Veritas NetBackup™

Device Configuration Guide

UNIX, Windows, Linux

Release 8.0 Draft

Veritas NetBackup™ Device Configuration Guide
Legal Notice
Copyright © 2016 Veritas Technologies LLC. All rights reserved.

Veritas, the Veritas Logo, and NetBackup are trademarks or registered trademarks of Veritas
Technologies LLC or its affiliates in the U.S. and other countries. Other names may be
trademarks of their respective owners.

This product may contain third party software for which Veritas is required to provide attribution
to the third party (“Third Party Programs”). Some of the Third Party Programs are available
under open source or free software licenses. The License Agreement accompanying the
Software does not alter any rights or obligations you may have under those open source or
free software licenses. Refer to the third party legal notices document accompanying this
Veritas product or available at:

https://www.veritas.com/about/legal/license-agreements

The product described in this document is distributed under licenses restricting its use, copying,
distribution, and decompilation/reverse engineering. No part of this document may be
reproduced in any form by any means without prior written authorization of Veritas Technologies
LLC and its licensors, if any.

THE DOCUMENTATION IS PROVIDED "AS IS" AND ALL EXPRESS OR IMPLIED

CONDITIONS, REPRESENTATIONS AND WARRANTIES, INCLUDING ANY IMPLIED
WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE OR
NON-INFRINGEMENT, ARE DISCLAIMED, EXCEPT TO THE EXTENT THAT SUCH
DISCLAIMERS ARE HELD TO BE LEGALLY INVALID. VERITAS TECHNOLOGIES LLC
SHALL NOT BE LIABLE FOR INCIDENTAL OR CONSEQUENTIAL DAMAGES IN
CONNECTION WITH THE FURNISHING, PERFORMANCE, OR USE OF THIS
DOCUMENTATION. THE INFORMATION CONTAINED IN THIS DOCUMENTATION IS
SUBJECT TO CHANGE WITHOUT NOTICE.

The Licensed Software and Documentation are deemed to be commercial computer software
as defined in FAR 12.212 and subject to restricted rights as defined in FAR Section 52.227-19
"Commercial Computer Software - Restricted Rights" and DFARS 227.7202, et seq.
"Commercial Computer Software and Commercial Computer Software Documentation," as
applicable, and any successor regulations, whether delivered by Veritas as on premises or
hosted services. Any use, modification, reproduction release, performance, display or disclosure
of the Licensed Software and Documentation by the U.S. Government shall be solely in
accordance with the terms of this Agreement.

Veritas Technologies LLC

500 E Middlefield Road
Mountain View, CA 94043

http://www.veritas.com

.
Technical Support
Technical Support maintains support centers globally. All support services will be delivered
in accordance with your support agreement and the then-current enterprise technical support
policies. For information about our support offerings and how to contact Technical Support,
visit our website:

https://www.veritas.com/support

You can manage your Veritas account information at the following URL:

https://my.veritas.com

If you have questions regarding an existing support agreement, please email the support
agreement administration team for your region as follows:

Worldwide (except Japan) [email protected]

Japan [email protected]

Documentation
The latest documentation is available on the Veritas website:

https://sort.veritas.com/documents

Documentation feedback
Your feedback is important to us. Suggest improvements or report errors or omissions to the
documentation. Include the document title, document version, chapter title, and section title
of the text on which you are reporting. Send feedback to:

[email protected]
You can also see documentation information or ask a question on the Veritas community site:

http://www.veritas.com/community/

Veritas Services and Operations Readiness Tools (SORT)

Veritas Services and Operations Readiness Tools (SORT) is a website that provides information
and tools to automate and simplify certain time-consuming administrative tasks. Depending
on the product, SORT helps you prepare for installations and upgrades, identify risks in your
datacenters, and improve operational efficiency. To see what services and tools SORT provides
for your product, see the data sheet:

https://sort.veritas.com/data/support/SORT_Data_Sheet.pdf
 Contents

Chapter 1 Introducing device configuration ..................................... 8

Using this guide ............................................................................. 8
General device configuration sequence .............................................. 9
Configuration cautions .............................................................. 9
About the NetBackup compatibility lists ............................................. 10

Section 1 Operating systems ................................................... 11

Chapter 2 AIX ......................................................................................... 12

Before you begin configuring NetBackup on AIX ................................. 12

RS/6000 AIX adapter number conventions ........................................ 13
About AIX persistent naming support ................................................ 14
About configuring robotic control device files in AIX ............................. 14
About device files for SAN Clients on AIX .......................................... 14
About configuring tape drive device files in AIX ................................... 15
About choosing a tape driver on AIX ........................................... 15
About non-QIC tape drives on AIX ............................................. 16
About extended-file marks for drives ........................................... 16
About fast-tape positioning (locate-block) on AIX .......................... 17
About no rewind device files on AIX ........................................... 17
Creating AIX no rewind device files for tape drives ........................ 17
Using multiple tape densities on AIX ........................................... 21
About SPC-2 SCSI reserve on AIX ............................................. 21
Disabling SPC-2 SCSI reserve in AIX ......................................... 22
About Sony AIT drives on AIX ......................................................... 22
AIX command summary ................................................................. 23

Chapter 3 HP-UX ................................................................................... 25

Before you begin configuring NetBackup on HP-UX ............................. 25
About robotic control on HP-UX ....................................................... 26
About HP-UX device addressing schemes ......................................... 26
HP-UX tape drive device file requirements for NetBackup ..................... 27
About device drivers and files for HP-UX persistent DSFs ..................... 28
About device drivers for persistent DSFs ..................................... 28
 Contents 5

About persistent DSFs for robotic control ..................................... 29

About persistent DSFs for tape drive access ................................ 29
About persistent DSF pass-through paths .................................... 29
About configuring persistent DSFs ................................................... 30
Creating HP-UX persistent DSFs ............................................... 30
Upgrading NetBackup to use HP-UX persistent DSFs .................... 31
Creating persistent DSF pass-through paths ................................ 32
About HP-UX legacy device drivers and files ...................................... 32
About device drivers for legacy device files .................................. 33
About legacy robotic control device files ...................................... 33
About legacy tape drive device files ............................................ 33
About legacy pass-through paths for tape drives ........................... 34
Creating device files for SAN Clients on HP-UX .................................. 35
About configuring legacy device files ................................................ 35
Creating legacy SCSI and FCP robotic controls on HP-UX .............. 36
About creating legacy tape drive device files ................................ 43
Creating tape drive pass-through device files ............................... 43
About SPC-2 SCSI reserve on HP-UX .............................................. 47
Disabling SPC-2 SCSI reserve in HP-UX ........................................... 47
About disabling the HP-UX EMS Tape Device Monitor for a SAN ............ 47
HP-UX command summary ............................................................ 48

Chapter 4 Linux ...................................................................................... 49

Before you begin on Linux .............................................................. 49

About the required Linux SCSI drivers .............................................. 50
About the st driver debug mode ................................................. 51
Verifying the Linux drivers .............................................................. 51
About configuring robot and drive control for Linux .............................. 51
About the Linux robotic control device files .................................. 52
About the Linux tape drive device files ........................................ 52
Verifying the device configuration on Linux ........................................ 52
About SAN clients on Linux ............................................................ 53
About SCSI persistent bindings for Linux ........................................... 54
About Emulex HBAs ..................................................................... 54
Utilities to test SCSI devices ........................................................... 54
Linux command summary .............................................................. 55

Chapter 5 Solaris ................................................................................... 56

Before you begin on Solaris ............................................................ 56

About the NetBackup sg driver ........................................................ 57
Determining if the NetBackup sg driver is installed .............................. 58
 Contents 6

Special configuration for the Oracle StorEdge Network Foundation

HBA driver ............................................................................ 59
About binding Fibre Channel HBA drivers .......................................... 59
Configuring Solaris 10 x86 for multiple drive paths .............................. 60
Installing/reinstalling the sg and the st drivers ..................................... 60
st.conf file example ................................................................. 63
sg.conf file example ................................................................ 63
sg.links file example ................................................................ 64
Configuring 6 GB and larger SAS HBAs in Solaris ............................... 65
Preventing Solaris driver unloading .................................................. 67
About Solaris robotic controls .......................................................... 68
About SCSI and FCP robotic controls on Solaris ........................... 68
Examples of SCSI and FCP robotic control device files on Solaris
..................................................................................... 69
About Solaris tape drive device files ................................................. 69
About Berkeley-style close ....................................................... 71
About no rewind device files on Solaris ....................................... 71
About fast-tape positioning (locate-block) on Solaris ...................... 71
About SPC-2 SCSI reserve on Solaris ........................................ 72
Disabling SPC-2 SCSI reserve on Solaris .................................... 72
About nonstandard tape drives .................................................. 73
Configuring Solaris SAN clients to recognize FT media servers .............. 73
Adding the FT device entry to the st.conf file ................................ 73
Modifying the st.conf file so that Solaris discovers devices on two
LUNS ............................................................................. 74
Uninstalling the sg driver on Solaris .................................................. 75
Solaris command summary ............................................................ 75

Chapter 6 Windows ............................................................................... 77

Before you begin configuring NetBackup on Windows .......................... 77
About tape device drivers on Windows .............................................. 78
Attaching devices to a Windows system ............................................ 78

Section 2 Robotic storage devices ....................................... 79

Chapter 7 Robot overview .................................................................. 80

NetBackup robot types .................................................................. 80

Robot attributes ........................................................................... 81
ACS robots ........................................................................... 81
TLD robots ............................................................................ 82
Table-driven robotics ..................................................................... 83
 Contents 7

Robotic test utilities ....................................................................... 84

Robotic processes ........................................................................ 84
Processes by robot type .......................................................... 85
Robotic process example ......................................................... 86

Chapter 8 Oracle StorageTek ACSLS robots ................................ 88

About Oracle StorageTek ACSLS robots ........................................... 89
Sample ACSLS configurations ........................................................ 89
Media requests for an ACS robot ..................................................... 93
About configuring ACS drives ......................................................... 93
Configuring shared ACS drives ....................................................... 95
Adding tapes to ACS robots ............................................................ 97
About removing tapes from ACS robots ............................................ 97
Removing tapes using the ACSLS utility ...................................... 98
Removing tapes using NetBackup ............................................. 98
Robot inventory operations on ACS robots ........................................ 98
Configuring a robot inventory filtering on ACS robots .................... 100
NetBackup robotic control, communication, and logging ...................... 101
NetBackup robotic control, communication, and logging for
Windows systems ........................................................... 101
NetBackup robotic control, communication, and logging for UNIX
systems ........................................................................ 102
ACS robotic test utility .................................................................. 106
acstest on Windows systems .................................................. 107
acstest on UNIX systems ....................................................... 107
Changing your ACS robotic configuration ......................................... 107
ACS configurations supported ....................................................... 108
Multiple ACS robots with one ACS library software host ................ 108
Multiple ACS robots and ACS library software hosts .................... 109
Oracle StorageTek ACSLS firewall configuration ............................... 110

Chapter 9 Device configuration examples ................................... 112

An ACS robot on a Windows server example .................................... 112

An ACS robot on a UNIX server example ......................................... 115

Index .................................................................................................................. 119

 Chapter 1
Introducing device
configuration
This chapter includes the following topics:

■ Using this guide

■ General device configuration sequence

■ About the NetBackup compatibility lists

Using this guide

Use this guide to help set up and configure the operating systems of the hosts you
use for NetBackup servers. Also use this guide for help with storage devices. This
guide provides guidance about NetBackup requirements; it does not replace the
vendor documentation.
This guide is organized as follows:
■ Information about operating systems.
■ Information about robotic storage devices.
Read the "Before you start" sections (if applicable) of the chapters in this guide.
These sections provide any important platform-specific instructions or may contain
specific instructions or limitations for server types.
Veritas tested the configuration file options in this guide; other configuration settings
may also work.
To minimize configuration errors, you can copy and paste configuration details from
a text file of the operating system chapters of this configuration guide. The format
of this text file is similar to the printed version of the guide. Be sure to review the
differences as explained at the beginning of the text file.
 Introducing device configuration 9
General device configuration sequence

The NetBackup_DeviceConfig_Guide.txt file is installed with NetBackup server

software in the following paths:
■ /usr/openv/volmgr (UNIX)

■ install_path\Veritas\Volmgr (Windows)

The Hardware Compatibility List contains information about supported devices.

See “About the NetBackup compatibility lists” on page 10.

General device configuration sequence

Use the following general sequence when you configure devices:
■ Physically connect the storage devices to the media server. Perform any
hardware configuration steps that the device vendor or the operating system
vendor specifies.
■ Create any required system device files for the drives and robotic control. Device
files are created automatically on Windows and on some UNIX platforms. Explicit
configuration of device files is required on some UNIX servers to make full use
of NetBackup features.
For SCSI controlled libraries, NetBackup issues SCSI commands to the robotic
devices. SCSI commands allow NetBackup to discover and configure devices
automatically. You may have to configure the server operating system to allow
device discovery.
■ Add the storage devices to NetBackup and configure them.
For instructions, see the NetBackup Administrator’s Guide, Volume I or the
NetBackup Administration Console help.
You can configure devices in NetBackup from the master server or the media
server to which the devices are attached (the device host). For more information,
see "To administer devices on other servers" in the NetBackup Administrator’s
Guide, Volume I or the NetBackup Administration Console help.

Configuration cautions
Observe the following cautions:
■ In multiple-initiator (multiple host bus adapter) environments, NetBackup uses
SCSI reservation to avoid tape drive usage conflicts and possible data loss
problems. SCSI reservation operates at the SCSI target level; the hardware that
bridges Fibre Channel to SCSI must work correctly.
By default, NetBackup uses SPC-2 SCSI reserve and release. Alternatively,
you can use SCSI persistent reserve or disable SCSI reservation entirely.
 Introducing device configuration 10
About the NetBackup compatibility lists

For information about the NetBackup use of SCSI reservation, see the following:
■ "Enable SCSI reserve" in the NetBackup Administrator’s Guide, Volume I.
■ "How NetBackup reserves drives" in the NetBackup Administrator’s Guide,
Volume II.
■ Veritas does not recommend or support the use of single-ended to differential
SCSI converters on NetBackup controlled devices. You may encounter problems
if you use these converters.

About the NetBackup compatibility lists

Veritas provides compatibility lists for the operating systems, peripherals, and
software with which NetBackup works.
See the NetBackup compatibility lists at the following webpage:
http://www.netbackup.com/compatibility
 Section 1
Operating systems

■ Chapter 2. AIX

■ Chapter 3. HP-UX

■ Chapter 4. Linux

■ Chapter 5. Solaris

■ Chapter 6. Windows
 Chapter 2
AIX
This chapter includes the following topics:

■ Before you begin configuring NetBackup on AIX

■ RS/6000 AIX adapter number conventions

■ About AIX persistent naming support

■ About configuring robotic control device files in AIX

■ About device files for SAN Clients on AIX

■ About configuring tape drive device files in AIX

■ About Sony AIT drives on AIX

■ AIX command summary

Before you begin configuring NetBackup on AIX

Observe the following points when you configure the operating system:
■ Verify that NetBackup supports your server platform and devices. Download
the NetBackup hardware and operating system compatibility lists.
See “About the NetBackup compatibility lists” on page 10.
■ Install and configure the IBM AIX Enhanced Device Driver (Atape device driver).
■ Attach all peripherals and reboot the system before you configure the devices
in NetBackup. When the computer is rebooted, AIX creates the device files for
the attached peripherals.
■ For many configuration steps, you can use the smit System Management
Interface Tool. For more information, see the smit(1) man page.
 AIX 13
RS/6000 AIX adapter number conventions

■ To verify that the devices are configured correctly, use smit and
/usr/sbin/lsdev command.
For the tape drives that you want to share among NetBackup hosts, ensure that
the operating system detects the devices on the SAN before you configure the
NetBackup Shared Storage Option.
■ To obtain error and debug information about devices and robotic software
daemons, the syslogd daemon must be active. See the AIX syslogd(1) man
page for more information.
After you configure the hardware, add the robots and the drives to NetBackup.

RS/6000 AIX adapter number conventions

The location code for an adapter consists of two pairs of digits with the format
AA-BB, as follows:
■ AA identifies the location code of the drawer that contains the adapter card, as
follows
■ If AA is 00, the adapter card is located in the CPU drawer or system unit,
depending on the type of system.
■ If AA is not 00, the card is located in an I/O expansion drawer, as follows:
■ The first digit identifies the I/O bus; 0 corresponds to the standard I/O
bus and 1 corresponds to the optional I/O bus.
■ The second digit identifies the slot on the I/O bus.

■ BB identifies the I/O bus and the slot that contains the card, as follows:
■ The first digit of BB identifies the I/O bus that contains the adapter card, as
follows:
■ If the card is in the CPU drawer or system unit, 0 represents the standard
I/O bus and 1 represents the optional I/O bus. If the card is in an I/O
expansion drawer, this digit is 0.
■ The second digit identifies the slot number on the I/O bus (or slot number
in the I/O expansion drawer) that contains the card.

The following are examples of adapter numbers:

■ 00-00 identifies the Standard I/O Planar.
■ 00-05 indicates an adapter card that is in slot 5 of the standard I/O board. The
board is in either the CPU drawer or the system unit, depending on the type of
system.
 AIX 14
About AIX persistent naming support

■ 00-12 indicates an adapter card that is in slot 2 of the optional I/O bus in the
CPU drawer.
■ 18-05 indicates an adapter card that is located in slot 5 of an I/O expansion
drawer. The drawer is connected to the asynchronous expansion adapter that
is located in slot 8 of the optional I/O bus in the CPU drawer.

About AIX persistent naming support

NetBackup requires that you enable persistent naming support for the AIX device
files. Doing so ensures that the device targets and LUNs do not change after a
system restart.
To enable persistent naming support, use the AIX SMIT utility or the chdev command
to change the logical names of the devices. Change the logical names after the
initial device configuration in AIX. For more information, see the IBM documentation.

About configuring robotic control device files in

AIX
For IBM robotic libraries, Veritas supports the IBM AIX Enhanced Device Driver
(Atape device driver) for use with NetBackup. NetBackup discovers the device files
when you configure devices.
For information about the driver and how to configure device files, see the IBM
documentation.
For robotic libraries other than IBM, Veritas recommends that you use an operating
system other than AIX as the robotic control host.

About device files for SAN Clients on AIX

NetBackup SAN clients use tape drivers and SCSI pass-through methods for Fibre
Transport traffic to NetBackup FT media servers. An AIX SAN Client that uses the
standard tape driver can discover Fibre Transport targets on the FT media servers.
The media server FT devices appear as ARCHIVE Python tape devices during SCSI
inquiry from the SAN client. However, they are not tape devices and do not appear
as tape devices in NetBackup device discovery.
During system startup, the AIX cfgmgr command configures all the devices that
are necessary to use the system. If a NetBackup SAN Client cannot discover the
FT devices, you can configure the device files on the client manually. Use the same
procedure that you use for tape devices.
 AIX 15
About configuring tape drive device files in AIX

See “Creating AIX no rewind device files for tape drives” on page 17.

About configuring tape drive device files in AIX

The following topics provide information about configuring tape drive device files
on AIX systems.

Table 2-1 AIX tape drive device file topics

Subject Topic

About choosing a tape driver See “About choosing a tape driver on AIX”
on page 15.

About non-QIC tape drives See “About non-QIC tape drives on AIX”
on page 16.

About extended-file marks for drives See “About extended-file marks for drives”
on page 16.

About fast-tape positioning (locate-block) on See “About fast-tape positioning

AIX (locate-block) on AIX” on page 17.

Creating no rewind device files for tape drives See “About no rewind device files on AIX”
on page 17.

See “Creating AIX no rewind device files for

tape drives” on page 17.

See “Example of creating a no rewind device

file” on page 19.

About multiple tape densities See “Using multiple tape densities on AIX”
on page 21.

About SPC-2 SCSI reserve on AIX See “About SPC-2 SCSI reserve on AIX”
on page 21.

See “Disabling SPC-2 SCSI reserve in AIX”

on page 22.

About choosing a tape driver on AIX

If you use IBM tape drives, Veritas recommends that you install the IBM AIX Atape
driver. For information about the driver, see the IBM documentation.
If you use other tape drives, Veritas recommends that you use the IBM AIX ost
(other SCSI tape) driver. For information about the driver, see the IBM
documentation.
 AIX 16
About configuring tape drive device files in AIX

For information about the drivers and how to configure device files, see the IBM
documentation.

About non-QIC tape drives on AIX

Variable length block and fixed length block refer to how the operating system reads
from and writes to a tape. Variable-mode devices allow more flexibility to read
previously written tapes. Many tape devices can be accessed in either mode.
NetBackup assumes variable length for non-quarter inch cartridge (QIC) drives.
For more information, see the chdev(1) and smit(1) man pages and the system
management guide. The smit application is the most convenient way to change
from fixed to variable-length-block devices manually.

Warning: For NetBackup, you must configure non-QIC tape drives as

variable-length-block devices. Otherwise NetBackup can write data but may not be
able to read it correctly. During a read, you may see a not in tar format error.

When you add a non-QIC tape drive to NetBackup, NetBackup issues the chdev
command to configure the drive as a variable length block device. For reference,
the following is the command that NetBackup uses to configure a drive for variable
mode:

/usr/sbin/chdev -l Dev -a block_size=0

Dev represents the logical identifier for the drive (for example: rmt0 or rmt1).
Therefore, you do not have to configure the drive manually for variable mode.

About extended-file marks for drives

If a tape drive supports extended file marks, you must configure it to use them (for
example, 8-mm drives). Otherwise, NetBackup may not be able to use those drives.
For information, see the AIX chdev(1) and smit(1) man pages.
When you add a tape drive to NetBackup, NetBackup issues the chdev command
to configure the drive to use extended file marks. For reference, the following is the
command that NetBackup uses:

/usr/sbin/chdev -l Dev -a extfm=yes

Replace Dev with the logical identifier for the drive (such as rmt0 or rmt1)
Therefore, you do not have to configure the drive manually for extended file marks.
 AIX 17
About configuring tape drive device files in AIX

About fast-tape positioning (locate-block) on AIX

Applies to AIT, DLT, Exabyte, and half-inch cartridge tape drives.
To position a tape to a specific block, NetBackup supports the SCSI locate-block
command.
NetBackup uses the locate-block command by default.
Veritas recommends that you do not disable locate-block positioning. If you need
to disable it, execute the following command:

touch /usr/openv/volmgr/database/NO_LOCATEBLOCK

With locate-block positioning disabled, NetBackup uses the forward-space-file/record

method.

