTTs.
Condition: When using FW extended tunnels at higher IO rates and higher WAN latency, there are IOs
that are passed through instead of having Transfer Ready generated.
Symptom: Spurious power supply voltage and current readings on BR6520 or BR7840 can lead to EM-
1229 and EM-1230 log messages to be displayed and the switch could be shutdown due to
perceived excessive power consumption.
Condition: Spurious power supply readings may happen when the BR6520 or BR7840 is running with a
single power supply and with FOS v8.2.2c.
Workaround: BR6520 or BR7840 running FOS 8.2.2c should always have two power supplies installed and
operating to avoid a potential spurious power reading.
Broadcom FOS-823e-RN
145
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: This occurs with multiple quick succession of Flogi's, as can happen with D-Port and
FAPWWN, if the area binding changes during the Flogi sequence the area route is not updated
correctly.
Recovery: Upgrade to FOS with fix or disable D-Port setting on the HBA, then bounce the port, with the
routing problem, to recover.
Symptom: Traffic disruption encountered when Encryption block errors occur due to errors, such as,
"frames too long".
Condition: This is encountered when Frames of size 2112 or longer traverse ISLs with In-flight
compression enabled. Note: 1. pre-FOS v8.0.2b, upgrades to any release are not impacted.
2. Upgrades, from FOS v8.0.2b through FOS v8.0.2f, to any FOS8.x, with the exception of
FOS 8.1.0GA, exposes this problem. FOS upgrade from impacted version to a desired version
may be done via FOS 8.1.0GA as an interim step to circumvent this issue. 3. Upgrades to
FOS 9.x are not impacted.
Workaround: Downgrade to v8.0.2a or prior and then upgrade to FOS v8.1.0GA before upgrading to a
desired higher FOS release without this fix, or upgrade to a FOS version containing this fix.
Broadcom FOS-823e-RN
146
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: HA state went out of sync after duplicate "zonecreate --peerzone" and cfgadd CLIs.
Condition: If two zone 'add' operations are performed in parallel from two different clients, there is a
chance that duplicate objects can be created in the zone database which causes the Standby
CP to VERIFY during syncing operation and will cause CPs to lose sync.
Condition: This occurs when memory that has just been freed is accessed during agshow CLI. Switch
panic happens when the freed memory is reallocated by others, so it should occur infrequently.
Broadcom FOS-823e-RN
147
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: Enhancement to provide periodic raslog warning messages on the impending End Of Support
(EOS) for the System and remove auditdump from EOS blocked commands.
Condition: Currently the system terminates support functionality at EOS without any warning prior to EOS
Date and auditdump is no longer available after EOS.
Symptom: FCIP Tunnel up, but all I/O stops flowing over the tunnel and application timesout.
Condition: FCIP tunnel in FCR configuration has a slow leak of end to end stream credits. When the
credits are exhausted, all IO is halted.
Symptom: Port is disabled with the reason of: Disabled (LD SFP is not supported in AG mode)
Condition: Issue will be seen when the port with LW SFP is coming online in AG on FOS v8.x and FOS
v9.x
Broadcom FOS-823e-RN
148
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: The chassis management module does not show the FOS version.
Condition: Issue is seen if invalid input is provided for "context" when making SNMPv3 query
Broadcom FOS-823e-RN
149
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: This is seen upon removal and reinsertion of PS with integrated FAN. And also during
disruptive testing of the FAN by deliberate insertion of an object to stop the FAN.
Symptom: If a user role is created with upper and lower case characters, it is stored as lowercase.
However, when specifying a role under the ldapcfg command it does not do the same
conversion.
Condition: When using ldapcfg command, the role is not converted to lowercase when it is used along
with map attributes.
Condition: After FX8-24 blade takes a panic with an invalid memory accessing
Broadcom FOS-823e-RN
150
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Symptom: A port blade failed and went missing from the "slotshow" CLI and switch monitor components
such as maps continue to report system status as healthy.
Condition: This happens when a blade cannot be seen on a PCI bus. Blade should be faulted in this rare
scenario such that switch monitor software can report proper status.
Symptom: Switch needs a manual reboot to recover from HA Out of Sync state after mdd daemon panic.
Observed raslog: [HAM-1013], , SLOT 1 | CHASSIS, CRITICAL, , Can't restart (md
(pid=27115)): System unready or LS trans in progress. Reboot/Failover manually if
necessary.
Condition: During a re-balance action to bring the port group into a balanced state when connecting to a
3rd party switch.
Symptom: Switch reports class-3-discards value as 4294967295 after clearing port stats.
Broadcom FOS-823e-RN
151
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Condition: Misbehaving device sending invalid values for FC header sequence count and sequence ID
leads to out-of-memory condition.
Workaround: Identify the misbehaving devices and remove them from the fabric.
Recovery: The reboot that results from out-of-memory recovers the switch. If high IU memory usage is
detected before the out-of-memory condition, hafailover can temporarily recover and prevent
the unexpected switch reboot.
Symptom: Raslog TS-1001, indicating NTP Query Failed, occurs daily or weekly.
Workaround: Configure switch's clock server to a stable NTP clock server or else confirm the network
stability if issue persists.
Broadcom FOS-823e-RN
152
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Workaround: Enable the EX-port trunk only when the switch is not busy, to avoid this rare timing issue.
Symptom: SNMP incorrectly reporting uptime of 497 days due to deprecation of switchuptime CLI
command.
Condition: This is a rare occurrence and may be encountered while performing a firmwarecleaninstall, or
when a new switch is first installed with FOS8.2.2x.
Broadcom FOS-823e-RN
153
Fabric OS v8.2.3e/v8.2.3e1 Fabric OS v8.2.3e Release Notes
Revision History
Broadcom FOS-823e-RN
154