About no rewind device files on AIX

By default, NetBackup uses no rewind device files. These SCSI device files are in
the /dev directory and have the following format:

/dev/rmtID.1

ID is the logical identifier assigned to the device by the system. The .1 extension
specifies the no rewind, no retension on open device file.
Normally, AIX creates tape drive device files automatically at boot time. Alternatively,
you can run the AIX cfgmgr command, which should create the device files. If they
do not exist, you must create them for the tape drives.
See “Creating AIX no rewind device files for tape drives” on page 17.

Creating AIX no rewind device files for tape drives

NetBackup uses no rewind device files for tape drives and for NetBackup SAN
Clients. During system startup, the AIX cfgmgr command configures all the devices
that are necessary to use the system. If necessary, you can use the following
procedure to check for and create a no rewind device file.
 AIX 18
About configuring tape drive device files in AIX

To check for and create a no rewind device file

1 Display the I/O controllers in the system by using the following command:

/usr/sbin/lsdev -C | grep I/O

The following sample output shows that SCSI controller 1 (00-01) has been
assigned the logical identifier scsi0.

scsi0 Available 00-01 SCSI I/O Controller

2 Display the SCSI and Fibre Channel devices in the system by using the
following command. For SCSI devices, use scsi for the type; for Fibre Channel
Protocol devices, use fcp for the type.

/usr/sbin/lsdev -C -s type

The following example shows two disk drives and a tape drive:

hdisk0 Available 00-01-00-0,0 400 MB SCSI Disk Drive

hdisk1 Available 00-01-00-1,0 400 MB SCSI Disk Drive
rmt0 Available 00-01-00-3,0 Other SCSI Tape Drive

If the device files for the tape drives exist, they appear in the output as rmt0,
rmt1, and so on. The previous example output shows rmt0.
3 If a device file does not exist for the wanted tape drive, create it by using the
following command:

/usr/sbin/mkdev -c tape -s scsi -t ost -p controller -w id,lun

The following are the arguments for the command:

■ controller is the logical identifier of the drive’s SCSI adapter, such as scsi0,
fscsi0, or vscsi1.
■ scsi_id is the SCSI ID of the drive connection.
■ lun is the logical unit number of the drive connection.
For example, the following command creates a device file for a non-IBM 8-mm
drive connected to controller scsi0 at SCSI address 5,0:

mkdev -c tape -s scsi -t ost -p scsi0 -w 5,0

 AIX 19
About configuring tape drive device files in AIX

4 To verify, display the SCSI device files by using the lsdev command, as follows:

/usr/sbin/lsdev -C -s scsi
hdisk0 Available 00-01-00-0,0 400 MB SCSI Disk Drive
hdisk1 Available 00-01-00-1,0 400 MB SCSI Disk Drive
rmt0 Available 00-01-00-3,0 Other SCSI Tape Drive
rmt1 Available 00-01-00-5,0 Other SCSI Tape Drive

The output shows that the rmt1 device file was created.
5 If the device files do not exist on an FCP controller, use the following command
to create them:

/usr/sbin/cfgmgr -l device

device is the controller number from step 1.

6 Ensure that the device is configured for variable-mode and extended file marks.
Use the chdev command, as follows (dev is the logical identifier for the drive
(for example, rmt1)).

/usr/sbin/chdev -l dev -a block_size=0

/usr/sbin/chdev -l dev -a extfm=yes

7 To configure the drive manually in NetBackup, enter the following device file
pathname:

/dev/rmt1.1

Example of creating a no rewind device file

This topic provides an example of how to create a no rewind device file on AIX for
NetBackup. Assume the device files for the wanted SCSI 8-mm tape drive (controller
1, SCSI ID 5) do not exist.
 AIX 20
About configuring tape drive device files in AIX

To create a no rewind device file for SCSI ID 5

1 Determine the logical identifier for the SCSI controller as follows:

/usr/sbin/lsdev -C -c adapter | grep SCSI

The following output shows that scsi0 is the logical name for SCSI controller
1:

scsi0 Available 00-01 SCSI I/O Controller

2 Determine if the device files exist for any device at SCSI ID 5.

/usr/sbin/lsdev -C -s scsi

The following output shows that some device files exist for tape and disk.
However, a device files does not exist for the 8-mm tape drive at controller 1
(scsi0) and SCSI ID 5 (5,0):

hdisk0 Available 00-01-00-0,0 400 MB SCSI Disk Drive

hdisk1 Available 00-01-00-1,0 400 MB SCSI Disk Drive
rmt0 Available 00-01-00-3,0 Other SCSI Tape Drive

3 Create the device file by using the following command:

mkdev -c tape -t ost -s scsi -p scsi0 -w 5,0

4 Display the device files by issuing the following command:

/usr/sbin/lsdev -C -s scsi
hdisk0 Available 00-01-00-0,0 400 MB SCSI Disk Drive
hdisk1 Available 00-01-00-1,0 400 MB SCSI Disk Drive
rmt0 Available 00-01-00-3,0 Other SCSI Tape Drive
rmt1 Available 00-01-00-5,0 Other SCSI Tape Drive

5 To ensure that the tape device is configured for variable-mode and extended
file marks, use the following commands:

chdev -l rmt1 -a block_size=0

chdev -l rmt1 -a extfm=yes

6 To configure the drive manually in NetBackup, enter the following device file
pathname:

/dev/rmt1.1
 AIX 21
About configuring tape drive device files in AIX

Using multiple tape densities on AIX

After you create the device files for your tape drives, you can configure densities
on the drives that support multiple densities. Exabyte 8500C is an example of a
tape drive for which you can use different densities.
AIX supports two configuration settings for density, although not all tape drives
support multiple densities. The default density for both density setting 1 and density
setting 2 is 0, which means maximum density.
The following procedure is an example of using the chdev command to modify the
density settings. Alternatively, you can use the System Manager Interface Tool
(SMIT).
To modify the density settings
◆ The following commands modify both of the tape drive device files:

chdev -l tapedev -a density_set_1=density

chdev -l tapedev -a density_set_2=density

The following describe the arguments to the command options:

■ tapedev is the logical identifier for the drive, such as rmt0 or rmt1.
■ density is a decimal number from 0 to 255 that represents the wanted
density. Zero (0) selects the default density for the tape drive, which is
usually the drive's high density setting. Permitted values and their meanings
vary with different types of tape drives.
To use density setting 1, use the following no rewind on close device file when
you configure the device in NetBackup:

/dev/rmt*.1

To use density setting 2, use the following no rewind on close device file when
you configure the device in NetBackup:

/dev/rmt*.5

About SPC-2 SCSI reserve on AIX

By default, NetBackup uses SPC-2 SCSI reserve and release for tape drive
reservations in shared drive environments. The NetBackup Shared Storage Option
provides shared drive functionality in NetBackup.
Alternatively, you can use SCSI persistent reserve for shared tape drive reservations
in NetBackup, as follows:
 AIX 22
About Sony AIT drives on AIX

■ For the tape drives that support SPC-3 Compatible Reservation Handling (CRH),
you can use SCSI persistent reserve by enabling it in NetBackup. No special
configuration in AIX is required.
■ For the tape drives that do not support CRH, you must disable SPC-2 SCSI
reserve in AIX for those drives. After you disable SPC-2 SCSI reserve, you can
use persistent reserve by enabling it in NetBackup. If the drive does not support
CRH and you do not disable SPC-2 SCSI reserve, access attempts to the drive
fail.
See “Disabling SPC-2 SCSI reserve in AIX” on page 22.

Warning: If the tape driver does not let you disable SPC-2 SCSI reserve, do
not use SCSI persistent reserve with the drives that do not support CRH. The
sctape driver is an example of a tape driver that lets you disable SPC-2 SCSI
reserve.

For more information about NetBackup and SCSI reservations, see the following:
■ The description of the Enable SCSI Reserve Media host property in the
NetBackup Administrator’s Guide, Volume I.
http://www.veritas.com/docs/DOC5332
■ The "How NetBackup reserves drives" topic in the NetBackup Administrator’s
Guide, Volume II.
http://www.veritas.com/docs/DOC5332

Disabling SPC-2 SCSI reserve in AIX

To disable SPC-2 SCSI reserve, use the AIX chdev command to change the reserve
attribute for the device file for the tape drive.
For information about the chdev command, see the AIX chdev man page.
To disable SPC-2 SCSI reserve in AIX
◆ Invoked the following command:

chdev -l name -a res_support=no

Replace name with the name of the device file, such as rmt0.

About Sony AIT drives on AIX

Sony S-AIT drives have DIP switches (SWA and SWB) located on the bottom of
the drive.
 AIX 23
AIX command summary

For the NetBackup Shared Storage Option, you must set the SWA-1 (No Reply
Unit Attention After Process Login) switch correctly. Its factory setting depends
on the serial number of the drive.
If the drive has an early serial number, switch SWA-1 may be OFF. If so, change
the switch to ON.
Drives with newer serial numbers have SWA-1 set to ON as the default.
These newer serial numbers are as follows:
■ SDZ-130 :01442007 and later
■ SDZ-130/L :01200696 and later
Also, for the drives that have a date of May 17, 2004 and later, the DIP switch is
set to ON.
The following table shows the dip switch settings for the newer serial number drives.

Table 2-2 DIP switch positions for AIT drives

Switch Setting (1 = ON and 0 = OFF)

SWA-1 1

SWA-2 0

SWA-3 0

SWA-4 0

SWA-5 0

SWA-6 0

SWA-7 1

SWA-8 0

AIX command summary

The following is a summary of commands that may be useful when you configure
devices. For examples of their usage, see the procedures in this chapter.
■ /usr/sbin/lsdev -C | grep I/O
Displays the adapters that are physically available on your server.
■ /usr/sbin/lsdev -C -s filetype
Displays the device files that have been created, where filetype defines the type
of file displayed: scsi displays SCSI files and fcp displays Fibre Channel files.
 AIX 24
AIX command summary

■ mkdev -c tape -s scsi -t ost -p controller -w id,lun

Creates the device files for tapes.
controller is the logical identifier of the drive SCSI adapter (such as scsi0 or
scsi1). id is the SCSI ID of the robotic connection. lun is the logical unit number
of the robotic connection.
■ /usr/sbin/chdev -l dev -a block_size=0
Configures the drive with logical identifier specified by dev (for example: rmt0)
to variable mode.
■ /usr/sbin/chdev -l dev -a extfm=yes
Configures the drive with logical identifier specified by dev (for example: rmt0)
for extended file marks.
■ /etc/lsattr -l dev -E -H
Displays the device information, where dev is the name of the device (for
example, rmt1).
■ /usr/sbin/cfgmgr -l device
Creates the device files on a Fibre Channel Protocol controller, where device
is the controller number (for example, fscsi0).
■ /usr/bin/odmget -q "name=rmtX" CuAt
Displays the device attributes for the device (rmtX). This command can be used
to determine SCSI target and LUN pairs when you configure Fibre Channel
devices.
Where rmtX is the name of the tape device (for example: rmt0 or rmt1).
 Chapter 3
HP-UX
This chapter includes the following topics:

■ Before you begin configuring NetBackup on HP-UX

■ About robotic control on HP-UX

■ About HP-UX device addressing schemes

■ HP-UX tape drive device file requirements for NetBackup

■ About device drivers and files for HP-UX persistent DSFs

■ About configuring persistent DSFs

■ About HP-UX legacy device drivers and files

■ Creating device files for SAN Clients on HP-UX

■ About configuring legacy device files

■ About SPC-2 SCSI reserve on HP-UX

■ Disabling SPC-2 SCSI reserve in HP-UX

■ About disabling the HP-UX EMS Tape Device Monitor for a SAN

■ HP-UX command summary

Before you begin configuring NetBackup on

HP-UX
Observe the following points when you configure the operating system:
■ Verify that NetBackup supports your server platform and devices. Download
the NetBackup hardware and operating system compatibility lists.
 HP-UX 26
About robotic control on HP-UX

See “About the NetBackup compatibility lists” on page 10.

■ For SCSI controlled libraries, NetBackup issues SCSI commands to the robotic
devices. For NetBackup to function correctly, the properly named device files
must exist.
■ To verify that the devices are configured correctly, use the HP-UX sam utility
and the ioscan -f command.
For the tape drives that you want to share among NetBackup hosts, ensure that
the operating system detects the devices on the SAN before you configure the
NetBackup Shared Storage Option.
■ Some HP SCSI adapters do not support SCSI pass through; therefore, devices
on such adapters are not discovered automatically.
After you configure the hardware, add the robots and the drives to NetBackup.

About robotic control on HP-UX

Several options exist for robotic control, as follows:
■ SCSI, Serial-attached SCSI (SAS), or Fibre Channel Protocol control.
SCSI control includes Fibre Channel Protocol (FCP), which is SCSI over Fibre
Channel. A robotic device in a library moves the media between storage slots
and the drives in the library.
See “About HP-UX device addressing schemes” on page 26.
■ API control over a LAN.
See the "ADIC Automated Media Library (AML)" topic.
See the "IBM Automated Tape Library (ATL)" topic.
See the "Oracle Sun StorageTek ACSLS robots" topic.
If you use API control, you still must configure tape drive device file access in
HP-UX.

About HP-UX device addressing schemes

NetBackup supports the following two HP-UX device addressing schemes for mass
storage devices:
■ Agile addressing, which was introduced in HP-UX 11i v3. Agile addressing uses
persistent device special files (DSFs) for devices.
NetBackup device discovery finds persistent DSFs only. Therefore, Veritas
recommends that you use persistent DSFs.
See “About device drivers and files for HP-UX persistent DSFs” on page 28.
See “About configuring persistent DSFs” on page 30.
 HP-UX 27
HP-UX tape drive device file requirements for NetBackup

■ The legacy naming model.

Legacy device files are supported in HP-UX 11i v3 and earlier.
If you use the legacy model, you must configure devices manually in NetBackup.
You cannot use NetBackup device discovery and automatic configuration.
See “About HP-UX legacy device drivers and files” on page 32.
See “About configuring legacy device files” on page 35.
When HP-UX 11i v3 is installed, both legacy and persistent DSFs are created on
the system. Both types of DSFs can coexist and may be used simultaneously to
access mass storage devices.
NetBackup requires specific device file capabilities for tape drives.
See “HP-UX tape drive device file requirements for NetBackup” on page 27.
You may be required to perform other HP-UX configuration.
See “About SPC-2 SCSI reserve on HP-UX” on page 47.
See “Creating device files for SAN Clients on HP-UX” on page 35.
See “About disabling the HP-UX EMS Tape Device Monitor for a SAN” on page 47.
See “Disabling SPC-2 SCSI reserve in HP-UX” on page 47.

HP-UX tape drive device file requirements for

NetBackup
The following table describes the requirements for tape drive device files.

Table 3-1 Tape drive device file requirements

Requirement Description

Berkeley-style NetBackup requires Berkeley-style close for tape drive device files. The
close letter b in the file name indicates Berkeley-style close device files.

In Berkeley-style close, the tape position remains unchanged by a

device close operation. (Conversely, in AT&T-style close, the drive
advances the tape to immediately after the next end-of-file (EOF)
marker.) To establish the correct position for the next tape operation,
applications must assume the tape's position after a close. NetBackup
assumes Berkeley-style close on HP-UX systems.
 HP-UX 28
About device drivers and files for HP-UX persistent DSFs

Table 3-1 Tape drive device file requirements (continued)

Requirement Description

Fast-tape HP-UX supports locate-block for most drive types for Fast Wide GSC
positioning SCSI adapters. For a list of drive types that NetBackup supports, see
(locate-block) the Veritas Hardware Compatibility List.

NetBackup uses the locate-block command by default if a pass-through

path is configured.

To disable locate-block positioning, execute the following command:

touch /usr/openv/volmgr/database/NO_LOCATEBLOCK

When locate-block positioning is disabled, NetBackup uses the

forward-space-file/record method.

No rewind on close NetBackup requires no rewind on close for tape devices. The letter n
in the file name indicates no rewind device files.

About device drivers and files for HP-UX

persistent DSFs
NetBackup supports persistent device special files (DSFs) for robotic tape libraries
and tape drives. Persistent DSFs are a component of a new storage stack
addressing model HP calls agile addressing. The device World Wide Identifier
(WWID) identifies a device; a device path does not identify a device. The device
file name of the logical unit is independent of the paths that lead to the LUN.
When HP-UX 11i v3 is installed, both persistent DSFs and legacy device files are
created on the system.
You can use persistent DSFs on some HP-UX servers and legacy device files on
others. However, if you use legacy device files, you must configure devices manually
in NetBackup.

Note: NetBackup device discovery finds persistent DSFs only. Therefore, Veritas
recommends that you use persistent DSFs.

See “About legacy pass-through paths for tape drives” on page 34.
See “About legacy tape drive device files” on page 33.

About device drivers for persistent DSFs

The following are the device drivers required to use persistent DSFs:
 HP-UX 29
About device drivers and files for HP-UX persistent DSFs

■ The eschgr driver for robotic control.

■ The estape driver for tape drives.
■ For IBM tape drives, the atdd driver. NetBackup requires a minimum atdd driver
level. See the Hardware Compatibility List for IBM atdd driver levels that are
supported and atdd configuration information.
http://www.veritas.com/docs/TECH76495
Also see the Operating System Compatibility List for the minimum OS patch
level that is required to run on HP-UX:
http://www.veritas.com/docs/TECH76648

About persistent DSFs for robotic control

The following is the persistent DSF name format for robotic control:
/dev/rchgr/autoch#

The # represents the instance number. For example, if HP-UX discovers two robotic
devices (auto-changers) and assigns them instance numbers 0 and 1 respectively,
HP-UX automatically creates the following device files:

/dev/rchgr/autoch0
/dev/rchgr/autoch1

About persistent DSFs for tape drive access

The following is the persistent DSF name format for tape drive read and write
access:
/dev/rtape/tape#_BESTnb

The following describe the DSF name format:

■ The # represents the instance number.
■ BEST indicates the highest density.
■ n indicates no rewind on close.
■ b indicates Berkeley-style close.

About persistent DSF pass-through paths

Although NetBackup requires the /dev/rtape DSFs to configure tape drives,
NetBackup uses pass-through device files for drive access.
 HP-UX 30
About configuring persistent DSFs

NetBackup creates pass-through paths for all valid /dev/rtape paths. NetBackup
creates the paths during device discovery or when you run the
/usr/openv/volmgr/bin/scan command. The following is the file name format:

/dev/pt/pt_tape#

The # represents the instance number that matches the number in the
/dev/rtape/tape#_BESTnb device file or that is retrieved from the ioscan output.

Although NetBackup uses the pass-through device files during tape drive operations,
you specify the /dev/rtape device files if you configure the drives manually in
NetBackup. NetBackup then uses the appropriate pass-through device files.
See “Creating persistent DSF pass-through paths” on page 32.

About configuring persistent DSFs

Configure persistent DSFs for the following:
■ Robotic control.
See “Creating HP-UX persistent DSFs” on page 30.
See “Upgrading NetBackup to use HP-UX persistent DSFs” on page 31.
■ Tape drive read and write accesss.
See “Creating HP-UX persistent DSFs” on page 30.
See “Creating persistent DSF pass-through paths” on page 32.

Creating HP-UX persistent DSFs

By default, new installations of HP-UX 11i v3 and later create both persistent DSFs
and legacy device files for the eschgr and estape drivers. However, you can reinstall
or create the persistent DSFs.
First connect the devices to the system. Then, create persistent DSFs on every
server on which you want to use the new persistent DSFs.
 HP-UX 31
About configuring persistent DSFs

To create the persistent DSFs automatically

◆ Depending on the driver, enter the following command as root:
For the eschgr autochanger driver, enter insf –d eschgr.
For the estape tape driver, enter insf –d estape.
For the IBM atdd tape driver, enter insf –d atdd
To refresh the device path for all devices that use a driver, adding the –e option
to the command line.
For more information about using the HP-UX insf command, see the man
page.

Upgrading NetBackup to use HP-UX persistent DSFs

Use the following procedure to configure an existing NetBackup environment to
use persistent DSFs after you upgrade a media server to HP-UX 11i v3.
The following procedure also removes the legacy paths from the NetBackup device
configuration. To save the legacy paths, you can use the NetBackup tpconfig
utility to disable the old paths but retain them in the device configuration.
If you run the NetBackup Device Configuration Wizard before you perform this
change, NetBackup adds the new DSF paths to the device configuration. However,
the legacy paths remain configured as alternate paths.
To change a NetBackup media server to use the new DSFs
1 Add the following entry to the /usr/openv/volmgr/vm.conf file (the syntax
must be as specified in all capital letters):
AUTO_PATH_CORRECTION=YES

The AUTO_PATH_CORRECTION entry directs NetBackup to scan for device paths

when the ltid device daemon starts.
2 After adding the entry to the vm.conf file and with no current jobs running on
the media server, run the following command:
/usr/openv/volmgr/bin/stopltid
 HP-UX 32
About HP-UX legacy device drivers and files

3 Wait a few minutes for the service to stop, then restart ltid by running the
following command:
/usr/openv/volmgr/bin/ltid

Upon start-up, ltid scans for device paths, adds the new DSFs, and then
purges the legacy DSFs from your NetBackup configuration for the media
server.
After ltid starts, only the new persistent DSF paths should be configured in
NetBackup.
4 After the services are started and device paths updated, you can (but do not
have to) remove the AUTO_PATH_CORRECTION=YES entry from the vm.conf file.

Creating persistent DSF pass-through paths

NetBackup creates pass-through paths for all valid /dev/rtape paths. NetBackup
creates the paths during device discovery or when you run the
/usr/openv/volmgr/bin/scan command.

However, you can create them manually. The following example shows how to
create a pass-through device file for a persistent DSF.
To create a pass-through path for a tape drive
◆ Enter the following command (# is the instance number of the device from the
ioscan output):

mksf –P –C tape –I #

For more information about using the HP-UX mksf command, see the man
page.

About HP-UX legacy device drivers and files

Legacy device files are the old style device files that are not part of the HP-UX agile
addressing scheme.
NetBackup does not discover legacy device files. If you use legacy device files,
you must configure devices manually in NetBackup.

Note: NetBackup device discovery finds persistent DSFs only. Therefore, Veritas
recommends that you use persistent DSFs.

See “About device drivers and files for HP-UX persistent DSFs” on page 28.
 HP-UX 33
About HP-UX legacy device drivers and files

About device drivers for legacy device files

The following are the drivers supported:
■ The sctl driver for robotic control.
■ The stape driver for tape drives.
■ For IBM tape drives, the atdd driver. NetBackup requires a minimum atdd driver
level. See the Hardware Compatibility List for IBM atdd driver levels that are
supported and atdd configuration information.
http://www.veritas.com/docs/TECH76495
NetBackup also supports the use of the IBM atdd tape driver on HP-UX 11i v3.
Also see the Operating System Compatibility List for the minimum OS patch
level that is required to run on HP-UX:
http://www.veritas.com/docs/TECH76648

About legacy robotic control device files

For SCSI robotic control, NetBackup can use the /dev/sctl device files. The device
file names have the following format:
/dev/sctl/cCARDtTARGETlLUN c Major 0xIITL00

Where:
■ CARD is the card instance number of the adapter.

■ TARGET is the SCSI ID of the robotic control.

■ LUN is the SCSI logical unit number (LUN) of the robot.

■ Major is the character major number (from the lsdev command).

■ II are two hexadecimal digits that represent the card instance number.

■ T is a hexadecimal digit that represents the SCSI ID of robotic control.

■ L is a hexadecimal digit that represents the SCSI LUN of the robotic control.

A library may have more than one robotic device. Each robotic device requires a
device file.
See “Creating legacy SCSI and FCP robotic controls on HP-UX” on page 36.

About legacy tape drive device files

NetBackup requires the /dev/rmt device files to configure tape drives.
The device file names have the following format:
/dev/rmt/c#t#d#BESTnb
 HP-UX 34
About HP-UX legacy device drivers and files

The following describe the device file names:

■ c# is the card instance number.

■ t# is the SCSI ID.

■ d# is the device LUN.

■ BEST indicates the highest density format and data compression the device
supports.
■ n indicates no rewind on close.

■ b indicates Berkeley-style close.

The following are examples of tape drive device files:

/dev/rmt/c7t0d0BESTnb
/dev/rmt/c7t1d0BESTnb
/dev/rmt/c7t4d0BESTnb
/dev/rmt/c7t5d0BESTnb

See “About creating legacy tape drive device files” on page 43.

About legacy pass-through paths for tape drives

Although NetBackup requires the /dev/rmt device files to configure tape drives,
NetBackup uses pass-through device files for drive access.
On media servers, NetBackup automatically creates pass-through device files if
the appropriate/dev/rmt tape drive device files exist. NetBackup creates the
pass-through device files in the /dev/sctl directory.
NetBackup does not modify or delete any existing pass-through paths.
NetBackup does not detect the type of adapter cards that are installed in the system.
Therefore, NetBackup creates pass-through paths for tape drives connected to the
adapter cards that do not support pass through. These pass-through paths do not
cause problems.
Although NetBackup uses the pass-through device files during tape drive operations,
you specify the /dev/rmt device files when you configure the drives in NetBackup.
NetBackup then uses the appropriate pass-through device files.
Usually, you do not have to create pass-through paths for drives. However,
instructions to do so are provided for reference.
NetBackup SAN clients require legacy pass-through device files.
See “Creating device files for SAN Clients on HP-UX” on page 35.
 HP-UX 35
Creating device files for SAN Clients on HP-UX

Note: Pass-through paths are not supported on HP-PB adapters such asHP28696A
- Wide SCSI or HP 28655A - SE SCSI.

See “Creating tape drive pass-through device files” on page 43.

See “About device drivers and files for HP-UX persistent DSFs” on page 28.

Creating device files for SAN Clients on HP-UX

NetBackup SAN clients use tape drivers and SCSI pass-through methods for Fibre
Transport traffic to NetBackup FT media servers. On HP-UX systems, NetBackup
SAN clients require the sctl driver and pass-through tape drive device files.
The following table describes the tasks that create the device files. Before you
create the device files, the NetBackup FT media server must be active and the SAN
must be zoned correctly as described in the NetBackup SAN Client and Fibre
Transport Guide.

Table 3-2 SAN Client device file tasks

Step Action Description

Step 1 If the sctl driver is not the default See the HP-UX scsi_ctl(7) man
pass-through driver on your system, page.
install and configure the sctl driver.

Step 2 Create the pass-through paths required. See “About legacy pass-through paths
for tape drives” on page 34.

See “Creating tape drive pass-through

device files” on page 43.

The media server FT devices appear as ARCHIVE Python tape devices during SCSI
inquiry from the SAN client. However, they are not tape devices and do not appear
as tape devices in NetBackup device discovery.

About configuring legacy device files

You can use legacy device files for the following:
■ Robotic control using SCSI or Fibre Channel Protocol control.
SCSI control includes Fibre Channel Protocol (FCP), which is SCSI over Fibre
Channel. A robotic device in a library moves the media between storage slots
and the drives in the library.
See “Creating legacy SCSI and FCP robotic controls on HP-UX” on page 36.
 HP-UX 36
About configuring legacy device files

■ Tape drive read and write access.

See “About creating legacy tape drive device files” on page 43.
See “Creating tape drive pass-through device files” on page 43.
■ SAN client pass-through paths for Fibre Transport traffic to NetBackup media
servers.
See “Creating device files for SAN Clients on HP-UX” on page 35.

Creating legacy SCSI and FCP robotic controls on HP-UX

You must create the robotic control device files for the sctl driver manually; they
are not created automatically when the system boots.
Before you create the device files, you must do the following:
■ Install and configure the sctl driver. For more information, see the HP-UX
scsi_ctl(7) man page.
The sctl driver may be the default pass-through driver on your system. If so,
you do not have to configure the kernel to use the sctl pass-through driver.
■ Install and configure the schgr device driver. For more information, see the
HP-UX autochanger(7) man page.
■ Attach the devices.
Examples of how to create the device files are available.
See “Example of how to create a sctl device file for SCSI (PA-RISC)” on page 37.
See “Example of how to create a sctl device file for FCP (PA-RISC)” on page 39.
See “Example of how to create sctl device files for FCP (Itanium)” on page 41.
To create sctl device files
1 Invoke the ioscan -f command to obtain SCSI bus and robotic control
information.
2 Examine the output for the card instance number and the SCSI ID and LUN
of the robotic device, as follows:
■ The instance number of the card is in the I column of the output.
■ The H/W Path column of the changer output (schgr) includes the SCSI ID
and LUN. Use the card's H/W Path value to filter the changer's H/W Path
entry; the SCSI ID and the LUN remain.
 HP-UX 37
About configuring legacy device files

3 Determine the character major number of the sctl driver by using the following
command:

lsdev -d sctl

Examine the output for an entry that shows sctl in the Driver column.
4 Use the following commands to create the device file for the SCSI robotic
control:

mkdir /dev/sctl
cd /dev/sctl
/usr/sbin/mknod cCARDtTARGETlLUN c Major 0xIITL00

Where:
■ CARD is the card instance number of the adapter.

■ TARGET is the SCSI ID of the robotic control.

■ LUN is the SCSI logical unit number (LUN) of the robot.

■ Major is the character major number (from the lsdev command).

■ II are two hexadecimal digits that represent the card instance number.

■ T is a hexadecimal digit that represents the SCSI ID of robotic control.

■ L is a hexadecimal digit that represents the SCSI LUN of the robotic control.

Example of how to create a sctl device file for SCSI

(PA-RISC)
In this example, the following robots exist:
■ An ADIC Scalar 100 library is on a SCSI bus with an instance number of 7, SCSI
ID 2, and LUN 0.
■ The robotic control for an IBM ULT3583-TL library is on the same SCSI bus at
SCSI ID 3 and LUN 0.
 HP-UX 38
About configuring legacy device files

To create SCSI robotic device files for HP-UX PA-RISC

1 Invoke the ioscan -f command, as follows:

ioscan -f
Class I H/W Path Driver S/W State H/W Type Description
=================================================================
ext_bus 7 0/7/0/1 c720 CLAIMED INTERFACE SCSI C896 Fast Wide LVD
target 10 0/7/0/1.0 tgt CLAIMED DEVICE
tape 65 0/7/0/1.0.0 stape CLAIMED DEVICE QUANTUM SuperDLT1
target 11 0/7/0/1.1 tgt CLAIMED DEVICE
tape 66 0/7/0/1.1.0 stape CLAIMED DEVICE QUANTUM SuperDLT1
target 12 0/7/0/1.2 tgt CLAIMED DEVICE
autoch 14 0/7/0/1.2.0 schgr CLAIMED DEVICE ADIC Scalar 100
target 13 0/7/0/1.3 tgt CLAIMED DEVICE
autoch 19 0/7/0/1.3.0 schgr CLAIMED DEVICE IBM ULT3583-TL
target 14 0/7/0/1.4 tgt CLAIMED DEVICE
tape 21 0/7/0/1.4.0 atdd CLAIMED DEVICE IBM ULT3580-TD1
target 15 0/7/0/1.5 tgt CLAIMED DEVICE
tape 19 0/7/0/1.5.0 atdd CLAIMED DEVICE IBM ULT3580-TD1

2 Examine the output for the card instance number and the SCSI ID and LUN
of the robotic device, as follows:
The card H/W Path is 0/7/0/1; the card instance number (I column) is 7. Apply
the H/W Path value as a mask. The ADIC robotic device (schgr) is at SCSI ID
2, LUN 0 on this bus. The IBM robotic device (schgr) is at SCSI ID 3, LUN 0
on this bus.
 HP-UX 39
About configuring legacy device files

3 Determine the character major number of the sctl driver by using the following
command:

lsdev -d sctl
Character Block Driver Class
203 -1 sctl ctl

The output from this command shows that the character major number for the
sctl driver is 203.

4 The commands to create the device files follow. For the ADIC robot, the card
instance number is 7, the target is 2, and the LUN is 0. For the IBM robot, the
card instance number is 7, the SCSI ID is 3, and the LUN is 0.

cd /dev/sctl
/usr/sbin/mknod c7t2l0 c 203 0x072000
/usr/sbin/mknod c7t3l0 c 203 0x073000

If you add the robots to NetBackup manually, you specify the following for ADIC
robotic control and IBM robotic control respectively:

/dev/sctl/c7t2l0
/dev/sctl/c7t3l0

Example of how to create a sctl device file for FCP

(PA-RISC)
The following example shows how create a sctl device file for an HP VLS9000 robot.
NetBackup uses the device file for robotic control.
 HP-UX 40
About configuring legacy device files

To create an FCP robotic device file for HP-UX PA-RISC

1 Invoke the ioscan -f command. The following output example is edited for
readability:

ioscan -f
Class I H/W Path Driver S/W State H/W Type Description
=============================================================================
fc 0 0/2/0/0 td CLAIMED INTERFACE HP Tachyon XL2 Fibre
Channel Mass Storage
Adapter
fcp 4 0/2/0/0.10 fcp CLAIMED INTERFACE FCP Domain
ext_bus 6 0/2/0/0.10.11.255.0 fcpdev CLAIMED INTERFACE FCP Device Interface
target 5 0/2/0/0.10.11.255.0.0 tgt CLAIMED DEVICE
autoch 2 0/2/0/0.10.11.255.0.0.0 schgr CLAIMED DEVICE HP VLS
tape 5 0/2/0/0.10.11.255.0.0.1 stape CLAIMED DEVICE HP Ultrium 4-SCSI
tape 6 0/2/0/0.10.11.255.0.0.2 stape CLAIMED DEVICE HP Ultrium 4-SCSI
tape 7 0/2/0/0.10.11.255.0.0.3 stape CLAIMED DEVICE HP Ultrium 4-SCSI

2 Examine the output for the card instance number and the SCSI ID and LUN
of the robotic device. In this example, the interface card instance number (the
I column) is 6. If you use the card's H/W Path value as a mask
(0/2/0/0.10.11.255.0), you see the following:
■ An HP VLS9000 robot is at SCSI ID 0, LUN 0.
■ Three Ultrium 4-SCSI drives are at SCSI ID 0 and LUN 1, LUN 2, and LUN
3.
 HP-UX 41
About configuring legacy device files

3 Determine the character major number of the sctl driver by using the lsdev
command, as follows:

lsdev -d sctl
Character Block Driver Class
203 -1 sctl ctl

The output from this command shows that the character major number for the
sctl driver is 203.

4 The commands to create the device file for the HP VLS9000 robotic control
are as follows. The card instance number is 6, the target is 0, and the LUN is
0.

cd /dev/sctl
/usr/sbin/mknod c6t0l0 c 203 0x060000

If you add the robot to NetBackup manually, specify the following pathname
for robotic control:

/dev/sctl/c6t0l0

Example of how to create sctl device files for FCP (Itanium)

With Fibre Channel, the hardware paths are longer than with SCSI.
In this example, the following devices are attached to the host.
■ An HP EML E-Series robot with four HP drives (two LTO2 and two LTO3 drives).
A separate path exists for each drive pair. The robotic control is through card
instance 12 (0/4/1/1.2.12.255.0).
■ An HP VLS 6000 robot with six drives. The robot is partitioned into two virtual
libraries, three Quantum SDLT320 drives in one library and three HP LTO3
drives in the other library. Separate robotic control exists for each library.
 HP-UX 42
About configuring legacy device files

To create FCP robotic device files for HP-UX Itanium

1 Invoke the ioscan -f command. The following is a command output excerpt
that shows the Fibre Channel devices on a host:

ext_bus 4 0/4/1/1.2.10.255.0 fcd_vbus CLAIMED INTERFACE FCP Device Interface

target 7 0/4/1/1.2.10.255.0.0 tgt CLAIMED DEVICE
tape 18 0/4/1/1.2.10.255.0.0.0 stape CLAIMED DEVICE HP Ultrium 3-SCSI
tape 20 0/4/1/1.2.10.255.0.0.1 stape CLAIMED DEVICE HP Ultrium 3-SCSI
ext_bus 13 0/4/1/1.2.11.255.0 fcd_vbus CLAIMED INTERFACE FCP Device Interface
target 8 0/4/1/1.2.11.255.0.0 tgt CLAIMED DEVICE
autoch 4 0/4/1/1.2.11.255.0.0.0 schgr CLAIMED DEVICE HP VLS
tape 22 0/4/1/1.2.11.255.0.0.1 stape CLAIMED DEVICE QUANTUM SDLT320
tape 23 0/4/1/1.2.11.255.0.0.2 stape CLAIMED DEVICE QUANTUM SDLT320
tape 24 0/4/1/1.2.11.255.0.0.3 stape CLAIMED DEVICE QUANTUM SDLT320
autoch 5 0/4/1/1.2.11.255.0.0.4 schgr CLAIMED DEVICE HP VLS
tape 25 0/4/1/1.2.11.255.0.0.5 stape CLAIMED DEVICE HP Ultrium 3-SCSI
tape 26 0/4/1/1.2.11.255.0.0.6 stape CLAIMED DEVICE HP Ultrium 3-SCSI
tape 27 0/4/1/1.2.11.255.0.0.7 stape CLAIMED DEVICE HP Ultrium 3-SCSI
ext_bus 12 0/4/1/1.2.12.255.0 fcd_vbus CLAIMED INTERFACE FCP Device Interface
target 6 0/4/1/1.2.12.255.0.0 tgt CLAIMED DEVICE
autoch 1 0/4/1/1.2.12.255.0.0.0 schgr CLAIMED DEVICE HP EML E-Series
tape 19 0/4/1/1.2.12.255.0.0.1 stape CLAIMED DEVICE HP Ultrium 2-SCSI
tape 21 0/4/1/1.2.12.255.0.0.2 stape CLAIMED DEVICE HP Ultrium 2-SCSI

2 Examine the output for the card instance number and the SCSI ID and LUN
of the robotic device.
In this example, the following devices are attached to this host:
■ The robotic control for the HP EML E-Series robot is through card instance
12 (0/4/1/1.2.12.255.0). Two of the drives are accessed through the same
path, and the other two are accessed through card instance 4
(0/4/1/1.2.10.255.0).
■ The robotic controls for the HP VLS 6000 robot partitions are through card
instance 13. Robotic control for one partition is at SCSI ID 0 and LUN 0.
Robotic control for the other partition is at SCSI ID 0 and LUN 4.
 HP-UX 43
About configuring legacy device files

3 Determine the character major number of the sctl driver by using the following
command:

lsdev -d sctl
Character Block Driver Class
203 -1 sctl ctl

The output from this command shows that the character major number for the
sctl driver is 203.

4 The commands to create the devices file for the robotic controls are as follows:

cd /dev/sctl
/usr/sbin/mknod c12t0l0 c 203 0x0c0000
/usr/sbin/mknod c13t0l0 c 203 0x0d0000
/usr/sbin/mknod c13t0l4 c 203 0x0d0400

If you add the robots to NetBackup manually, you specify the following
pathnames for robotic control. The first device file is for the HP EML E-Series
robot. The second and third device files are for the VLS 6000 robot (two robotic
devices).

/dev/sctl/c12t0l0
/dev/sctl/c13t0l0
/dev/sctl/c13t0l4

About creating legacy tape drive device files

By default, HP-UX creates tape drive device files when the system is booted.
However, the tape driver must be installed and configured, and the devices must
be attached and operational.
Alternatively, you can create tape drive device files manually. To do so, use either
the HP-UX System Administration Manager (SAM) utility or the insf(1M) command.
For information, see the HP-UX documentation.

Creating tape drive pass-through device files

On media servers, NetBackup creates pass-through paths for tape drives
automatically. However, you can create them manually.
NetBackup also uses the tape drive pass-through device files for SAN Client.
Use one of the following two procedures:
■ Create pass-through tape drive device files
See “To create pass-through tape drive device files” on page 44.
 HP-UX 44
About configuring legacy device files

■ Create SAN client pass-through device files

See “To create SAN client pass-through device files” on page 46.
To create pass-through tape drive device files
1 Determine the devices that are attached to the SCSI bus by using the HP-UX
ioscan -f command, as follows:

ioscan -f
Class I H/W Path Driver S/W State H/W Type Description
=================================================================
ext_bus 7 0/7/0/ c720 CLAIMED INTERFACE SCSI C896 Fast Wide LVD
target 10 0/7/0/1.0 tgt CLAIMED DEVICE
tape 65 0/7/0/1.0.0 stape CLAIMED DEVICE QUANTUM SuperDLT1
target 11 0/7/0/1.1 tgt CLAIMED DEVICE
tape 66 0/7/0/1.1.0 stape CLAIMED DEVICE QUANTUM SuperDLT1
target 12 0/7/0/1.2 tgt CLAIMED DEVICE
autoch 14 0/7/0/1.2.0 schgr CLAIMED DEVICE ADIC Scalar 100
target 13 0/7/0/1.3 tgt CLAIMED DEVICE
autoch 19 0/7/0/1.3.0 schgr CLAIMED DEVICE IBM ULT3583-TL
target 14 0/7/0/1.4 tgt CLAIMED DEVICE
tape 21 0/7/0/1.4.0 atdd CLAIMED DEVICE IBM ULT3580-TD1
target 15 0/7/0/1.5 tgt CLAIMED DEVICE
tape 19 0/7/0/1.5.0 atdd CLAIMED DEVICE IBM ULT3580-TD1

This example output shows the following:

■ The robotic control for an ADIC Scalar 100 library is on a SCSI bus with an
instance number of 7. The SCSI ID is 2, and the LUN is 0. The robotic
control for an IBM ULT3583-TL library is on the same SCSI bus at SCSI
ID 3 and LUN 0.
■ The ADIC library contains two Quantum Super DLT drives. One has a SCSI
ID of 0 and a LUN of 0. The other has a SCSI ID of 1 and a LUN of 0.
■ The IBM library contains two IBM Ultrium LTO drives. One has a SCSI ID
of 4 and a LUN of 0. The other has a SCSI ID of 5 and a LUN of 0.
Use the IBM atdd driver when you configure IBM tape drives on HP-UX.
Configure atdd and BEST device paths according to the IBM driver
documentation. Do not configure atdd for robotic control of IBM robots. For
 HP-UX 45
About configuring legacy device files

the latest recommended atdd driver version from IBM, check the Veritas
support Web site.

2 Create the pass-through device files for the tape drives, as follows:

cd /dev/sctl
/usr/sbin/mknod c7t0l0 c 203 0x070000
/usr/sbin/mknod c7t1l0 c 203 0x071000
/usr/sbin/mknod c7t4l0 c 203 0x074000
/usr/sbin/mknod c7t5l0 c 203 0x075000

When you use the HP-UX mknod command for tape drives, the target is the
SCSI ID of the tape drive. It is not the SCSI ID of the robotic control.
The previous commands create the following pass-through device files.

/dev/sctl/c7t0l0
/dev/sctl/c7t1l0
/dev/sctl/c7t4l0
/dev/sctl/c7t5l0

Although the pass-through device files for tape drives are used during
NetBackup operation, they are not used during NetBackup configuration. During
NetBackup tape drive configuration, use the following device files to configure
the tape drives.

/dev/rmt/c7t0d0BESTnb
/dev/rmt/c7t1d0BESTnb
/dev/rmt/c7t4d0BESTnb
/dev/rmt/c7t5d0BESTnb
 HP-UX 46
About configuring legacy device files

To create SAN client pass-through device files

1 Determine the devices that are attached to the SCSI bus by using the HP-UX
ioscan -f command, as follows:

ioscan -f
Class I H/W Path Driver S/W State H/W Type Description
=================================================================================
ext_bus 9 0/3/1/0.1.22.255.0 fcd_vbus CLAIMED INTERFACE FCP Device Interface
target 4 0/3/1/0.1.22.255.0.0 tgt CLAIMED DEVICE
tape 6 0/3/1/0.1.22.255.0.0.0 stape CLAIMED DEVICE ARCHIVE Python
tape 7 0/3/1/0.1.22.255.0.0.1 stape CLAIMED DEVICE ARCHIVE Python

This example output shows that the instance number of the Fibre Channel
HBA is 9. It also shows that the target mode drivers on the Fibre Transport
media server appear as ARCHIVE Python devices. One has a SCSI ID of 0
and a LUN of 0; the other has a SCSI ID of 0 and a LUN of 1.
2 Determine the character major number of the sctl driver by using the following
command:

lsdev -d sctl
Character Block Driver Class
203 -1 sctl ctl

The output from this command shows that the character major number for the
sctl driver is 203.
3 Create the pass-through device files, as follows:

cd /dev/sctl
/usr/sbin/mknod c9t0l0 c 203 0x090000
/usr/sbin/mknod c9t0l1 c 203 0x090100

The following describes the device file name:

■ c9 defines the instance number of the interface card.

■ t0 defines the SCSI ID (the target).

■ l1 defines the LUN (the first character is the letter “l”).

4 Verify that the device files were created, as follows:

# ls -l /dev/sctl
total 0
crw-r--r-- 1 root sys 203 0x090000 Nov 1 13:19 c9t0l0
crw-r--r-- 1 root sys 203 0x090100 Nov 1 13:19 c9t0l1
 HP-UX 47
About SPC-2 SCSI reserve on HP-UX

About SPC-2 SCSI reserve on HP-UX

By default, NetBackup uses SPC-2 SCSI reserve and release for tape drive
reservations in shared drive environments. However, you must disable SPC-2 SCSI
reserve and release in HP-UX. If you do not, conflicts occur between the operating
system and NetBackup. The NetBackup Shared Storage Option provides shared
drive functionality in NetBackup.
Veritas recommends that you use the SAM utility to disable SPC-2 SCSI reserve
and release.
As an alternative to SCSI reserve and release, you can use SCSI persistent reserve
in NetBackup for shared tape drive reservations. If you use SCSI persistent reserve,
you still must disable SPC-2 SCSI reserve and release in HP-UX.
For more information about NetBackup and SCSI reservations, see the following:
■ The description of the Enable SCSI Reserve Media host property in the
NetBackup Administrator’s Guide, Volume I.
■ The "How NetBackup reserves drives" topic in the NetBackup Administrator’s
Guide, Volume II.

Disabling SPC-2 SCSI reserve in HP-UX

Use the following procedure to disable SPC-2 SCSI reserve in HP-UX.
To disable SPC-2 SCSI reserve in HP-UX
1 Set the st_ats_enabled kernel parameter to 0 (zero) .
2 Reboot the system.

About disabling the HP-UX EMS Tape Device

Monitor for a SAN
You should configure the Tape Device Monitor (dm_stape) so it does not run on
HP-UX hosts in a SAN configuration. The Tape Device Monitor is a component of
the Event Monitoring System (EMS). The EMS service periodically polls the tape
devices to monitor their conditions. When a server polls the devices while another
server uses a tape device, backup operations may time out and fail.
You can avoid the situation as follows:
■ To disable EMS completely, run the HP-UX Hardware Monitoring Request
Manager and select (K) kill (disable) Monitoring.
 HP-UX 48
HP-UX command summary

Invoke the Hardware Monitoring Request Manager by using the

/etc/opt/resmon/lbin/monconfig command.

■ To configure EMS so it does not log any events or poll devices, set the
POLL_INTERVAL value to 0 (zero). The POLL_INTERVAL parameter is in the
following HP-UX configuration file:
/var/stm/config/tools/monitor/dm_stape.cfg
EMS runs but does not send any SCSI commands.

HP-UX command summary

The following is a summary of commands that may be useful when you configure
and verify devices. For usage examples, see the procedures in this chapter.
■ ioscan -C class -f
Shows the information about the physical interfaces. Numeric information is
displayed in decimal.
class is the type of interface, as follows:
■ tape specifies tape drives.

■ ext_bus specifies SCSI controllers.

■ mknod /dev/spt/cCARDtTARGETlLUN c Major 0xIITL00

Creates the device files for SCSI robotic or tape drive controls.
The following describe the device file names:
■ CARD is the card instance number of the adapter.

■ TARGET is the SCSI ID of the robotic control.

■ LUN is the SCSI logical unit number (LUN) of the robot.

■ Major is the character major number (from the lsdev command).

■ II are two hexadecimal digits that represent the card instance number.

■ T is a hexadecimal digit that represents the SCSI ID of robotic control.

■ L is a hexadecimal digit that represents the SCSI LUN of the robotic control.

■ lsdev -d driver
Displays the information about the SCSI robotic control drivers.
■ mksf -C tape -H hw-path -b BEST -u -n
Creates the device files for tape drives. The hw-path is the hardware path of
the tape drive, as specified by the ioscan command.
 Chapter 4
Linux
This chapter includes the following topics:

■ Before you begin on Linux

■ About the required Linux SCSI drivers

■ Verifying the Linux drivers

■ About configuring robot and drive control for Linux

■ Verifying the device configuration on Linux

■ About SAN clients on Linux

■ About SCSI persistent bindings for Linux

■ About Emulex HBAs

■ Utilities to test SCSI devices

■ Linux command summary

Before you begin on Linux

Observe the following important points when you configure the operating system:
■ Verify that NetBackup supports your server platform and devices. The Veritas
support Web site contains server platform compatibility information. For the
compatibility information, see the NetBackup compatibility lists on the NetBackup
landing page on the Veritas Support website.
■ For SCSI controlled libraries, NetBackup issues SCSI commands to the robotic
devices. For NetBackup to function correctly, the properly named device files
must exist. Information about how to configure device files is available.
See “About configuring robot and drive control for Linux” on page 51.
 Linux 50
About the required Linux SCSI drivers

■ Verify that a SCSI low-level driver is installed for each HBA in your system, as
follows:
■ Follow the HBA vendor's installation guide to install or load the driver in the
kernel.
■ Configure the kernel for SCSI tape support and SCSI generic support.
■ Probe all LUNs on each SCSI device and enable the SCSI low-level driver
for the HBA.
■ Enable multi-LUN support for the kernel according to the Linux
documentation.
For more information, refer to your HBA vendor documentation.
■ Multipath configurations (multiple paths to robots and drives) are supported only
with the following configurations:
■ Native path (/dev/nstx, /dev/sgx)
■ The sysfs file system that is mounted on /sys

After you configure the hardware, add the robots and the drives to NetBackup.

About the required Linux SCSI drivers

To use SCSI tape drives and robotic libraries, the following drivers must be
configured in the kernel or loaded as modules:
■ Linux SCSI generic (sg) driver. This driver allows pass-through commands to
SCSI tape drives and control of robotic devices.
If you do not use a pass-through driver, performance suffers. NetBackup and
its processes use the pass-through driver as follows:
■ To scan drives
■ For SCSI reservations
■ For SCSI locate-block operations
■ For SAN error recovery
■ For Quantum SDLT performance optimization
■ To collect robot and drive information
■ To collect Tape Alert information from tape drives
■ For WORM tape support
■ For future features and enhancements
 Linux 51
Verifying the Linux drivers

■ SCSI tape (st) driver. This driver allows the use of SCSI tape drives.
■ Standard SCSI driver.
■ SCSI-adapter driver.
The standard Enterprise Linux releases have the sg and the st modules available
for loading. The modules are loaded as needed. Also, you can load these modules
if they are not in the kernel. Use the following commands:

/sbin/modprobe st
/sbin/modprobe sg

About the st driver debug mode

You can enable debug mode for the st tape driver. Debug mode echoes each
command and its result to the system log. For details, see the Linux documentation.

Verifying the Linux drivers

NetBackup requires specific Linux drivers.
See “About the required Linux SCSI drivers” on page 50.
You can use the /sbin/lsmod command to verify that the st and the sg drivers are
loaded in the kernel.
To verify that the drivers are installed and loaded in the kernel
◆ Invoke the lsmod command as follows:

lsmod
Module Size Used by
sg 14844 0
st 24556 0

About configuring robot and drive control for

Linux
NetBackup supports SCSI control and API control of robotic devices. SCSI control
includes Fibre Channel Protocol (FCP), which is SCSI over Fibre Channel.
You must configure the control method, as follows:
■ SCSI or Fibre Channel Protocol control.
 Linux 52
Verifying the device configuration on Linux

NetBackup uses device files to configure control for SCSI tape devices, including
robotic devices. (A robotic device in a library moves the media between storage
slots and the drives in the library.)
See “About the Linux robotic control device files ” on page 52.
See “About the Linux tape drive device files” on page 52.
■ API control over a LAN.
See the "ADIC Automated Media Library (AML)" topic of this guide.
See the "IBM Automated Tape Library (ATL)" topic of this guide.
See the "Sun StorageTek ACSLS robots" topic of this guide.

About the Linux robotic control device files

For robotic devices, NetBackup uses the /dev/sgx device files, where x is a decimal
number from 0 to 255. Linux should create the device files automatically. If the
device files do not exist, see the Linux documentation for information about how to
create them.
If you use device discovery, NetBackup looks for /dev/sgx robotic control device
files. NetBackup discovers the robotic control device files (and hence the devices)
automatically. Alternatively, if you add a robot manually in NetBackup, you must
enter the pathname to the device file for that robotic device.

About the Linux tape drive device files

For tape drive device files, NetBackup uses the /dev/nstx files (n indicates the no
rewind device file). The Linux driver should create the device files automatically. If
the device files do not exist, see the Linux documentation for information about how
to create them.
If you use device discovery in NetBackup, NetBackup looks for /dev/nstx device
files. NetBackup discovers the device files (and hence the devices) automatically.
Alternatively, if you add a drive manually in NetBackup, you must enter the pathname
to the device file for that drive.
The NetBackup avrd daemon establishes a default tape driver operating mode. If
you change the default mode, NetBackup may not read and write tapes correctly,
which results in data loss.

Verifying the device configuration on Linux

The /proc/scsi/scsi file shows all of the devices that the SCSI driver detects.
If the operating system detects the SCSI devices, NetBackup can discover them.
 Linux 53
About SAN clients on Linux

To verify that the operating system can see the devices

◆ Run the following command from a terminal window:
cat /proc/scsi/scsi

The output that is displayed should be similar to the following:

Attached devices:
Host: scsi0 Channel: 00 Id: 01 Lun: 00
Vendor: HP Model: C7200-8000 Rev: 1040
Type: Medium Changer ANSI SCSI revision: 03
Host: scsi0 Channel: 00 Id: 02 Lun: 00
Vendor: QUANTUM Model: DLT8000 Rev: 010F
Type: Sequential-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 03 Lun: 00
Vendor: QUANTUM Model: DLT8000 Rev: 010F
Type: Sequential-Access ANSI SCSI revision: 02

About SAN clients on Linux

NetBackup SAN clients on Linux hosts require the SCSI Generic (sg) driver and
pass-through tape drive device files for traffic to NetBackup FT media servers. The
media server FT devices appear as ARCHIVE Python tape devices during SCSI
inquiry from the SAN client. (However, they are not tape devices and do not appear
as tape devices in NetBackup device discovery.)
You should verify that you have the correct driver and device files.
See “Verifying the Linux drivers” on page 51.
If your Linux operating system does not add all of the SCSI device files automatically,
you can do so manually. The following is an example of code you can include in
the /etc/rc.local file to add LUN 1, targets 0-7 on Controllers 0-2. Note that the
last line is the MAKEDEV command, which makes the required device files. The code
you include in your /etc/rc.local file depends on how your hardware environment.

# Add the troublesome device on LUN 1 for the FT server

echo "scsi add-single-device 0 0 0 1" > /proc/scsi/scsi
echo "scsi add-single-device 0 0 1 1" > /proc/scsi/scsi
echo "scsi add-single-device 0 0 2 1" > /proc/scsi/scsi
echo "scsi add-single-device 0 0 3 1" > /proc/scsi/scsi
echo "scsi add-single-device 0 0 4 1" > /proc/scsi/scsi
echo "scsi add-single-device 0 0 5 1" > /proc/scsi/scsi
echo "scsi add-single-device 0 0 6 1" > /proc/scsi/scsi
echo "scsi add-single-device 0 0 7 1" > /proc/scsi/scsi
 Linux 54
About SCSI persistent bindings for Linux

echo "scsi add-single-device 1 0 0 1" > /proc/scsi/scsi

echo "scsi add-single-device 1 0 1 1" > /proc/scsi/scsi
echo "scsi add-single-device 1 0 2 1" > /proc/scsi/scsi
echo "scsi add-single-device 1 0 3 1" > /proc/scsi/scsi
echo "scsi add-single-device 1 0 4 1" > /proc/scsi/scsi
echo "scsi add-single-device 1 0 5 1" > /proc/scsi/scsi
echo "scsi add-single-device 1 0 6 1" > /proc/scsi/scsi
echo "scsi add-single-device 1 0 7 1" > /proc/scsi/scsi
echo "scsi add-single-device 2 0 0 1" > /proc/scsi/scsi
echo "scsi add-single-device 2 0 1 1" > /proc/scsi/scsi
echo "scsi add-single-device 2 0 2 1" > /proc/scsi/scsi
echo "scsi add-single-device 2 0 3 1" > /proc/scsi/scsi
echo "scsi add-single-device 2 0 4 1" > /proc/scsi/scsi
echo "scsi add-single-device 2 0 5 1" > /proc/scsi/scsi
echo "scsi add-single-device 2 0 6 1" > /proc/scsi/scsi
echo "scsi add-single-device 2 0 7 1" > /proc/scsi/scsi
/dev/MAKEDEV sg

About SCSI persistent bindings for Linux

Veritas recommends that you use persistent bindings to lock the mappings between
the SCSI targets that are reported to Linux and the specific devices. The Linux
kernel device manager udev is often used to configure persistent bindings.
If you cannot use binding with the HBA in your configuration, add an
ENABLE_AUTO_PATH_CORRECTION entry in the /usr/openv/volmgr/vm.conf file on
all Linux media servers.

About Emulex HBAs

If you use a /usr/openv/volmgr/AVRD_DEBUG touch file on a system with an Emulex
HBA driver, the system log may contain entries similar to the following:

Unknown drive error on DRIVENAME (device N, PATH) sense[0] = 0x70,

sense[1] = 0x0, sensekey = 0x5

You can ignore these messages.

Utilities to test SCSI devices

You can manipulate tape devices with the operating system mt command. For more
information, see the mt(1) man page.
 Linux 55
Linux command summary

You can use theNetBackup robtest utility to test robots. The robtest utility resides
in /usr/openv/volmgr/bin.
A set of SCSI utilities are available from the Linux SCSI Generic (sg) driver home
page.

Linux command summary

The following is a summary of commands that were used in this topic:
■ /sbin/lsmod
Lists the modules that are loaded.
■ /sbin/modprobe
Installs loadable kernel modules.
■ /usr/sbin/reboot
Stops and restarts the system.
■ /bin/mknod /dev/sgx c 21 N
Creates SCSI generic device files; x is a decimal number from 0 to 255.
 Chapter 5
Solaris
This chapter includes the following topics:

■ Before you begin on Solaris

■ About the NetBackup sg driver

■ Determining if the NetBackup sg driver is installed

■ Special configuration for the Oracle StorEdge Network Foundation HBA driver

■ About binding Fibre Channel HBA drivers

■ Configuring Solaris 10 x86 for multiple drive paths

■ Installing/reinstalling the sg and the st drivers

■ Configuring 6 GB and larger SAS HBAs in Solaris

■ Preventing Solaris driver unloading

■ About Solaris robotic controls

■ About Solaris tape drive device files

■ Configuring Solaris SAN clients to recognize FT media servers

■ Uninstalling the sg driver on Solaris

■ Solaris command summary

Before you begin on Solaris

Observe the following points when you configure the operating system:
■ Verify that NetBackup supports your server platform and devices. Download
the NetBackup hardware and the operating system compatibility lists.
 Solaris 57
About the NetBackup sg driver

See “About the NetBackup compatibility lists” on page 10.

■ For SCSI controlled libraries, NetBackup issues SCSI commands to the robotic
devices.
For NetBackup to function correctly, the properly named device files must exits,
as follows:
■ NetBackup installs its own pass-through driver, the SCSI generic sg driver.
You must configure this driver properly to create device files for any device
NetBackup uses.
■ The Solaris tape and disk driver interfaces also create a device file for each
tape drive device. These device files must exist for all read or write I/O
capability.
See “About Solaris robotic controls” on page 68.
■ Verify that the Solaris st driver is installed.
■ Verify that the devices are configured correctly. To do so, use the Solaris mt
command and the NetBackup /usr/openv/volmgr/bin/sgscan utility.
For the tape drives that you want to share among NetBackup hosts, ensure that
the operating system detects the devices on the SAN.
■ When you configure devices, you should attach all peripherals and restart the
system with the reconfigure option (boot -r or reboot -- -r).
■ If you remove or replace adapter cards, remove all device files that are
associated with that adapter card.
■ If you use the Automated Cartridge System (ACS) robotic software, you must
ensure that the Solaris Source Compatibility Package is installed. The package
is required so that the ACS software can use the shared libraries in /usr/ucblib.
■ Oracle systems with parallel SCSI host bus adapters do not support 16-byte
SCSI commands on any devices that are attached to these HBAs. Therefore,
those HBAs do not support WORM media. To override this limitation, create a
touch file as follows:
touch /usr/openv/volmgr/database/SIXTEEN_BYTE_CDB

After you configure the hardware, add the robots and the drives to NetBackup.

About the NetBackup sg driver

NetBackup provides its own SCSI pass-through driver to communicate with
SCSI-controlled robotic peripherals. This driver is called the SCSA (generic SCSI
pass-through driver), also referred to as the sg driver.
 Solaris 58
Determining if the NetBackup sg driver is installed

For full feature support, NetBackup requires the sg driver and SCSI pass-through
device paths.
Install the NetBackup sg driver on each Solaris NetBackup media server that hosts
tape devices. Each time you add or remove a device, you should reinstall the sg
driver again.
If you do not use a pass-through driver, performance suffers.
NetBackup uses the pass-through driver for the following:
■ By avrd and robotic processes to scan drives.
■ By NetBackup to position tapes by using the locate-block method.
■ By NetBackup for SAN error recovery.
■ By NetBackup for Quantum SDLT performance optimization.
■ By NetBackup for SCSI reservations.
■ By NetBackup device configuration to collect robot and drive information.
■ To collect Tape Alert information from tape devices allowing support of functions
such as tape drive cleaning.
■ For WORM tape support.
■ Future NetBackup features and enhancements

Note: Because NetBackup uses its own pass-through driver, NetBackup does not
support the Solaris sgen SCSI pass-through driver.

See “Installing/reinstalling the sg and the st drivers” on page 60.

Determining if the NetBackup sg driver is installed

Use the following procedure to determine if the sg driver is installed and loaded.
More information about the driver is available.
See “About the NetBackup sg driver” on page 57.
To determine if the sg driver is installed and loaded
◆ Invoke the following command:
/usr/sbin/modinfo | grep sg

If the driver is loaded, output includes a line similar to the following:

141 fc580000 2d8c 116 1 sg (SCSA Generic Revision: 3.5e)
 Solaris 59
Special configuration for the Oracle StorEdge Network Foundation HBA driver

Special configuration for the Oracle StorEdge

Network Foundation HBA driver
When you configure the sg driver, it binds the StorEdge Network Foundation host
bus adapter World Wide Port Names for use by the sg driver.
See “Installing/reinstalling the sg and the st drivers” on page 60.
The configuration process uses the Solaris luxadm command to probe for HBAs
that are installed in the system. Ensure that the luxadm command is installed and
in the shell path. For Solaris 11 and later, NetBackup uses the Solaris sasinfo
command to probe for SAS attached devices.
To determine if a host contains a StorEdge Network Foundation HBA, you can run
the following command:
/usr/openv/volmgr/bin/sgscan

If the script detects a StorEdge Network Foundation HBA, it produces output similar
to the following example:

#WARNING: detected StorEdge Network Foundation connected devices not

in sg configuration file:
#
# Device World Wide Port Name 21000090a50001c8
#
# See /usr/openv/volmgr/NetBackup_DeviceConfig_Guide.txt topic
# "Special configuration for Sun StorEdge Network Foundation
# HBA/Driver" for information on how to use sg.build and
# sg.install to configure these devices

Each time you add or remove a device, you should configure the NetBackup sg
driver and the Sun st driver again.
See “About the NetBackup sg driver” on page 57.
For 6 GB and larger serial attached SCSI (SAS) HBAs, also configure class 08 and
0101 for the sg driver.
See “Configuring 6 GB and larger SAS HBAs in Solaris” on page 65.

About binding Fibre Channel HBA drivers

For Fibre Channel HBAs other than Sun StorEdge Network Foundation, you must
bind the devices to specific target IDs on the NetBackup host. When you bind
devices to targets, the target ID does not change after a system reboot or a Fibre
Channel configuration change.
 Solaris 60
Configuring Solaris 10 x86 for multiple drive paths

In some instances, Veritas products are configured to use a specific target ID. If
you change the ID, the products fail until you configure the ID correctly.
How you bind devices to targets is vendor and product specific. For information
about how to modify the HBA configuration files to bind devices to targets, see the
documentation for the HBA.
The binding may be based on the following:
■ Fibre Channel World Wide Port Name (WWPN)
■ World Wide Node Name (WWNN)
■ The destination target ID and LUN
After you bind the devices to target IDs, continue with the Solaris configuration in
the same manner as for parallel SCSI installations.
See “Installing/reinstalling the sg and the st drivers” on page 60.
Each time you add or remove a device, you must update the bindings and then
configure the sg and the st drivers again.

Configuring Solaris 10 x86 for multiple drive paths

To use multiple paths to the same tape drive, NetBackup requires that Solaris
Multiplexed I/O (MPxIO) be disabled. MPxIO is enabled by default on Solaris 10
x86 systems.
Use the following procedure to disable MPxIO.
To disable MPxIO
1 Use a text editor to open the following file:
/kernel/drv/fp.conf

2 Change the mpxio-disable value from no to yes. After the change, the line
in the file should appear as follows:
mpxio-disable="yes"

3 Save the changes and exit from the text editor.

Installing/reinstalling the sg and the st drivers

You must install the NetBackup sg driver and the Sun st driver on each Solaris
NetBackup media server that hosts tape devices.
 Solaris 61
Installing/reinstalling the sg and the st drivers

Each time you add or remove a device, you should configure the NetBackup sg
driver and the Sun st driver again. For 6 GB and larger serial-attached SCSI (SAS)
HBAs, also configure class 08 and 0101 for the sg driver.
See “Configuring 6 GB and larger SAS HBAs in Solaris” on page 65.
Before you configure the sg and the st drivers, ensure that all devices are turned
on and connected to the HBA.
See “About the NetBackup sg driver” on page 57.
The sg.build command uses the Solaris sasinfo command to probe for SAS
attached device paths. This command is only available on Solaris 11 and later. On
Solaris 10 and earlier, you must configure the sg driver manually.
To install and configure the sg and the st drivers
1 Invoke the following two commands to run the NetBackup sg.build script:

cd /usr/openv/volmgr/bin/driver
/usr/openv/volmgr/bin/sg.build all -mt target -ml lun

The following describes the options:

■ The all option creates the following files and populates them with the
appropriate entries:
■ /usr/openv/volmgr/bin/driver/st.conf
See “st.conf file example” on page 63.
■ /usr/openv/volmgr/bin/driver/sg.conf
See “sg.conf file example” on page 63.
■ /usr/openv/volmgr/bin/driver/sg.links
See “sg.links file example” on page 64.

■ The -mt target option and argument specify the maximum target ID that
is in use on the SCSI bus (or bound to an FCP HBA). The maximum value
is 126. By default, the SCSI initiator target ID of the adapter is 7, so the
script does not create entries for target ID 7.
 Solaris 62
Installing/reinstalling the sg and the st drivers

■ The -ml lun option and argument specify the maximum number of LUNs
that are in use on the SCSI bus (or by an FCP HBA). The maximum value
is 255.

2 Replace the following seven entries in the /kernel/drv/st.conf file with all
of the entries from the /usr/openv/volmgr/bin/driver/st.conf file:

name="st" class="scsi" target=0 lun=0;

name="st" class="scsi" target=1 lun=0;
name="st" class="scsi" target=2 lun=0;
name="st" class="scsi" target=3 lun=0;
name="st" class="scsi" target=4 lun=0;
name="st" class="scsi" target=5 lun=0;
name="st" class="scsi" target=6 lun=0;

You should make a backup copy of the /kernel/drv/st.conf file before you
modify it.
3 Reboot the system with the reconfigure option (boot -r or reboot -- -r).
During the boot process, the system probes all targets in the st.conf file for
devices. It should create device files for all of the devices it discovers.
4 Verify that Solaris created the device nodes for all the tape devices by using
the following command:
ls -l /dev/rmt/*cbn

5 Install the new sg driver configuration by invoking the following two commands:

/usr/bin/rm -f /kernel/drv/sg.conf
/usr/openv/volmgr/bin/driver/sg.install

The NetBackup sg.install script does the following:

■ Installs and loads the sg driver.
■ Copies the /usr/openv/volmgr/bin/driver/sg.conf file to
/kernel/drv/sg.conf.

■ Creates the /dev/sg directory and nodes.

■ Appends the /usr/openv/volmgr/bin/driver/sg.links file to the
/etc/devlink.tab file.

6 Verify that the <command>sg</command> driver finds all of the robots and
tape drives.
 Solaris 63
Installing/reinstalling the sg and the st drivers

st.conf file example

The following /usr/openv/volmgr/bin/driver/st.conf file example shows targets
0-15 and LUNs 0-7.

name="st" class="scsi" target=0 lun=0;

name="st" class="scsi" target=0 lun=1;
name="st" class="scsi" target=0 lun=2;
name="st" class="scsi" target=0 lun=3;
name="st" class="scsi" target=0 lun=4;
name="st" class="scsi" target=0 lun=5;
name="st" class="scsi" target=0 lun=6;
name="st" class="scsi" target=0 lun=7;
name="st" class="scsi" target=1 lun=0;
name="st" class="scsi" target=1 lun=1;
name="st" class="scsi" target=1 lun=2;
.
<entries omitted for brevity>
.
name="st" class="scsi" target=15 lun=5;
name="st" class="scsi" target=15 lun=6;
name="st" class="scsi" target=15 lun=7;

sg.conf file example

The following /usr/openv/volmgr/bin/driver/sg.conf file example shows targets 0-15
and LUNs 0-8. It also includes target entries for three StorEdge Network Foundation
HBA ports.
The sg.build -mt option does not affect FCP targets, but the -ml option does. The
Solaris luxadm command detected three ports (identified by their World Wide
Names). Therefore, the sg.build script created entries for LUNs 0 through 7 for
those three ports.

name="sg" class="scsi" target=0 lun=0;

name="sg" class="scsi" target=0 lun=1;
name="sg" class="scsi" target=0 lun=2;
name="sg" class="scsi" target=0 lun=3;
name="sg" class="scsi" target=0 lun=4;
name="sg" class="scsi" target=0 lun=5;
name="sg" class="scsi" target=0 lun=6;
name="sg" class="scsi" target=0 lun=7;
name="sg" class="scsi" target=1 lun=0;
name="sg" class="scsi" target=1 lun=1;
 Solaris 64
Installing/reinstalling the sg and the st drivers

name="sg" class="scsi" target=1 lun=2;

...
<entries omitted for brevity>
...
name="sg" class="scsi" target=15 lun=5;
name="sg" class="scsi" target=15 lun=6;
name="sg" class="scsi" target=15 lun=7;
name="sg" parent="fp" target=0 lun=0 fc-port-wwn="500104f0008d53c3";
name="sg" parent="fp" target=0 lun=1 fc-port-wwn="500104f0008d53c3";
name="sg" parent="fp" target=0 lun=0 fc-port-wwn="500104f0008d53c6";
name="sg" parent="fp" target=0 lun=1 fc-port-wwn="500104f0008d53c6";
name="sg" parent="fp" target=0 lun=0 fc-port-wwn="500104f0008d53c9";
name="sg" parent="fp" target=0 lun=1 fc-port-wwn="500104f0008d53c9";
name="sg" parent="fp" target=0 lun=0 fc-port-wwn="500104f0008d53cc";
name="sg" parent="fp" target=0 lun=1 fc-port-wwn="500104f0008d53cc";
name="sg" parent="fp" target=0 lun=0 fc-port-wwn="500104f0008d53b9";
name="sg" parent="fp" target=0 lun=1 fc-port-wwn="500104f0008d53b9";
name="sg" parent="fp" target=0 lun=0 fc-port-wwn="500104f0008d53c3";
name="sg" parent="fp" target=0 lun=1 fc-port-wwn="500104f0008d53c3";
name="sg" parent="fp" target=0 lun=0 fc-port-wwn="500104f0008d53c6";
name="sg" parent="fp" target=0 lun=1 fc-port-wwn="500104f0008d53c6";
name="sg" parent="fp" target=0 lun=0 fc-port-wwn="500104f0008d53c9";
name="sg" parent="fp" target=0 lun=1 fc-port-wwn="500104f0008d53c9";
name="sg" parent="fp" target=0 lun=0 fc-port-wwn="500104f0008d53cc";
name="sg" parent="fp" target=0 lun=1 fc-port-wwn="500104f0008d53cc";
name="sg" parent="fp" target=0 lun=0 fc-port-wwn="500104f0008d53b9";
name="sg" parent="fp" target=0 lun=1 fc-port-wwn="500104f0008d53b

sg.links file example

The following /usr/openv/volmgr/bin/driver/sg.links file example shows
targets 0-15 and LUNs 0-7. It also includes entries for three StorEdge Network
Foundation HBA ports.
The sg.build -mt option does not affect FCP targets, but the -ml option does. The
Solaris luxadm command detected three ports (identified by their World Wide
Names). Therefore, the sg.build script created entries for LUNs 0 through 7 for
those three ports.
The field separator between the addr=x, y; field and the sg/ field is a tab. The
addr= field uses hexadecimal notation, and the sg/ field uses decimal values.

# begin SCSA Generic devlinks file - creates nodes in /dev/sg

type=ddi_pseudo;name=sg;addr=0,0; sg/c\N0t0l0
 Solaris 65
Configuring 6 GB and larger SAS HBAs in Solaris

type=ddi_pseudo;name=sg;addr=0,1; sg/c\N0t0l1
type=ddi_pseudo;name=sg;addr=0,2; sg/c\N0t0l2
type=ddi_pseudo;name=sg;addr=0,3; sg/c\N0t0l3
type=ddi_pseudo;name=sg;addr=0,4; sg/c\N0t0l4
type=ddi_pseudo;name=sg;addr=0,5; sg/c\N0t0l5
type=ddi_pseudo;name=sg;addr=0,6; sg/c\N0t0l6
type=ddi_pseudo;name=sg;addr=0,7; sg/c\N0t0l7
type=ddi_pseudo;name=sg;addr=1,0; sg/c\N0t1l0
type=ddi_pseudo;name=sg;addr=1,1; sg/c\N0t1l1
...
<entries omitted for brevity>
...
type=ddi_pseudo;name=sg;addr=f,5; sg/c\N0t15l5
type=ddi_pseudo;name=sg;addr=f,6; sg/c\N0t15l6
type=ddi_pseudo;name=sg;addr=f,7; sg/c\N0t15l7
type=ddi_pseudo;name=sg;addr=w500104f0008d53c3,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f0008d53c3,1; sg/c\N0t\A1l1
type=ddi_pseudo;name=sg;addr=w500104f0008d53c6,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f0008d53c6,1; sg/c\N0t\A1l1
type=ddi_pseudo;name=sg;addr=w500104f0008d53c9,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f0008d53c9,1; sg/c\N0t\A1l1
type=ddi_pseudo;name=sg;addr=w500104f0008d53cc,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f0008d53cc,1; sg/c\N0t\A1l1
type=ddi_pseudo;name=sg;addr=w500104f0008d53b9,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f0008d53b9,1; sg/c\N0t\A1l1
type=ddi_pseudo;name=sg;addr=w500104f0008d53c3,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f0008d53c3,1; sg/c\N0t\A1l1
type=ddi_pseudo;name=sg;addr=w500104f0008d53c6,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f0008d53c6,1; sg/c\N0t\A1l1
type=ddi_pseudo;name=sg;addr=w500104f0008d53c9,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f0008d53c9,1; sg/c\N0t\A1l1
type=ddi_pseudo;name=sg;addr=w500104f0008d53cc,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f0008d53cc,1; sg/c\N0t\A1l1
type=ddi_pseudo;name=sg;addr=w500104f0008d53b9,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f0008d53b9,1; sg/c\N0t\A1l1
# end SCSA devlinks

Configuring 6 GB and larger SAS HBAs in Solaris

Use the procedure in this topic to configure the NetBackup sg driver for Oracle 6
GB and larger SAS HBAs on Solaris.
A separate topic describes how to install the NetBackup sg and Sun st drivers.
 Solaris 66
Configuring 6 GB and larger SAS HBAs in Solaris

See “Installing/reinstalling the sg and the st drivers” on page 60.

Note: Support for Solaris 6 GB serial-attached SCSI (SAS) HBAs for tape devices
requires a specific Solaris patch level. Ensure that you install the required patches.
For supported Solaris versions, see the Oracle Support website.

To configure 6 GB and larger SAS HBAs in Solaris

1 Verify that the 6 GB SAS tape device path exists by running the following
command in a shell window:
ls -l /dev/rmt | grep cbn

6 GB SAS tape devices should have iport@ in the name path. The following
is an example of the output (the tape drive address is highlighted):

1cbn -> ../../devices/pci@400/pci@0/pci@9/LSI,sas@0/iport@8/tape@w500104f000ba856a,0:cbn

2 Edit the /etc/devlink.tab file

Include the following lines for every 6 GB SAS tape drive in the
/etc/devlink.tab file. Replace drive_address with the tape drive address;
see the output from step 1 for the tape drive addresses.

type=ddi_pseudo;name=sg;addr=wdrive_address,0,1; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=wdrive_address,1,1; sg/c\N0t\A1l1

Include the following lines for every 6 GB SAS robotic library in the
/etc/devlink.tab file. Replace drive_address with the tape drive address;
see the output from step 1 for the tape drive address.

type=ddi_pseudo;name=medium-changer;addr=wdrive_address,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=medium-changer;addr=wdrive_address,1; sg/c\N0t\A1l1

The following are example entries for the devlink.tab file:

# SCSA devlinks for SAS-2 drives:

type=ddi_pseudo;name=sg;addr=w500104f000ba856a,0,1; sg/c\N0t\A1l0
type=ddi_pseudo;name=sg;addr=w500104f000ba856a,1,1; sg/c\N0t\A1l1
# SCSA devlinks for SAS-2 libraries:
type=ddi_pseudo;name=medium-changer;addr=w500104f000ba856a,0; sg/c\N0t\A1l0
type=ddi_pseudo;name=medium-changer;addr=w500104f000ba856a,1; sg/c\N0t\A1l1
 Solaris 67
Preventing Solaris driver unloading

3 Verify that the sg driver SCSI classes are 08 and 0101 by running the following
command:
grep sg /etc/driver_aliases

The following is an example of the output:

sg "scsiclass,0101"
sg "scsiclass,08"

4 If the sg driver SCSI classes are not 08 and 0101, reinstall the sg driver by
using the following commands:

rem_drv sg
update_drv -d -i "scsiclass,08" sgen
add_drv -m '* 0600 root root' -i '"scsiclass,0101" "scsiclass,08"' sg

5 Restart the host.

6 Verify that the sg drivers exist by running the following command:
ls -l /dev/sg

The following is an example of the output (the output was modified to fit on the
page):

c0tw500104f000ba856al0 ->
../../devices/pci@400/pci@0/pci@9/LSI,sas@0/iport@8/sg@w500104f000ba856a,0,1:raw
c0tw500104f000ba856al1 ->
../../devices/pci@400/pci@0/pci@9/LSI,sas@0/iport@8/medium-changer@w500104f000ba856a,1:raw

7 Verify that the NetBackup sgscan utility recognizes the tape devices by entering
the following command:
/usr/openv/volmgr/bin/sgscan

The following is an example of the output:

/dev/sg/c0tw500104f000ba856al0: Tape (/dev/rmt/1): "HP Ultrium 5-SCSI"

/dev/sg/c0tw500104f000ba856al1: Changer: "STK SL500"

Preventing Solaris driver unloading

When system memory is limited, Solaris unloads unused drivers from memory and
reloads drivers as needed. Tape drivers are often unloaded because they are used
less often than disk drivers.
 Solaris 68
About Solaris robotic controls

The drivers NetBackup uses are the st driver (from Sun), the sg driver (from Veritas),
and Fibre Channel drivers. Problems may occur depending on when the driver
loads and unloads. These problems can range from a SCSI bus not able to detect
a device to system panics.
Veritas recommends that you prevent Solaris from unloading the drivers from
memory.
The following procedures describe how to prevent Solaris from unloading the drivers
from memory.
To prevent Solaris from unloading the drivers from memory
◆ Add the following forceload statements to the /etc/system file:

forceload: drv/st
forceload: drv/sg

To prevent Solaris from unloading the Fibre Channel drivers from memory
◆ Add an appropriate forceload statement to the /etc/system file.
Which driver you force to load depends on your Fibre Channel adapter. The
following is an example for a Sun Fibre Channel driver (SunFC FCP
v20100509-1.143):
forceload: drv/fcp

About Solaris robotic controls

NetBackup supports SCSI control and API control of robotic devices. A robotic
device in a library moves the media between the storage slots and the drives in the
library.
Robotic control varies, as follows:
■ SCSI or Fibre Channel Protocol control.
See “About SCSI and FCP robotic controls on Solaris” on page 68.
■ API control over a LAN.
See the "ADIC Automated Media Library (AML)" topic in this guide.
See the "IBM Automated Tape Library (ATL)" topic in this guide.
See the "Sun StorageTek ACSLS robots" topic in this guide.

About SCSI and FCP robotic controls on Solaris

When you configure the NetBackup sg driver, a NetBackup script creates the device
files for the attached robotic devices.
 Solaris 69
About Solaris tape drive device files

See “About the NetBackup sg driver” on page 57.

If you use device discovery in NetBackup, NetBackup discovers the robotic control
device files in the /dev/sg directory (and hence the devices) automatically. If you
add a robot manually in NetBackup, you must enter the pathname to the device
file.
To display the device files that the sg driver can use, use the NetBackup sgscan
command with the all parameter. The word "Changer" in the sgscan output
identifies robotic control device files.
Examples are available.
See “Examples of SCSI and FCP robotic control device files on Solaris” on page 69.

Examples of SCSI and FCP robotic control device files on Solaris

The following is an example of sgscan all output from a host, to which the
examples refer:

# /usr/openv/volmgr/bin/sgscan all
/dev/sg/c0t6l0: Cdrom: "TOSHIBA XM-5401TASUN4XCD"
/dev/sg/c1tw500104f0008d53b9l0: Changer: "STK SL500"
/dev/sg/c1tw500104f0008d53c3l0: Tape (/dev/rmt/0): "HP Ultrium 3-SCSI"
/dev/sg/c1tw500104f0008d53c6l0: Tape (/dev/rmt/1): "HP Ultrium 3-SCSI"
/dev/sg/c1tw500104f0008d53c9l0: Tape (/dev/rmt/2): "IBM ULTRIUM-TD3"
/dev/sg/c1tw500104f0008d53ccl0: Tape (/dev/rmt/3): "IBM ULTRIUM-TD3"
/dev/sg/c2t1l0: Changer: "STK SL500"
/dev/sg/c2t2l0: Tape (/dev/rmt/22): "HP Ultrium 3-SCSI"
/dev/sg/c2t3l0: Tape (/dev/rmt/10): "HP Ultrium 3-SCSI"
/dev/sg/c2tal0: Tape (/dev/rmt/18): "IBM ULTRIUM-TD3"
/dev/sg/c2tbl0: Tape (/dev/rmt/19): "IBM ULTRIUM-TD3"
/dev/sg/c3t0l0: Disk (/dev/rdsk/c1t0d0): "FUJITSU MAV2073RCSUN72G"
/dev/sg/c3t3l0: Disk (/dev/rdsk/c1t3d0): "FUJITSU MAV2073RCSUN72G"

You can filter the sgscan output for device types by using other sgscan options.
The following is the sgscan usage statement:
sgscan [all|basic|changer|disk|tape] [conf] [-v]

About Solaris tape drive device files

NetBackup uses the tape drive device files that support compression, no rewind
on close, and Berkeley style close.
 Solaris 70
About Solaris tape drive device files

When you configure the Solaris st driver, Solaris creates the device files for the
attached tape devices
See “Installing/reinstalling the sg and the st drivers” on page 60.
The device files are in the /dev/rmt directory, and they have the following format:
/dev/rmt/IDcbn

The following describe the device file names:

■ ID is the logical drive number as shown by the NetBackup sgscan command.

■ c indicates compression.

■ b indicates Berkeley-style close.

■ n indicates no rewind on close.

If you use device discovery in NetBackup, NetBackup discovers the device files
and hence the devices. If you add a tape drive to a NetBackup configuration
manually, you must specify the pathname to the device file. NetBackup requires
compression, no rewind on close, and Berkeley-style close device files.
To display the tape device files that are configured on your system, use the sgscan
command with the tape parameter, as follows:

# /usr/openv/volmgr/bin/sgscan tape
/dev/sg/c1tw500104f0008d53c3l0: Tape (/dev/rmt/0): "HP Ultrium 3-SCSI"
/dev/sg/c1tw500104f0008d53c6l0: Tape (/dev/rmt/1): "HP Ultrium 3-SCSI"
/dev/sg/c1tw500104f0008d53c9l0: Tape (/dev/rmt/2): "IBM ULTRIUM-TD3"
/dev/sg/c1tw500104f0008d53ccl0: Tape (/dev/rmt/3): "IBM ULTRIUM-TD3"
/dev/sg/c2t2l0: Tape (/dev/rmt/22): "HP Ultrium 3-SCSI"
/dev/sg/c2t3l0: Tape (/dev/rmt/10): "HP Ultrium 3-SCSI"
/dev/sg/c2tal0: Tape (/dev/rmt/18): "IBM ULTRIUM-TD3"
/dev/sg/c2tbl0: Tape (/dev/rmt/19): "IBM ULTRIUM-TD3"

The following are examples of no-rewind, compression, Berkeley-style close device

files from the preceding sgscan example output:
■ For the Ultrium3 SCSI drive at LUN 0 of World Wide Node Name (WWNN)
500104f0008d53c3, the device file pathname is:
/dev/rmt/0cbn

■ For the HP Ultrium3 SCSI drive at SCSI ID 2 of adapter 2, the device file
pathname is:
/dev/rmt/22cbn
 Solaris 71
About Solaris tape drive device files

You can show all device types by using the all option. The output can help you
associate tape devices with other SCSI devices that may be configured on the same
adapter. The following is the sgscan usage statement:
sgscan [all|basic|changer|disk|tape] [conf] [-v]

About Berkeley-style close

NetBackup requires Berkeley-style close for tape drive device files. The letter b in
the file name indicates Berkeley-style close device files.
In Berkeley-style close, the tape position remains unchanged by a device close
operation. (Conversely, in AT&T-style close, the drive advances the tape to just
after the next end-of-file (EOF) marker.) To establish the correct position for the
next tape operation, applications must assume the tape's position after a close.
NetBackup assumes Berkeley-style close on Solaris systems.

About no rewind device files on Solaris

NetBackup requires no rewind on close device files for tape drives.
With no rewind on close, the tape is not rewound after a close operation. It remains
in position for the next write operation.
The letter n in the device file names in the /dev/rmt directory specifies no rewind
on close.

About fast-tape positioning (locate-block) on Solaris

Applies to AIT, DLT, Exabyte, DTF, and half-inch tape drives.
To position a tape to a specific block, NetBackup supports the SCSI locate-block
command. It requires the NetBackup sg driver.
NetBackup uses the locate-block command by default.
Veritas recommends that you do not disable locate-block positioning. If you need
to disable it, execute the following command:
touch /usr/openv/volmgr/database/NO_LOCATEBLOCK

If locate-block positioning is disabled, NetBackup uses the forward-space-file/record

method.
 Solaris 72
About Solaris tape drive device files

About SPC-2 SCSI reserve on Solaris

By default, NetBackup uses SPC-2 SCSI reserve and release for tape drive
reservations in shared drive environments. The NetBackup Shared Storage Option
provides shared drive functionality in NetBackup.
Alternatively, you can use SCSI persistent reserve for shared tape drive reservations
in NetBackup, as follows:
■ For the tape drives that support SPC-3 Compatible Reservation Handling (CRH),
you can use SCSI persistent reserve by enabling it in NetBackup. No special
configuration in Solaris is required.
■ For the tape drives that do not support CRH, you must disable SPC-2 SCSI
reserve in Solaris for those drives. After you disable SPC-2 SCSI reserve, you
can use persistent reserve by enabling it in NetBackup. If the drive does not
support CRH and you do not disable SPC-2 SCSI reserve, access attempts to
the drive fail.
See “Disabling SPC-2 SCSI reserve on Solaris” on page 72.
For more information about NetBackup and SCSI reservations, see the following:
■ The description of the Enable SCSI Reserve Media host property in the
NetBackup Administrator’s Guide, Volume I.
■ The "How NetBackup reserves drives" topic in the NetBackup Administrator’s
Guide, Volume II.

Disabling SPC-2 SCSI reserve on Solaris

Use the following procedure to disable SPC-2 SCSI reserve.
More information about reservations is available.
See “About SPC-2 SCSI reserve on Solaris” on page 72.
To disable SPC-2 SCSI reserve
◆ Modify the Solaris st.conf file on the NetBackup media server. In the
tape-config-list section of the st.conf file, set the ST_NO_RESERVE_RELEASE
configuration value (0x20000) in the appropriate data-property-name entry.
For example, the following entry disables SCSI reserve and release for all tape
drives that use the DLT7k-data configuration values:

DLT7k-data = 1,0x38,0,0x20000,4,0x82,0x83,0x84,0x85,2;

For more information about the st.conf file, see the Solaris st(7D) man page.
 Solaris 73
Configuring Solaris SAN clients to recognize FT media servers

About nonstandard tape drives

Solaris includes the device drivers that support most standard devices.
To receive the most current support for devices, you should install the latest Solaris
patch for the st driver.
However, if you have a device that Solaris does not support, the device manufacturer
should provide the software to install and administer the device properly. In addition,
the device vendor should contact Sun Microsystems to add support for the device
to Solaris.
For more information about what you need for unsupported devices, contact the
device vendor. Also see the Solaris devices and file systems documentation.

Configuring Solaris SAN clients to recognize FT

media servers
NetBackup SAN clients use tape drivers and SCSI pass-through methods for Fibre
Transport traffic to NetBackup FT media servers. The media server FT devices
appear as ARCHIVE Python tape devices during SCSI inquiry on the SAN client.
However, they are not tape devices and do not appear as tape devices in NetBackup
device discovery.
Veritas owns the ARCHIVE brand name and Python product name. Therefore,
st.conf file changes to ARCHIVE Python do not affect an existing tape drive product.

Table 5-1 is an overview of procedures to configure the Solaris operating system

so that it recognizes the NetBackup FT devices on the NetBackup media servers.

Table 5-1 Configuring SAN clients to recognize FT media servers

Step Task Procedure

1 Add the Fibre Transport device entry to See “Adding the FT device entry to the
the st.conf file st.conf file” on page 73.

2 Modify the st.conf file so that Solaris See “Modifying the st.conf file so that
discovers devices on two LUNS Solaris discovers devices on two LUNS”
on page 74.

Adding the FT device entry to the st.conf file

The following procedure describes how to add the FT device entry to the st.conf
file.
 Solaris 74
Configuring Solaris SAN clients to recognize FT media servers

To add the FT device entry to the st.conf file

1 In the /kernel/drv/st.conf file, find the tape-config-list= section or create
it if it does not exist.
2 Examine the tape-config-list= section for a line that begins with ARCHIVE
Python and contains ARCH_04106. If such a line exists, ensure that it begins
with a comment character (#).
3 Add the following line to the tape-config-list= section:
"ARCHIVE Python", "FT Pipe", "ARCH_04106";

4 Find the line that begins with ARCH_04106, copy it, and paste it after the
tape-config-list= line. Delete the comment character (#) from the beginning
of the line. The following is an example of the line:
ARCH_04106 = 1, 0x2C, 0, 0x09639, 4, 0x00, 0x8C, 0x8c, 0x8C, 3;

Modifying the st.conf file so that Solaris discovers devices on two

LUNS
The following procedure describes how to modify the st.conf file so that Solaris
discovers devices on two LUNS.
 Solaris 75
Uninstalling the sg driver on Solaris

To modify the st.conf file so that Solaris discovers devices on two LUNS
1 Find the following line in the st.conf file:
name="st" class="scsi" target=0 lun=0;

2 Replace that line and the following lines through target 5 with the following.
Doing so modifies the st.conf file to include searches on non-zero LUNs.

name="st" class="scsi" target=0 lun=0;

name="st" class="scsi" target=0 lun=1;
name="st" class="scsi" target=1 lun=0;
name="st" class="scsi" target=1 lun=1;
name="st" class="scsi" target=2 lun=0;
name="st" class="scsi" target=2 lun=1;
name="st" class="scsi" target=3 lun=0;
name="st" class="scsi" target=3 lun=1;
name="st" class="scsi" target=4 lun=0;
name="st" class="scsi" target=4 lun=1;
name="st" class="scsi" target=5 lun=0;
name="st" class="scsi" target=5 lun=1;
name="st" parent="fp" target=0;
name="st" parent="fp" target=1;
name="st" parent="fp" target=2;
name="st" parent="fp" target=3;
name="st" parent="fp" target=4;
name="st" parent="fp" target=5;
name="st" parent="fp" target=6;

Uninstalling the sg driver on Solaris

You can uninstall the sg driver. If you do, NetBackup performance suffers. The
following procedure describes how to uninstall the sg driver.
To uninstall the sg driver
◆ Invoke the following command:
/usr/sbin/rem_drv sg

Solaris command summary

The following is a summary of commands that may be useful when you configure
and verify devices:
 Solaris 76
Solaris command summary

■ /usr/sbin/modinfo | grep sg
Displays whether or not the sg driver is installed.
■ /usr/openv/volmgr/bin/driver/sg.install
Installs the sg driver or updates the sg driver.
■ /usr/sbin/rem_drv sg
Uninstalls the sg driver. This command usually is not necessary because
sg.install uninstalls the old driver before it upgrades a driver.

■ /usr/openv/volmgr/bin/sg.build all -mt max_target -ml max_lun

Updates st.conf, sg.conf, and sg.links, and generates SCSI Target IDs with
multiple LUNs.
■ /usr/openv/volmgr/bin/sgscan all
Scans all connected devices with an SCSI inquiry and provides correlation
between physical and the logical devices that use all device files in /dev/sg.
Also checks for the devices that are connected to the Sun StorEdge Network
Foundation HBA that are not configured for use by Veritas products.
■ boot -r or reboot -- -r
Reboot the system with the reconfigure option (-r). The kernel’s SCSI disk (sd)
driver then recognizes the drive as a disk drive during system initialization.
See the procedures in this chapter for examples of their usage.
 Chapter 6
Windows
This chapter includes the following topics:

■ Before you begin configuring NetBackup on Windows

■ About tape device drivers on Windows

■ Attaching devices to a Windows system

Before you begin configuring NetBackup on

Windows
Observe the following points when performing the configurations described in this
chapter:
■ Verify that NetBackup supports your server platform and devices. Download
the NetBackup hardware and operating system compatibility lists.
See “About the NetBackup compatibility lists” on page 10.
■ For NetBackup to recognize and communicate with connected devices and for
device discovery to discover devices, NetBackup issues SCSI pass-through
commands to the devices in a configuration.
A tape driver must exist for each tape device. Attached devices appear in the
registry.
■ Use the Microsoft Windows device applications to verify that the devices are
configured correctly. The device applications available on your server may differ
depending on your Windows operating system. Make sure that Windows detects
the devices on the SAN before you configure the NetBackup Shared Storage
Option.
■ If you have multiple devices connected to a fibre bridge, Windows may only see
one LUN. This will normally be the device with the lowest-ordered LUN.
 Windows 78
About tape device drivers on Windows

This limitation occurs because of the default install settings for the device driver
for some fibre channel HBAs. See your vendor documentation to verify the
settings.
■ Information about how to configure API robot control over a LAN is available
See the "ADIC Automated Media Library (AML)" topic in this guide.
See the "IBM Automated Tape Library (ATL)" topic in this guide.
See the "Sun StorageTek ACSLS robots" topic in this guide.
After configuring the hardware, add the drives and robots to NetBackup.

About tape device drivers on Windows

Veritas does not provide device drivers for Windows hosts. If you require drivers,
contact Microsoft or the tape drive vendor.

Attaching devices to a Windows system

The following procedure describes a general method for attaching devices to a
Windows computer. The Microsoft Windows device applications available on the
server that you use in these steps may differ depending on your Windows operating
system.
To attach devices to a Windows system
1 Use the appropriate Windows application to obtain information on any currently
attached SCSI devices.
2 If you attach a new robotic library or drive to a NetBackup media server, follow
the vendor’s instructions for attaching the device.
Shut down the server and physically attach the supported device. Ensure that
SCSI targets and termination settings are consistent with adapter card and
peripheral vendor recommendations.
3 Reboot the server and answer the prompts for adapter card peripheral
configuration options. Watch the display to ensure that the adapter card
recognizes the attached peripherals.
4 If you add drives, install the tape drivers and use the appropriate Windows
application to verify that the drive was recognized.
 Section 2
Robotic storage devices

■ Chapter 7. Robot overview

■ Chapter 8. Oracle StorageTek ACSLS robots

■ Chapter 9. Device configuration examples

 Chapter 7
Robot overview
This chapter includes the following topics:

■ NetBackup robot types

■ Robot attributes

■ Table-driven robotics

■ Robotic test utilities

■ Robotic processes

NetBackup robot types

A robot is a peripheral device that moves tape volumes into and out of tape drives.
NetBackup uses robotic control software to communicate with the robot firmware.
NetBackup classifies robots according to one or more of the following characteristics:
■ The communication method the robotic control software uses; SCSI and API
are the two main methods.
■ The physical characteristics of the robot. Library refers to a large robot, in terms
of slot capacity or number of drives.
■ The media type commonly used by that class of robots. HCART (1/2-inch
cartridge tape) is an example of a media type.
Table 7-1 lists the NetBackup robot types that are supported in release 8.0 Draft,
with drive and slot limits for each type.
To determine which robot type applies to the model of robot that you use, see the
hardware compatibility list for your NetBackup version:
http://www.netbackup.com/compatibility
 Robot overview 81
Robot attributes

Table 7-1 NetBackup robot types in release 8.0 Draft

Robot type Description Drive limits Slot limits Note

ACS Automated Cartridge System 1680 No limit API control. The ACS library
software host determines the
drive limit.

TLD Tape library DLT No limit 32000 SCSI control.

Note: The user interface for NetBackup may show configuration options for the
peripheral devices that are not supported in that release. Those devices may be
supported in an earlier release, and a NetBackup master server can manage the
hosts that run earlier NetBackup versions. Therefore, the configuration information
for such devices must appear in the user interface. The NetBackup documentation
also may describe the configuration information for such devices. To determine
which versions of NetBackup support which peripheral devices, see the hardware
compatibility list:
http://www.netbackup.com/compatibility

For configuration information about the robots that NetBackup supports in releases
earlier than 8.0 Draft, see the NetBackup Device Configuration Guide for that
release, available through the following URL:
http://www.veritas.com/docs/DOC5332

Robot attributes
NetBackup configures and controls robots differently depending on the robot type.
The following tables list the attributes that dictate how these robot types differ.
For more detailed information about supported devices, firmware levels, and
platforms, see the hardware compatibility list for your NetBackup version:
http://www.netbackup.com/compatibility

ACS robots
Unlike other robot types, NetBackup does not track slot locations for the media in
ACS robots. The ACS library software tracks slot locations and reports them to
NetBackup.
The following table describes the ACS robot attributes.
 Robot overview 82
Robot attributes

Table 7-2 ACS robot attributes

Attribute NetBackup server

API robot Yes

SCSI control No

LAN control Yes

Remote Robot control No. Each host that has ACS drives that are attached to it has
robotic control.

NDMP support Yes

Shared drives support Yes

Drive cleaning support No. The ACS library software manges drive cleaning.

Media access port support Yes, for eject only.

NetBackup tracks slots No

Media type support DLT, DLT2, DLT3, HCART, HCART2, and HCART3.

Hosts Supported Windows, UNIX, and Linux.

Windows servers require STK LibAttach software. See the

Veritas support web site for the latest compatibility information
and obtain the appropriate LibAttach software from STK.

Barcode Support Yes. Depends on ACS library software to obtain NetBackup

media IDs.

Barcodes must be the same as the media ID (1 to 6

characters).

Robot Examples Oracle SL500, Oracle SL3000, and Oracle SL8500

TLD robots
The following table describes the tape library DLT attributes.

Table 7-3 TLD robot attributes

Attribute NetBackup Server NetBackup server

API robot No No

SCSI control Yes Yes

 Robot overview 83
Table-driven robotics

Table 7-3 TLD robot attributes (continued)

Attribute NetBackup Server NetBackup server

LAN control Not Applicable No

Remote robot control Not Applicable Yes

NDMP support Yes Yes

Shared drives support Not Applicable Yes

Drive cleaning support Yes Yes

Media access port Yes Yes

support

NetBackup tracks slots Yes Yes

Hosts supported Windows, UNIX, and Linux. Windows, UNIX, and Linux.

Media type support DLT, DLT2, DLT3, DTF, 8MM, DLT, DLT2, DLT3, DTF, 8MM,
8MM2, 8MM3, QIC, HCART, 8MM2, 8MM3, QIC, HCART,
HCART2, HCART3 HCART2, HCART3

Barcode support Yes. Barcodes can be from 1 to Yes. Barcodes can be from 1 to
16 characters in length. The 16 characters in length. The
Media Manager media ID is six Media Manager media ID is six
or fewer characters. or fewer characters.

Robot examples HP MSL, Fujitsu FibreCAT HP MSL, Fujitsu FibreCAT

TX48, IBM TotalStorage3583, TX48, IBM TotalStorage3583,
Spectra Logic T680, Sun/Oracle Spectra Logic T680, Sun/Oracle
SL3000 SL3000

Table-driven robotics
Table-driven robotics provides support for new robotic library devices without the
need to modify any library control binary files. This feature uses a device mapping
file for supported robots and drives.
You may be able to add support for new or upgraded devices without waiting for a
maintenance patch from Veritas. The device mapping file includes the information
that relates to the operation and control of libraries. Therefore, you can download
an updated mapping file to obtain support for newly NetBackup-certified devices.
For the device mappings file downloads, see the following URL:
http://www.netbackup.com/compatibility
 Robot overview 84
Robotic test utilities

Robotic test utilities

You can use robotic test utilities for testing robots already configured in NetBackup.
Invoke the test utilities as follows:
■ /usr/openv/volmgr/bin/robtest (UNIX and Linux)

■ install_path\Veritas\Volmgr\bin\robtest.exe (Windows)

From each test utility, you can obtain a list of available test commands by entering
a question mark (?).
Use the drstat command to determine the drive addressing parameters for the
ACS robot type. This command is available in the robotic test utilities for these robot
types.
NetBackup addresses drives as follows:
■ For ACS robot types, by ACS, LSM, Panel, and Drive number
■ For other robot types, by the robot drive number

Robotic processes
A NetBackup robotic process and possibly a robotic control process exist on a
NetBackup media server for each robot that you install, as follows:
■ Every media server that has a drive in a robotic library has a robotic process
for that robotic library. The robotic process receives requests from the NetBackup
Device Manager (ltid) and sends necessary information directly to the robotics
or to a robotic control process.
■ Robotic control processes exist only for the robot types that support library
sharing (or robot sharing).
When the NetBackup Device Manager starts, it starts the robotic processes and
the robotic control processes for all of the configured robots on that host. When the
Device Manager stops, the robotic processes and the robotic control processes
stop. (On UNIX, the name is Media Manager Device daemon.)
You can start and stop the Device Manager manually by using the NetBackup
Administration Console as follows:
■ In the NetBackup Activity Monitor Daemons tab, select it and then select
Actions Start Daemon or Stop Daemon.
■ Select Device Monitor, Media, or Devices in the right pane and then select
Actions Stop/Restart Media Manager Device Daemon.
 Robot overview 85
Robotic processes

In addition, the NetBackup Commands Reference Guide describes commands to

control the robotic processes that run on Windows media servers.
You can determine if a robotic process or robotic control process is active by using
the NetBackup Activity Monitor Processes tab.
You can determine the control state of NetBackup by using the Device Monitor
Drive paths pane or Drive status pane. If the value in the Control column for a
drive shows the control mode, the robotic process is running and the drive is usable.
For example, for a TLD robot the control mode is TLD.
Other values such as AVR or DOWN may indicate that the drive is unusable. For
the possible values and their explanations, see the online Help for the Device
Monitor.
See “Processes by robot type” on page 85.
See “Robotic process example” on page 86.

Processes by robot type

The following table describes the robotic processes and robotic control processes
for each robot type.

Table 7-4 Robotic processes and robotic control processes

Robot type Process Description

Automated Cartridge acsd The NetBackup ACS daemon acsd provides robotic control to mount and
System (ACS) dismount volumes. It also requests inventories of the volumes that are under
the control of ACS library software.

acssel The NetBackup ACS storage server interface (SSI) event logger acssel
logs events. UNIX and Linux only.

acsssi The NetBackup ACS storage server interface (SSI) acsssi communicates
with the ACS library software host. acsssi processes all RPC
communications from acsd or from the ACS robotic test utility that are
intended for the ACS library software. UNIX and Linux only.

Tape library DLT (TLD) tldd The tape library DLT daemon tldd runs on a NetBackup server that has a
drive in the tape library DLT. This process receives NetBackup Device
Manager requests to mount and unmount volumes, and sends these requests
to the robotic-control process, tldcd.
 Robot overview 86
Robotic processes

Table 7-4 Robotic processes and robotic control processes (continued)

Robot type Process Description

tldcd The tape library DLT Control daemon tldcd communicates with the tape
library DLT robotics through a SCSI interface.

For library sharing, tldcd runs on the NetBackup server that has the robotic
control.

Tape library 4MM (TL4) tl4d The tape library 4MM daemon tl4d runs on the host that has a tape library
4MM. This process receives NetBackup Device Manager requests to mount
and unmount volumes and communicates these requests to the robotics
through a SCSI interface.

Tape library 8MM (TL8) tl8d The tape library 8MM daemon tl8d runs on a NetBackup server that has a
drive in the tape library 8MM. This process receives NetBackup Device
Manager requests to mount and unmount volumes, and sends these requests
to the robotic-control process, tl8cd.

tl8cd The tape library 8MM Control daemon tl8cd communicates with the TL8
robotics through a SCSI interface.

For library sharing, tl8cd runs on the NetBackup server that has the robotic
control.

Tape library Half-inch tlhd The tape library Half-inch daemon tlhd runs on each NetBackup server
(TLH) that has a drive in the tape library Half-inch. This process receives NetBackup
Device Manager requests to mount and unmount volumes and sends these
requests to the robotic-control process.

tlhcd The tape library half-inch control daemon tlhcd runs on the NetBackup
server that has the robotic control. It communicates with the TLH robotics
through a SCSI interface

Tape library Multimedia tlmd tlmd runs on the NetBackup server and communicates mount and unmount
(TLM) requests to the host that controls the TLM robotics.

See “NetBackup robot types” on page 80.

Robotic process example

Each drive in a tape library DLT (TLD) robot can be attached to a different host,
and a tldd process runs on each host. However, only one host controls the robotics,
and the tldcd robotic control process runs on that host only. To mount a tape, the
tldd process on the host to which the drive is attached sends control information
to the tldcd process on the robotic control host.
The following figure shows the processes and where they run for a TLD robot.
 Robot overview 87
Robotic processes

Figure 7-1 TLD robot control process example

Host A
Host B
Robotic control host

Device Manager Device Manager

tldd TLD robot tldd

Robotics
tldcd
Drive 1

SCSI SCSI
Drive 2

The following describes this example:

■ Each host connects to one drive, and a tldd robotic process runs on each host.
■ The robotic control and therefore the robotic control process, tldcd, is on host
A.
The NetBackup Device Manager services on host A and B start tldd. The tldd
process on host A also starts tldcd. Requests to mount tapes from host B go to
tldd on host B, which then sends the robotic command to tldcd on host A.
 Chapter 8
Oracle StorageTek ACSLS
robots
This chapter includes the following topics:

■ About Oracle StorageTek ACSLS robots

■ Sample ACSLS configurations

■ Media requests for an ACS robot

■ About configuring ACS drives

■ Configuring shared ACS drives

■ Adding tapes to ACS robots

■ About removing tapes from ACS robots

■ Robot inventory operations on ACS robots

■ NetBackup robotic control, communication, and logging

■ ACS robotic test utility

■ Changing your ACS robotic configuration

■ ACS configurations supported

■ Oracle StorageTek ACSLS firewall configuration

 Oracle StorageTek ACSLS robots 89
About Oracle StorageTek ACSLS robots

About Oracle StorageTek ACSLS robots

Note: If you use the access control feature of Oracle StorageTek ACSLS controlled
robots and the NetBackup media sharing feature, do the following: ensure that all
servers in the NetBackup media server share group have the same ACSLS
permissions to all the same ACSLS media and ACSLS drives. Any mismatches
can cause failed jobs and stranded tapes in drives.

Oracle StorageTek Automated Cartridge System Library Software controlled robots

are NetBackup robot type ACS.
ACS robots are API robots (a NetBackup robot category in which the robot manages
its own media).
Unlike other robot types, NetBackup does not track slot locations for the media in
ACS robots. The Automated Cartridge System Library Software tracks slot locations
and reports them to NetBackup.
The term automated cartridge system (ACS) can refer to any of the following:
■ A type of NetBackup robotic control.
■ The Oracle StorageTek system for robotic control.
■ The highest-level component of the Oracle StorageTek ACSLS. It refers to one
robotic library or to multiple libraries that are connected with a media
pass-through mechanism.
The ACS library software component can be either of the following Oracle
StorageTek products:
■ Oracle StorageTek Automated Cartridge System Library Software (ACSLS)
■ Oracle StorageTek Library Station

Sample ACSLS configurations

The sample ACSLS configurations show the following:
■ A typical UNIX ACSLS configuration.
See Figure 8-1 on page 90.
■ A typical Windows ACSLS configuration.
See Figure 8-2 on page 91.
■ The major components in typical configurations.
See Table 8-1 on page 92.
 Oracle StorageTek ACSLS robots 90
Sample ACSLS configurations

The following figure shows a typical UNIX ACSLS configuration.

Figure 8-1 Typical ACSLS configuration on UNIX

ACSLS
NetBackup media server Administrative Utility

ascd IPC Robotic requests

using RPC
acsssi ACS Library Software

acssel
Database
Device Drivers
SCSI SCSI

Library Management
Unit (LMU)

Robotics

Data Drive Library Storage

Control
Module (LSM)
CAP

Unit (CU)
Drive
Drive

The following figure shows a typical Windows ACSLS configuration.

 Oracle StorageTek ACSLS robots 91
Sample ACSLS configurations

Figure 8-2 Typical ACSLS configuration on Windows

ACSLS
NetBackup media server Administrative Utility

Sun Robotic requests

acsd IPC StorageTek using RPC
LibAttach ACS Library Software
Service

Database
Device Drivers

SCSI SCSI

Library Management
Unit (LMU)

Robotics

Data Control Drive Library Storage

Unit

CAP
Module (LSM)
(CU) Drive
Drive

The following table describes the components of the ACSLS configuration.

 Oracle StorageTek ACSLS robots 92
Sample ACSLS configurations

Table 8-1 ACSLS configuration component description

Component Description

NetBackup media server Specifies a host that has NetBackup media server software and is a client to the ACS
library software host.
The NetBackup ACS robotic daemon (acsd) formulates requests for mounts, unmounts,
and inventories. An API then uses IPC communication to routes these requests to:

■ (UNIX) The NetBackup ACS storage server interface (acsssi). The requests are
converted into RPC-based communications and sent to the ACS library software.
■ (Windows) the Oracle StorageTek LibAttach service. This service sends the requests
to the ACS library software.

Oracle StorageTek Specifies that Library Attach for Windows, an ACS library software client application,
LibAttach Service enables Windows servers to use the StorageTek Nearline enterprise storage libraries.

Windows computers only LibAttach provides the connection between Windows and ACS library software through
a TCP/IP network.

Obtain the appropriate LibAttach software from Oracle. See the Veritas support Web site
for the latest compatibility information.

The following ACS library Receives the robotic requests from NetBackup and uses the Library Management Unit
software: to find and mount or unmount the correct cartridge on media management requests.

■ Automated Cartridge On compatible host platforms, you may be able to configure ACS library software and
System Library NetBackup media server software on the same host.
Software (ACSLS)
■ Sun StorageTek Library
Station

Library Management Unit Provides the interface between the ACS library software and the robot. A single LMU
(LMU) can control multiple ACSLS robots.

Library Storage Module Contains the robot, drives, or media.

(LSM)

Control Unit (CU) Specifies that the NetBackup media server connects to the drives through device drivers
and a control unit (tape controller). The control unit may have an interface to multiple
drives. Some control units also allow multiple hosts to share these drives.

Most drives do not require a separate control unit. In these cases, the media server
connects directly to the drives.

CAP Specifies the Cartridge Access Port.

 Oracle StorageTek ACSLS robots 93
Media requests for an ACS robot

Media requests for an ACS robot

The following is the sequence of events for a media request for an ACS robot:
■ The Media Manager device daemon (UNIX) or NetBackup Device Manager
service (Windows) ltid receives the request from bptm.
■ ltid sends a mount request to the NetBackup ACS process acsd.

■ acsd formulates the request.

An API then uses Internal Process Communications (IPC) to send the request
on the following systems:
■ UNIX. The NetBackup ACS storage server interface acsssi. The request is
then converted into RPC-based communications and sent to the ACS library
software.
■ Windows. The Oracle StorageTek LibAttach service. This service sends the
request to the ACS library software.

■ If the Library Storage Module (LSM) in which the media resides is offline, the
ACS library software reports this offline status to NetBackup. NetBackup assigns
the request a pending status. NetBackup retries the request hourly until the LSM
is online and the ACS library software can satisfy the media request.
■ The ACS library software locates the media and sends the necessary information
to the Library Management Unit (LMU).
■ The LMU directs the robotics to mount the media in the drive. When the LibAttach
service (Windows) or acsssi (UNIX) receives a successful response from the
ACS library software, it returns the status to acsd.
■ The acsd child process (that is associated with the mount request) scans the
drive. When the drive is ready, acsd sends a message to ltid that completes
the mount request. NetBackup then begins to send data to or read data from
the drive.

About configuring ACS drives

An ACS robot supports DLT or 1/2-inch cartridge tape drives. If an ACS robot
contains more than one type of DLT or 1/2-inch cartridge tape drive, you can
configure an alternate drive type. Therefore, there can be up to three different DLT
and three different 1/2-inch cartridge drive types in the same robot. If you use
alternate drive types, configure the volumes by using the same alternate media
type. Six drive types are possible: DLT, DLT2, DLT3, HCART, HCART2, and
HCART3.
 Oracle StorageTek ACSLS robots 94
About configuring ACS drives

Before you configure drives in NetBackup, configure the operating system tape
drivers and device files for those drives. For information about how to do so, refer
to the operating system documentation. For guidance about the NetBackup
requirements, see the information about the host operating system in this guide
Use the same methods to create or identify device files for these drives as for other
drives. If the drives are SCSI and connect to the robot through a shared control
unit, the drives share the same SCSI ID. Therefore, you must specify the same
logical unit number (LUN) for each drive.
When you configure ACS drives as robotic in NetBackup, you must include the
ACS drive coordinate information.
The following table shows the ACS drive coordinates.

Table 8-2 ACS drive coordinates

ACS drive coordinate Description

ACS number Specifies the index, in ACS library software terms, that
identifies the robot that has this drive.

LSM number Specifies the Library Storage Module that has this drive.

Panel number Specifies the panel where the drive is located.

Drive number Specifies the physical number of the drive in ACS library
software terms.

The following figure shows the location of this information in a typical ACS robot.
 Oracle StorageTek ACSLS robots 95
Configuring shared ACS drives

Figure 8-3 ACSLS robot and drive configuration information

ACS Library Software

Host ACS Library Software

ACS number (0-126)

Library Management Unit

(LMU)

LSM number (0-23)

ID
SI
Panel number (0-19)

SC
Robotics

ve
ri
D
SCSI ID Control Drive
Unit (CU) Library Storage SCSI ID
Module (LSM) Drive
Drive

Drive number
(0-19)

Configuring shared ACS drives

If the ACSLS server does not support serialization, use the following procedure to
configure shared drives. Shared drives require the NetBackup Shared Storage
Option license. Oracle StorageTek ACSLS versions before 6.1 do not support
serialization.) If the server supports serialization, use the NetBackup Device
Configuration Wizard to configure shared drives.
This procedure can significantly reduce the amount of manual configuration that is
required in an SSO environment. For example, for 20 drives that 30 hosts share,
 Oracle StorageTek ACSLS robots 96
Configuring shared ACS drives

these configuration steps require that you configure only 20 device paths rather
than 600 device paths.
During the setup phase, the NetBackup Device Configuration Wizard tries to
discover the tape drives available. The wizard also tries to discover the positions
of the drives within the library (if the robot supports serialization).
A SAN (including switches rather than direct connection) can increase the possibility
of errors. If errors occur, you can define the tape drive configuration manually by
using the NetBackup Administration Console or NetBackup commands.
Take care to avoid any errors. With shared drives, the device paths must be correct
for each server. Also, ensure that the drives are defined correctly to avoid errors.
(A common error is to define a drive as ACS index number 9 rather than ACS index
0.)
Use the following procedure to configure shared drives in a nonserialized
configuration.
To configure shared drives in a nonserialized configuration
1 Run the NetBackup Device Configuration Wizard on one of the hosts to which
drives in an ACS-controlled library are attached. Allow the drives to be added
as stand-alone drives.
2 Add the ACS robot definition and update each drive to indicate its position in
the robot. Make each drive robotic and add the ACS, LSM, Panel, and Drive
information.
Information about how to determine the correct drive addresses and how to
verify the drive paths is available. See "Correlating device files to physical
drives" in the NetBackup Administrator’s Guide, Volume I.
3 After you verify the drive paths on one host, run the Device Configuration
Wizard again. Scan all hosts that have ACS drives in the library.
The wizard adds the ACS robot definition and the drives to the other hosts and
uses the correct device paths.
For this process to work correctly, the following must be true:
■ The wizard discovered the devices and their serial numbers successfully
the first time.
■ You configured the drive paths correctly on the first host.
 Oracle StorageTek ACSLS robots 97
Adding tapes to ACS robots

Adding tapes to ACS robots

ACS robotic control software supports the following characters in a volume ID that
are not valid NetBackup media ID characters. (Volume ID is the ACS term for media
ID).
Therefore, do not use any of the following characters when you configure ACS
volumes:
■ Dollar sign ($)
■ Pound sign (#)
■ The yen symbol
■ Leading and trailing spaces
The following tables is an overview of how to add tapes to an ACS robot and then
add those tapes to NetBackup.

Table 8-3 Adding tapes to ACS robots process

Task Description

Add barcode labels to the The Library Manager reads the bar codes and classifies the media by media type. A
media and insert the media category is assigned to each volume. Some volume categories restrict application
into the robot by using the access to certain volumes. The Library Manager tracks volume locations.
media access port.

Define the media in To define the media, do one of the following:

NetBackup by using the ACS
■ Update the volume configuration by using the robot inventory function.
volume IDs as media IDs.
■ Add new volumes by using the Volume Configuration Wizard.

See the NetBackup Administrator’s Guide, Volume I:

http://www.veritas.com/docs/DOC5332

Because the ACS volume IDs and bar codes are the same, NetBackup has a record
of the bar codes for the media. Note that you do not enter slot locations because the
ACS library software manages slot locations.

Verify the volume Use Show Contents and Compare Contents with Volume Configuration from the
configuration Robot Inventory dialog.

About removing tapes from ACS robots

You can remove tapes by using the Sun StorageTek utility or by using NetBackup.
See “Removing tapes using the ACSLS utility” on page 98.
 Oracle StorageTek ACSLS robots 98
Robot inventory operations on ACS robots

See “Removing tapes using NetBackup” on page 98.

Removing tapes using the ACSLS utility

If you remove media from an ACS robot, you must move the media logically to
stand alone in NetBackup.
If you do not move media logically, NetBackup does not know that the media were
moved. NetBackup may issue mount requests for it, which causes a misplaced
tape error.
However, you can move media from one location to another within the robot. The
ACS library software finds the requested media if its database is current.
To remove tapes using the SCSLS utility
◆ Do one of the following:
■ Update the volume configuration by using the NetBackup robot inventory
function.
See the NetBackup Administrator’s Guide, Volume I.
http://www.veritas.com/docs/DOC5332
■ Move the volumes.
See the NetBackup Administrator’s Guide, Volume I.
http://www.veritas.com/docs/DOC5332

Removing tapes using NetBackup

To remove tapes using NetBackup
◆ Use one of the following methods:
■ Select Actions > Eject Volumes From Robot in the NetBackup
Administration Console.
■ Use the NetBackup vmchange command.
See the NetBackup Commands Reference Guide.
http://www.veritas.com/docs/DOC5332
Both of these methods performs the logical move and the physical move.

Robot inventory operations on ACS robots

If the ACS library software host is a Sun StorageTek Library Station, an Inventory
Robot Filter (INVENTORY_FILTER) entry may be required in the vm.conf file. Old
versions of Library Station do not support queries of all volumes in an ACS robot.
 Oracle StorageTek ACSLS robots 99
Robot inventory operations on ACS robots

In NetBackup, the ACS robot type supports bar codes.

The following sequence of events occurs when you inventory an ACS robot in
NetBackup:
■ NetBackup requests volume information from the ACS library software.
■ The ACS library software provides a listing of the volume IDs, media types, ACS
location, and LSM location from its database.
See Table 8-4 on page 99.
■ NetBackup maps the volume IDs into media IDs and bar codes. For example
in the previous table, volume ID 100011 becomes media ID 100011 and the
barcode for that media ID is also 100011.
■ If the operation does not require a volume configuration update, NetBackup
uses the media type defaults for ACS robots when it creates its report.
■ If the operation requires a volume configuration update, NetBackup does the
following:
■ Maps the ACS media types to the default NetBackup media types.
■ Adds the ACS and the LSM locations for new volumes to the EMM database.
This location information is used for media and drive selection.

Information about the default media type mappings and how to configure media
type mappings is available.
See the NetBackup Administrator’s Guide, Volume I.
The following table shows an example of the ACS drive coordinates that NetBackup
receives.

Table 8-4 ACS drive coordinates

ACS volume ID ACS media type ACS LSM

100011 DLTIV 0 0

200201 DD3A 0 0

412840 STK1R 0 1

412999 STK1U 0 1

521212 JLABEL 0 0

521433 STK2P 0 1

521455 STK2W 0 1

770000 LTO_100G 0 0
 Oracle StorageTek ACSLS robots 100
Robot inventory operations on ACS robots

Table 8-4 ACS drive coordinates (continued)

ACS volume ID ACS media type ACS LSM

775500 SDLT 0 0

900100 EECART 0 0

900200 UNKNOWN 0 0

Configuring a robot inventory filtering on ACS robots

If you want NetBackup to use only a subset of the volumes under ACS library
control, you can filter the volume information from the library. To do so, you use
the ACSLS administrative interface to assign the volumes you want to use to a
scratch pool or pools. Then you configure NetBackup to use only the volumes in
those scratch pools.
A NetBackup robot inventory includes the volumes that exist in the ACS scratch
pool. The ACS library software moves each volume from the scratch pool after it
is mounted.
A partial inventory also includes those volumes that NetBackup can validate exist
in the robotic library, including volumes not in the ACS scratch pool. To prevent
losing track of previously mounted volumes, the library reports the complete list of
volumes that exist in the robotic library.
The following procedure is an example of how to configure an inventory filter.
To configure an inventory filter (example)
1 Use the ACSLS administrative interface (ACSSA) command to create a scratch
pool. Assign ID 4 and 0 to 500 as the range for the number of volumes, as
follows:

ACSSA> define pool 0 500 4

2 Use the ACSLS administrative interface (ACSSA) command to define the

volumes in scratch pool 4:

ACSSA> set scratch 4 600000-999999

3 On the NetBackup media server from which you invoke the inventory operation,
add an INVENTORY_FILTER entry to the vm.conf file. The following is the
usage statement:

INVENTORY_FILTER = ACS robot_number BY_ACS_POOL acs_scratch_pool1

[acs_scratch_pool2 ...]
 Oracle StorageTek ACSLS robots 101
NetBackup robotic control, communication, and logging

The following define the options and arguments:

■ robot_number is the number of the robot in NetBackup.
■ acs_scratch_pool1 is the scratch pool ID as configured in the ACS library
software.
■ acs_scratch_pool2 is a second scratch pool ID (up to 10 scratch pools are
allowed).
For example, the following entry forces ACS robot number 0 to query scratch
volumes from Sun StorageTek pool IDs 4 and 5.

INVENTORY_FILTER = ACS 0 BY_ACS_POOL 4 5

NetBackup robotic control, communication, and

logging
How NetBackup uses robotic control, communication, and logging during tape
operations depends on the operating system type as follows:
■ Windows systems
See “NetBackup robotic control, communication, and logging for Windows
systems” on page 101.
■ UNIX systems
See “NetBackup robotic control, communication, and logging for UNIX systems”
on page 102.

NetBackup robotic control, communication, and logging for Windows

systems
The NetBackup acsd process provides robotic control to mount and dismount
volumes. It also requests inventories of the volumes that are under the control of
ACS library software. The NetBackup Device Manager service ltid starts the acsd
process and communicates with it.
The acsd process requests SCSI tape unloads through the device host’s tape driver
before it uses the ACS API to request that tape dismounts. This request process
accommodates the configurations that have SCSI multiplexors. Loaded tapes are
not ejected forcibly when a dismount operation occurs.
 Oracle StorageTek ACSLS robots 102
NetBackup robotic control, communication, and logging

NetBackup robotic control, communication, and logging for UNIX

systems
On UNIX systems, several NetBackup daemons and processes provide robotic
control, communication, and logging.

NetBackup ACS daemon (acsd)

The NetBackup ACS daemon acsd provides robotic control to mount and dismount
volumes. It also requests inventories of the volumes that are under the control of
ACS library software. the Media Manager device daemon ltid starts the acsd
daemon and communicates with it. If ltid is active already, you can start acsd
manually.
The acsd daemon requests SCSI tape unloads through the device host’s tape driver
before it uses the ACS API to request that tape dismounts. This control process
accommodates the configurations that have SCSI multiplexors. Loaded tapes are
not ejected forcibly when a dismount operation occurs.
When acsd starts, it first starts the NetBackup acssel process and then starts the
acsssi process. When it starts acsssi, acsd passes the ACS library software host
name to acsssi. One copy of acsssi starts for each ACS library software host that
appears in the NetBackup device configuration for the media server. If multiple
media servers share drives in an ACS robot, acsssi must be active on each media
server.

NetBackup ACS SSI event logger (acssel)

The NetBackup ACS storage server interface (SSI) event logger acssel is modeled
after the Sun StorageTek mini_el event logger. Therefore, its functional model
differs from other NetBackup robotic controls.
The NetBackup acsd daemon starts acssel automatically. You also can start it
manually. Event messages are logged to the following file:
/usr/openv/volmgr/debug/acsssi/event.log

Note: Veritas recommends that acssel run continuously because it tries to connect
on the event logger's socket for its message logging. If acsssi cannot connect to
acssel, NetBackup cannot process requests immediately. Therefore, retry and
error recovery situations can occur.
 Oracle StorageTek ACSLS robots 103
NetBackup robotic control, communication, and logging

On UNIX systems, only the kill command stops acssel. The NetBackup
bp.kill_all utility (UNIX ) stops the acssel process. On Windows systems, the
bpdown.exe program stops the acssel process.

The full path to the event logger is /usr/openv/volmgr/bin/acssel. The usage

format is as follows:

acssel [-d] -s socket_name

The following describes the options:

■ -d displays debug messages (by default, debug messages are disabled).
■ socket_name is the socket name (or IP port) to listen on for messages.

Using acssel with a different socket name

If the vm.conf file does not contain an ACS_SEL_SOCKET entry, acssel listens on
socket name 13740 by default.
You can change this default by using one of the following methods:
■ Modify the vm.conf configuration file.
See To change the default by modifying the vm.conf configuration file.
■ Add environment variables. This method assumes that one ACS robot is
configured and that the SSI default socket name has not been changed. (The
vm.conf ACS_SEL_SOCKET entry can change the default).
See To change the default by adding environment variables.
acssel also has a command line option to specify the socket name. However,
because acsssi needs to know the event logger socket name, setting an
environment variable is preferred.
To change the default by modifying the vm.conf configuration file
1 Edit the vm.conf file and add an ACS_SEL_SOCKET entry. The following is an
example:
ACS_SEL_SOCKET = 13799

2 Stop the acsd, acsssi, and acssel processes by invoking the following script.
(This script stops all NetBackup processes.)
/usr/openv/NetBackup/bin/bp.kill_all

3 Restart the NetBackup daemons and processes by invoking the following script:
/usr/openv/NetBackup/bin/bp.start_all
 Oracle StorageTek ACSLS robots 104
NetBackup robotic control, communication, and logging

To change the default by adding environment variables

1 Stop the acsd, acsssi, and acssel processes by invoking the following script.
(This script stops all NetBackup processes.)
/usr/openv/NetBackup/bin/bp.kill_all

2 Set the wanted socket name in an environment variable and export it. The
following is an example:

ACS_SEL_SOCKET = 13799
export ACS_SEL_SOCKET

3 Start the event logger in the background.

/usr/openv/volmgr/bin/acssel &

4 Set the ACS library software host name for acsssi in an environment variable.

CSI_HOSTNAME = einstein
export CSI_HOSTNAME

5 Start acsssi as follows:

/usr/openv/volmgr/bin/acsssi 13741 &

6 Optionally, start acstest by using the robtest utility or by using the following
command:
/usr/openv/volmgr/bin/acstest -r einstein -s 13741

If you request SCSI unloads, you also must specify drive paths on the acstest
command line.
See “ACS robotic test utility” on page 106.
The robtest utility specifies drive paths automatically if ACS drives have been
configured.
7 Start ltid as follows, which starts acsd. You can use the -v option for verbose
message output.
/usr/openv/volmgr/bin/ltid

During initialization, acsd obtains the SSI Event Logger socket name from
vm.conf and sets ACS_SEL_SOCKET in the environment before it starts acssel.
If acsssi is started manually, it has to use (listen on) the same SSI socket that
acsd uses to send data.
 Oracle StorageTek ACSLS robots 105
NetBackup robotic control, communication, and logging

NetBackup ACS storage server interface (acsssi)

The NetBackup ACS storage server interface (SSI) acsssi communicates with the
ACS library software host. acsssi processes all RPC communications from acsd
or from the ACS robotic test utility that are intended for the ACS library software.
One copy of acsssi must run for each unique ACS library software host that is
configured on a NetBackup media server. acsd tries to start copies of acsssi for
each host. However, if an acsssi process for a specific ACS library software host
exists already, the new acsssi processes for that host fails during initialization.
In normal operations, acsssi runs in the background and sends log messages to
acssel.

You can specify the socket name (IP port) used by acsssi in any of the following
ways:
■ On the command line when you start acsssi.
■ By using an environment variable (ACS_SSI_SOCKET).
■ Through the default value.
If you configure acsssi to use a nondefault socket name, you also must configure
the ACS daemon and ACS test utility to use the same socket name.
The ACS library software host name is passed to acsssi by using the CSI_HOSTNAME
environment variable.
acsssi is based on the Sun StorageTek storage server interface. Therefore, it
supports environment variables to control most aspects of operational behavior.
See “Optional environment variables” on page 106.

About the ACS_SSI_SOCKET configuration option

By default, acsssi listens on unique, consecutive socket names; the socket names
begin at 13741. To specify socket names on an ACS library software host basis,
you can add a configuration entry in the NetBackup vm.conf file.
Use the following format:
ACS_SSI_SOCKET = ACS_library_software_hostname socket_name

The following is an example entry (do not use the IP address of the ACS library
host for this parameter):
ACS_SSI_SOCKET = einstein 13750
 Oracle StorageTek ACSLS robots 106
ACS robotic test utility

Starting acsssi manually

This method is not the recommended method to start acsssi. Normally, acsd starts
acsssi.

Before you can start acsssi manually, you must configure the CSI_HOSTNAME
environment variable. The following is a Bourne shell example:

CSI_HOSTNAME=einstein
export CSI_HOSTNAME
/usr/openv/volmgr/bin/acsssi 13741 &

Use the following procedure to start acsssi.

To start acsssi
1 Start the event logger, acssel.
2 Start acsssi. The format is acsssi socket_name.

Optional environment variables

If you want individual acsssi processes to operate differently, you can set
environment variables before the acsssi processes are started.
The following table describes the optional environment variables.

Table 8-5 Optional environment variables

Environment Description
variable

SSI_HOSTNAME Specifies the name of the host where ACS library software RPC
return packets are routed for ACS network communications. By
default, the local host name is used.

CSI_RETRY_TIMEOUT Set this variable to a small positive integer. The default is 2 seconds.

CSI_RETRY_TRIES Set this variable to a small positive integer. The default is five retries.

CSI_CONNECT_AGETIME Set this variable to a value between 600 seconds and 31536000
seconds. The default is 172800 seconds.

ACS robotic test utility

The acstest utility lets you verify ACS communications and provides a remote
system administrative interface to an ACS robot. It can also be used to query, enter,
 Oracle StorageTek ACSLS robots 107
Changing your ACS robotic configuration

eject, mount, unload, and dismount volumes. In addition, acstest lets you define,
delete, and populate ACS library software scratch pools.
While acsd services requests, do not use acstest. Communication problems may
occur if acsd and acstest process ACS requests at the same time.

acstest on Windows systems

acstest depends on the Sun StorageTek LibAttach service being started
successfully. You can verify that this service is started by using the Services tool
available in administrative tools in the Windows control panel. acstest attempts to
communicate with ACS library software by using the LibAttach service.
The usage format follows:
acstest -r ACS_library_software_hostname [-d device_name ACS, LSM,
panel, drive] ... [-C sub_cmd]

The following example assumes that the LibAttach service started:

install_path\Volmgr\bin\acstest -r einstein -d Tape0 0,0,2,1

acstest on UNIX systems

acstest depends on acsssi being started successfully. You can use the UNIX
netstat -a command to verify that a process listens on the SSI socket. acstest
attempts to communicate with ACS library software using acsssi and connects on
an existing socket.
The usage format follows. You can pass the socket name on the command line.
Otherwise, the default socket name (13741) is used.
acstest -r ACS_library_software_hostname [-s socket_name] [-d
drive_path ACS, LSM, panel, drive] ... [-C sub_cmd]

The following example assumes that the acsssi process has been started by using
socket 13741:
/usr/openv/volmgr/bin/acstest -r einstein -s 13741

Changing your ACS robotic configuration

UNIX and Linux systems only.
If you change your ACS robot configuration, you should update NetBackup so that
acsssi can successfully communicate with acsd, acstest, and ACS library software.
 Oracle StorageTek ACSLS robots 108
ACS configurations supported

Any acsssi processes must be canceled after your changes are made and before
the Media Manager device daemon ltid is restarted. Also, for the acstest utility
to function, acsssi for the selected robot must be running.
Use the following procedure to update NetBackup after you change your
configuration.
To update NetBackup after you change your configuration
1 Make your configuration changes.
2 Use /usr/openv/NetBackup/bin/bp.kill_all to stop all running processes.
3 Restart the NetBackup daemons and processes by invoking the following script:
/usr/openv/NetBackup/bin/bp.start_all

ACS configurations supported

UNIX and Linux systems only.
NetBackup supports the following ACS configurations:
■ Multiple robots that are controlled from a single ACS host
See “Multiple ACS robots with one ACS library software host” on page 108.
■ Multiple robots that are controlled from multiple ACS hosts
See “Multiple ACS robots and ACS library software hosts” on page 109.

Multiple ACS robots with one ACS library software host

NetBackup supports the following configuration:
■ A NetBackup server is connected to drives in multiple ACS robots.
■ The robots are controlled from a single ACS library software host.
The following figure shows multiple ACS robots that are controlled from a single
ACS library software host.
 Oracle StorageTek ACSLS robots 109
ACS configurations supported

Figure 8-4 Multiple ACS robots, one ACS library software host

Robot 1
Sun
NetBackup Server
StorageTek
ACS 0
ACS(10) controls drive 1
ACS Library
ACS(20) controls drive 2
Software
Host
Robot 2
Sun
StorageTek
ACS 1

Network Communications (RPC)

Inventory requests include: the volumes that are configured on the ACS library
software host that resides on the ACS robot that is designated in the drive address.
In this example, assume the following about drive 1:
■ Has an ACS drive address (ACS, LSM, panel, drive) of 0,0,1,1 in the NetBackup
device configuration
■ Is under control of robot number 10 (ACS(10)).
If any other robot ACS(10) drives have a different ACS drive address (for example,
1,0,1,0), the configuration is invalid.
NetBackup supports configurations of multiple LSMs in a single ACS robot if a
pass-through port exists.

Multiple ACS robots and ACS library software hosts

NetBackup supports the following configuration:
■ A NetBackup server is connected to drives in multiple ACS robots.
■ The robots are controlled from separate ACS library software hosts.
The following figure shows multiple ACS robots that are controlled from multiple
ACS library software hosts.
 Oracle StorageTek ACSLS robots 110
Oracle StorageTek ACSLS firewall configuration

Figure 8-5 Multiple ACS robots, multiple ACS library software hosts

Robot 1 ACS Library

NetBackup Server Sun Software
StorageTek Host A
ACS(10) controls drive 1 ACS 0
ACS(20) controls drive 2

Robot 2 ACS Library

Sun Software
StorageTek Host B
ACS 0

Network Communications (RPC)

Inventory requests include the volumes that are configured on the ACS library
software hosts (Host A for Robot 1 and Host B for Robot 2). The software hosts
reside on the robot (ACS 0 for each) that is designated in the Sun StorageTek drive
address.
In this example, assume the following about drive 1:
■ Has an ACS drive address (ACS, LSM, panel, drive) of 0,0,1,1 in the NetBackup
device configuration
■ Is under control of robot number 10 (ACS(10))
If any other robot ACS(10) drives have a different ACS drive address (for example,
1,0,1,0), the configuration is invalid.
NetBackup supports configurations of multiple LSMs in a single ACS robot if a
pass-through port exists.

Oracle StorageTek ACSLS firewall configuration

To configure an ACS robot in an Sun StorageTek ACSLS firewall environment, use
the following NetBackup vm.conf file configuration entries to designate TCP port
connections:
■ ACS_CSI_HOSTPORT

■ ACS_SSI_INET_PORT

■ ACS_TCP_RPCSERVICE

More information about vm.conf entries is available.

See the NetBackup Administrator’s Guide, Volume I.
 Oracle StorageTek ACSLS robots 111
Oracle StorageTek ACSLS firewall configuration

The Sun StorageTek ACSLS server configuration options must match the entries
in the vm.conf file. For example, in a typical ACSLS firewall configuration, you
would change the following settings as shown:
■ Changes to alter use of TCP protocol...
Set to TRUE - Firewall-secure ACSLS runs across TCP.
■ Changes to alter use of UDP protocol...
Set to FALSE - Firewall-secure ACSLS runs across TCP.
■ Changes to alter use of the portmapper...
Set to NEVER - Ensures that the ACSLS server does not query the portmapper
on the client platform.
■ Enable CSI to be used behind a firewall...
Set to TRUE - Allows specification of a single port for the ACSLS server.
■ Port number used by the CSI...
The port that the user chooses. The 30031 default value is used most often.
This port number must match the port number that you specify in the NetBackup
vm.conf file.
For complete information about setting up a firewall-secure ACSLS server, refer to
your vendor documentation.
 Chapter 9
Device configuration
examples
This chapter includes the following topics:

■ An ACS robot on a Windows server example

■ An ACS robot on a UNIX server example

An ACS robot on a Windows server example

The following figure shows a Windows server and ACS robot configuration.
 Device configuration examples 113
An ACS robot on a Windows server example

Figure 9-1 Windows server and ACS robot configuration example

Windows server shark ACSLS host whale

acsd

Automated Cartridge System

Library Software – (ACS 0)
STK LibAttach

SCSI

Library Management Unit (LMU)

Data
Panel 2 Robotics

lun 0
Drive 0

CAP
Library Storage
Control unit (cu)
lun 1 Module (LSM 0)
Drive 1

This configuration uses an Automated Cartridge System (ACS) robot for storage.
Server shark can be a Windows NetBackup master server or media server.
The following are items to note when you review this example:
■ The Oracle StorageTek ACSLS host (in the Add Robot dialog) is host whale,
where the ACS library software resides. In this example, Automated Cartridge
System Library Software (ACSLS) is installed as the ACS library software.
On some server platforms, you can run NetBackup media server software and
ACS library software on the same server. Therefore, you need only one server.
■ The ACS, LSM, PANEL, and DRIVE numbers are part of the ACS library software
configuration and must be obtained from the administrator of that host.
■ Robot number and ACS number are different terms. Robot number is the robot
identifier used in NetBackup. ACS number is the robot identifier in ACS library
software. These numbers can be different, although they both default to zero.
 Device configuration examples 114
An ACS robot on a Windows server example

■ If you connnect the drives through an independent control unit, you must use
the correct Logical Unit Numbers (LUNs) so that the correct tape name is used.
■ The Add Robot dialog entries include an ACSLS Host entry so that the ACS
library software host communicates by using STK LibAttach software. This
software must be installed on each Windows server that has the ACS drives
attached to it.
The following table shows the robot attributes for the remote host shark.

Table 9-1 Add Robot dialog entries (remote host)

Dialog box field Value

Device Host shark

Robot Type ACS (Automated Cartridge System)

Robot Number 0

Robot control is handled by a remote host Set (cannot be changed for this robot type)

ACSLS Host whale

The following table shows the drive 0 attributes.

Table 9-2 Add Drive dialog entries (drive 0)

Dialog box field Value

Device Host shark

Drive Type 1/2" Cartridge (hcart)

Drive Name shark_drive_0

Path Information [5,0,1,0]

Drive is in a Robotic Library Yes

Robotic Library ACS(0) - whale

ACS ACS: 0

LSM: 0

PANEL: 2

DRIVE: 0

The following table shows the drive attributes for drive 1.

 Device configuration examples 115
An ACS robot on a UNIX server example

Table 9-3 Add Drive dialog entries (drive 1)

Dialog box field Value

Device Host shark

Drive Type 1/2" Cartridge (hcart)

Drive Name shark_drive_1

Path Information [4,0,1,1]

Drive is in a Robotic Library Yes

Robotic Library ACS(0) - whale

ACS ACS: 0

LSM: 0

PANEL: 2

DRIVE: 1

An ACS robot on a UNIX server example

The following figure shows a UNIX server and ACS robot configuration.
 Device configuration examples 116
An ACS robot on a UNIX server example

Figure 9-2 UNIX server and ACS robot configuration example

UNIX server shark ACSLS host whale

acsd

Automated Cartridge System

Library Software – (ACS 0)
acsssi

SCSI

Library Management Unit (LMU)

Data
Panel 2 Robotics

lun 0
Drive 0

CAP
Library Storage
Control unit (cu)
lun 1 Module (LSM 0)
Drive 1

This configuration uses an Automated Cartridge System (ACS) robot for storage.
Host shark can be a UNIX NetBackup master server or media server.
The following are some items to note when you review this example:
■ The ACSLS Host (in the Add Robot dialog) is server whale, where the ACS
library software resides. In this example, Automated Cartridge System Library
Software (ACSLS) is installed as the ACS library software.
On some server platforms, you can run NetBackup media server software and
ACS library software on the same server. Therefore, you need only one server.
■ The ACS, PANEL, LSM, and DRIVE numbers are part of the ACS library software
configuration and must be obtained from that system.
■ Robot number and ACS number are different terms. Robot number is the robot
identifier used in NetBackup. ACS number is the robot identifier in ACS library
software. These numbers can be different, although they both default to zero.
 Device configuration examples 117
An ACS robot on a UNIX server example

■ If you connnect the drives through an independent control unit, you must use
the correct Logical Unit Numbers (LUNs) so that the correct tape name is used.
■ The Add Robot dialog entries include an ACSLS Host entry. That entry configures
NetBackup to use the ACS Storage Server Interface (acsssi) to communicate
with the ACS library software host.
The following table shows the robot attributes.

Table 9-4 Add Robot dialog entries (remote host)

Dialog box field Value

Device Host shark

Robot Type ACS (Automated Cartridge System)

Robot Number 0

Robot control is handled by a remote host Set (cannot be changed for this robot type)

ACSLS Host whale

The following table shows the drive 0 attributes.

Table 9-5 Add Drive dialog entries (drive 0)

Dialog box field Value

Device Host shark

Drive Name shark_drive_0

Drive Type 1/2" Cartridge (hcart)

No Rewind Device /dev/rmt1.1

Drive is in a Robotic Library Yes

Robotic Library ACS(0) - whale

ACS ACS Number: 0

LSM Number: 2

PANEL Number: 0

DRIVE Number: 0

The following table shows the drive 1 attributes.

 Device configuration examples 118
An ACS robot on a UNIX server example

Table 9-6 Add Drive dialog entries (drive 1)

Dialog box field Value

Device Host shark

Drive Name shark_drive_1

Drive Type 1/2" Cartridge (hcart)

No Rewind Device /dev/rmt1.1

Drive is in a Robotic Library Yes

Robotic Library ACS(0) - whale

ACS ACS Number: 0

LSM Number: 2

PANEL Number: 0

DRIVE Number: 1
 Index

A acsssi (continued)
ACS. See Automated Cartridge System starting manually 106
ACS daemon (acsd) acstest 104, 107
NetBackup 102 on UNIX systems 107
ACS drives on Windows systems 107
configuring 93 adding tapes
ACS robot to ACS robots 97
ACSLS firewall configuration 110 agile addressing 26
changing configuration 107 AIX
ACS robot on a UNIX server adapter number conventions 13
configuration example 116 adapter numbers 13
ACS robot on a Windows server command summary 23
configuration example 113 configuring robotic control device files for IBM
ACS robot type 81 robots 14
ACS robotic test utility 107 configuring tape drive device files 15
ACS robots 81 disabling SPC-2 SCSI reserve 22
adding tapes 97 introduction 12
removing tapes 97 locate-block 17
robot inventory filtering 100 smit tool 12
robot inventory operations 99 tape drive configuration
with mulitple ACS hosts 109 extended file marks 16
with single ACS host 108 make device files 17
ACS SSI event logger (acssel) multiple densities 21
NetBackup 102 variable mode devices 16
using with a different socket name 103 AL-PA destination ID
ACS storage server interface (acsssi) Solaris 60
NetBackup 105 alternate media types
starting manually 106 ACS robots 93
ACS_SSI_SOCKET API robots 89
configuration option 105 atdd driver
acsd daemon 102 HP-UX 45
acsd process attaching devices
NetBackup 101 to a Windows system 78
ACSLS attributes
configurations 89 robot 81
ACSLS utility Automated Cartridge System
removing tapes 98 adding volumes 97
acssel 102 barcode operations 99
using with a different socket name 103 configuration example 113, 116
acsssi 105 configurations supported 108
environment variables 106 Library Server (ACSLS) 89, 92
media requests 93
 Index 120

Automated Cartridge System (continued) configuring (continued)

multiple ACS robots with multiple ACS hosts 109 tape drive device files in AIX 15
multiple ACS robots with single ACS host 108 control unit
robot inventory filtering 100 ACS 92
Solaris 57 conventions
special characters 97 RS/6000 AIX adapter number 13
STK Library Station 89, 92 creating
Automated Cartridge System (ACS) HP-UX persistent DSFs 30
removing tapes 97 legacy SCSI and FCP robotic controls on
HP-UX 36
B legacy tape drive device files 43
no rewind device files for tape drives 17
Berkeley-style close
persistent DSF pass-through paths 32
HP-UX 27
sctI device file for FCP (Itanium) 41
Solaris 71
sctI device file for FCP (PA-RISC) 39
boot -r
sctl device file for SCSI (PA-RISC) 37
Solaris 76
tape drive pass-through device files 43
creating device files
C for SAN clients on AIX 14
cfgmgr command 24 for SAN clients on HP-UX 35
chdev command 16, 24
choosing
tape driver 15
D
debug mode
command summary
for st tape driver 51
for AIX 23
device
HP-UX 48
configuration wizard 95
Linux 55
device addressing schemes
Solaris 75
HP-UX 26
configuration
device configuration sequence 9
changing ACS robot 107
device discovery 9
for Sun StorEdge Network Foundation HBA
device drivers
driver 59
for legacy device files 33
configuration example
for persistent DSFs 28
ACS robot on a UNIX server 116
sg
ACS robot on a Windows server 113
Linux 50
configuration guidelines
Solaris 57
HP-UX 25
st
configuration option
Linux 51
ACS_SSI_SOCKET 105
device drivers and files
configurations supported
for persistent DSFs 28
Automated Cartridge System 108
device files
configuring
creating for SAN clients on AIX 14
ACS drives 93
creating for SAN clients on HP-UX 35
legacy device files 35
creating no rewind 17
persistent DSFs 30
for legacy tape drives 33
robotic control device files for IBM robots in
Linux (2.6 kernel) robotic control 52
AIX 14
Linux (2.6 kernel) tape drive 52
SAN clients to recognize FT media servers 73
no rewind 17
shared ACS drives 95
 Index 121

device special files forward-space-file/record

persistent 28 HP-UX 28
disable
HP-UX EMS Tape Device Monitor for a SAN 47 H
SPC-2 SCSI reserve in HP-UX 47
Hardware Compatibility List (HCL) 80
disabling
HP-UX
Solaris Multiplexed I/O (MPxIO) 60
command summary 48
SPC-2 SCSI reserve in AIX 22
configuration guidelines 25
SPC-2 SCSI reserve on Solaris 72
creating legacy SCSI and FCP robotic controls 36
driver unloading
creating persistent DSFs 30
Solaris 67
device addressing schemes 26
drives
disable EMS Tape Device Monitor for a SAN 47
Sony S-AIT 22
disable SPC-2 SCSI reserve 47
drstat command 84
legacy device drivers and files 32
DSFs. See device special files 28
robotic control 26
SAM utility 47
E SCSI robotic controls 33
environment variables SPC-2 SCSI reserve 47
for acsssi processes 106 Tape drive configuration
example Berkeley-style close 27
robotic process 86 make device files 28
sg.conf file 63 upgrading NetBackup to use persistent DSFs 31
sg.links file 64
st.conf file 63 I
example configuration
introduction
ACSLS 89
Linux 49
examples
Solaris 56
no rewind device file 19
ioscan command
SCSI and FCP robotic control device files 69
HP-UX 48
extended file marks
for drives 16
L
legacy device drivers and files
F HP-UX 32
fabric assigned destination ID
legacy device files
Solaris 60
configuring 35
fast-tape positioning. See locate-block
device drivers supported 33
Fibre Channel
legacy pass-through paths
drivers 67
for tape drives 34
HP-UX configuration example 39, 41
legacy tape drive device files
fibre channel
creating 43
binding process
legacy tape drives
Solaris 59
device file names 33
Fibre Channel HBA drivers
Library Management Unit 92
binding 59
Library Management Unit (LMU) 93
firewall configuration
Library Storage Module 92
Sun StorageTek ACSLS 110
Library Storage Module (LSM) 93
fixed length block 16
Linux
command summary 55
 Index 122

Linux (continued) NetBackup sg driver

introduction 49 verify installation 58
loading drivers 51 no rewind device file example 19
robotic controls 51 no rewind device files 17
SAN clients 53 creating 17
SCSI robotic controls 51 Solaris 71
sg driver 50 nonstandard tape drives 73
utilities to test SCSI devices 54
verifying the device configuration 52 O
Linux (2.6 kernel)
odmget command 24
robotic controls 52
Oracle
tape drive device files 52
configuring shared ACS drives 95
LMU. See Library Management Unit
StorageTek ACSLS robots 89
locate-block
on AIX 17
on Solaris 71 P
lsattr command 24 pass-through paths
lsdev command for persistent DSFs 29
HP-UX 48 persistent DSF pass-through paths
on AIX 23 creating 32
LSM. See Library Storage Module persistent DSFs
lsmod command configuring 30
Linux 51 device drivers 28
device drivers and files 28
for robotic control 29
M for tape drive access 29
media requests
pass-through paths 29
for an ACS robot 93
processes
mknod command
by robotic type 85
HP-UX 48
robotic 84
modinfo command
robotic control 84
Solaris 76
modprobe command
Linux 51 R
mt command rem_drv command
Linux 54 Solaris 76
multiple tape densities removing
using 21 tapes from ACS robots 97
tapes using NetBackup 98
tapes using the ACSLS utility 98
N requirements
NetBackup
tape drive device files 27
ACS daemon (acsd) 102
robot
ACS SSI event logger (acssel) 102
attributes 81
ACS storage server interface (acsssi) 105
process 84
acsd process 101
robot inventory operations
removing tapes 98
on ACS robots 99
robotic control, communication, and logging 101
robot types 80
using acssel with different socket name 103
robotic processes 85
 Index 123

robotic SCSI
control process 84 pass through driver
test utilities 84 Solaris 57
test utilities, ACS 107 robotic control
robotic control HP-UX 33
for persistent DSFs 29 Linux (2.6 kernel) 52
on HP-UX 26 on Linux 51
on UNIX systems 102 on Solaris 69
on Windows systems 101 SCSI persistent bindings 54
robotic control device files SCSI reservations
for IBM robots in AIX 14 data integrity 9
robotic control, communication, and logging disabling 9
during tape operations 101 disabling SPC-2 reserve in AIX 21
robotic controls disabling SPC-2 reserve in HP-UX 47
SCSI disabling SPC-2 reserve on Solaris 72
HP-UX 33 sctl device file
Linux (2.6 kernel) 52 creating for FCP (Itanium) 41
on AIX 51, 69 creating for FCP (PA-RISC) 39
Solaris 68 creating for SCSI (PA-RISC) 37
robotic inventory sg driver
filtering 100 Linux 50
robotic process Solaris 57
example 86 uninstalling 75
robotics sg.build command
table-driven 83 Solaris 76
robots sg.conf file
ACS 81 example 63
Oracle StorageTek ACSLS 89 sg.install script
TLD 82 Solaris 60, 76
robtest 84, 104 sg.links file
robtest utility example 64
Linux 55 shared ACS drives
configuring 95
S smit command 16
Solaris
SAM utility
adapter card removal 57
HP-UX 47
binding Fibre Channel HBA drivers 59
SAN clients
command summary 75
about drivers for Linux 53
configuring SAN clients 73
configuring drivers on AIX 14
disabling MPxIO 60
configuring drivers on HP-UX 35
disabling SPC-2 reserve on Solaris 72
configuring drivers on Solaris 73
examples of SCSI and FCP robotic control device
schgr device driver
files 69
HP-UX 36
introduction 56
Scripts
locate-block 71
sg.install
no rewind device files 71
Solaris 60
preventing driver unloading 67
scripts
robotic controls 68–69
sgscan 69, 76
SCSI pass through driver 57
 Index 124

Solaris (continued) Tape drive configuration

SCSI robotic controls 69 HP-UX
sg driver install or reconfigure 60 no rewind 28
sg.install script 60 tape drive configuration
SPC-2 SCSI reserve 72 on AIX
tape drive configuration 70 make device files 17
Berkeley-style close 71 Solaris 70
using ACS 57 tape drive device files
Solaris Multiplexed I/O (MPxIO) Linux (2.6 kernel) 52
disable 60 requirements 27
Sony S-AIT drives 22 tape drive device files in AIX
SPC-2 SCSI reserve configuring 15
disable in HP-UX 47 tape drive pass-through device files
disabling in AIX 22 creating 43
Solaris 72 tape driver
SSO choosing 15
configuring shared ACS drives 96 tape drives
st driver creating no rewind device files 17
Linux 51 legacy pass-through paths 34
st tape driver nonstandard 73
debug mode 51 tapes
st.conf file removing from ACS robots 97
example 63 removing using NetBackup 98
Sun removing using the ACSLS utility 98
acstest utility on UNIX 107 text version of this guide 8
acstest utility on Windows 107 TLD robots 82
Automated Cartridge System
multiple ACS robots with multiple ACS U
hosts 109
uninstalling
multiple ACS robots with single ACS
the sg driver 75
host 108
UNIX
Automated Cartridge System (ACS)
acstest utility 107
configurations supported 108
upgrading
Sun StorageTek ACSLS
NetBackup to use HP-UX persistent DSFs 31
firewall configuration 110
using
Sun StorEdge Network Foundation HBA driver
guide 8
configuration 59
using the passthru driver
switch settings
capabilities 58
Sony S-AIT
on AIX 22
V
variable length block 16
T variable-mode devices
table-driven robotics 83
on AIX 16
tape device drivers
verifying
Windows 78
device configuration on Linux 52
tape drive access
persistent DSFs 29
 Index 125

W
Windows
acstest utility 107
attaching devices 78
tape device drivers 78
world wide node names (WWNN) 60
world wide port names (WWPN) 59–60