Netbackup Advanced Troubleshooting
Netbackup Advanced Troubleshooting
September 2005
September 2005
Disclaimer
The information contained in this publication is subject to change without notice. VERITAS Software
Corporation makes no warranty of any kind with regard to this manual, including, but not limited to,
the implied warranties of merchantability and fitness for a particular purpose. VERITAS Software
Corporation shall not be liable for errors contained herein or for incidental or consequential damages
in connection with the furnishing, performance, or use of this manual.
Third-Party Copyrights
For a list of third-party copyrights, see that Appendix in the NetBackup Release Notes for UNIX,
Windows, and Linux.
ii NetBackup Release Notes - Additional Operational Notes for UNIX, Windows, and Linux
Contents
Chapter 1. Additional NetBackup 6.0 Operational Notes . . . . . . . . . . . . . . . . . . . . . . . 1
NetBackup General Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
NetBackup Documentation Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
NetBackup Installation and Upgrade Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
NetBackup Backward Compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Cluster Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
NetBackup Firewall and Port Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
NetBackup Windows Clients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Localization and Internationalization Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
NetBackup Port Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
NetBackup Access Control (NBAC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
NetBackup Advanced Client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
bpadm Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
FlashBackup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Network Appliance (NetApp) SnapVault . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Network Appliance (NetApp) NearStore Disk Storage Unit . . . . . . . . . . . . . . . . . . . . . 27
NetBackup Bare Metal Restore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
NetBackup Catalog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
NetBackup Database Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
DB2 Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
DB2 Agent With the Advanced Client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Informix Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Lotus Notes Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
iii
Microsoft Exchange Server Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
NetBackup for Exchange with Advanced Client . . . . . . . . . . . . . . . . . . . . . . . . . 32
Oracle Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
SAP Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
SharePoint Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
SQL Server Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Sybase Server Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
NetBackup Disk Backup Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Disk Media Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Inline Copy limitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
NetBackup Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
General NetBackup Java Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
NetBackup-Java Backup Policy Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
NetBackup Remote Administration Console . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
NetBackup Media Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Media and Device Selection (MDS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
General Media Manager Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Enhanced Device Discovery and Auto Configuration . . . . . . . . . . . . . . . . . . . . . 47
Media Manager Cluster Support Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Storage Migrator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
NDMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
NetBackup Operations Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
NetBackup Vault . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
NetBackup Operations Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Platforms Supported for the NOM Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
NetBackup Managed Master Server Platforms Supported by NOM . . . . . . . . 53
Web Browsers Supported by NOM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2. VERITAS logview (vxlogview) hangs when you try to display logs from a large
number of files.
This situation occurs because of a file-descriptors quota limitation that exists on UNIX
platforms. While attempting to view a large number of files, if you do not assign a
large enough file-descriptors quota, then vxlogview hangs when it tries to acquire
more descriptors.
The following example demonstrates how you can determine the number of files that
vxlogview is trying to display.
If you are trying to display logs for product 1234 using the command, vxlogview -p
1234, you can use the same condition or query with vxlogmgr to list the files and file
count. For example, use the command, vxlogmgr -s -p 1234, to list the files.
To find the file-descriptors quota allocated to you, use the command, ulimit -n, to
display the file-descriptor quota.
1
NetBackup General Notes
To avoid this issue, you can increase the file-descriptors quota using the command
ulimit -n file count displayed by vxlogmgr + 10
3. When configuring NetBackup 6.0 for HP Tru64 systems, you may need to adjust
kernel parameters if you receive the following error:
Insufficient system resources - failed to allocate a SYSV
semaphore
This error indicates that you have an insufficient number of semaphores because you
may be exceeding the number of clients and engine processes that can be run at any
one time. If this is the case, then you should increase the sem_mni and sem_msl
semaphore values.
Run the following commands to first determine how your current kernel parameters
are set, and then to change these values.
Run the following command to verify your system kernel parameters:
sysconfig -q ipc
A list of the kernel parameters will appear.
Run the following two commands to change the sem_mni and sem_msl semaphore
values respectively:
sysconfig -r ipc sem_mni=1000
sysconfig -r ipc sem_msl=1000
Note If the ipc subsystem does not support dynamic reconfiguration and you receive
an error when running the sysconfig command, use the dxkerneltuner
application to change the ipc settings.
a. Type: cd /usr/openv/netbackup/bin
a. Type: cd install_path/netbackup/bin
5. Access control lists (ACLs) are supported on Solaris, HP-UX and AIX systems in this
release of NetBackup, however, ACLs are not supported on Linux systems.
7. For NetBackup to support Block-level Increment (BLI) on an AIX platform, you must
have VxFS 3.4.6.2 installed.
8. Checkpoint Restart is not supported for restores from Backup Exec media. The restore
job may be suspended and resumed; however, a resumed restore job will restart at the
beginning of the job, since no checkpoints for the restore are available. Failed restore
jobs will be in the incomplete state rather than the done state. Like a suspended job,
the incomplete jobs may be resumed, but will restart at the beginning of the job.
9. When restarting NetBackup on system running Solaris 10 you may receive errors in
/var/adm/messages. For example, you may encounter the following message:
pam_dial_auth: terminal-device not specified by login,
returning error in underlying service module
This message does not hinder NetBackup from functioning; therefore, you can safely
ignore it.
10. After upgrading from NetBackup 5.x to NetBackup 6.0, it may be possible to run
back-level vmquery commands on a back-level host and see the old volume database
records that existed before the upgrade. Enhancements have been made in NetBackup
6.0 that limit the changes to the old volume database no longer in use, however, these
enhancements do not limit a user from viewing the old volume database records.
Note The user should use the -h emm_server option to point the vm command at
the proper host to get the correct records.
12. Backup selections for VSS snapshot backups must use the same volume management
software. For the VSS snapshot method, the files to be backed up, cannot be mixed: all
must be in Storage Foundation 4.0 volumes, or all must not be in Storage Foundation
4.0 volumes.
13. There may be a delay in detecting of loss of connection from a NetBackup master
server to a media server.
In certain situations, there can be a delay on a NetBackup master server before it
detects that the connection to a media server has been aborted. For example, if a
media server goes down while running a backup, there may be a delay on the master
server before it detects that the media server is no longer available.
While at first it may appear that there is a problem with the NetBackup master server,
this delay is actually a result of a certain TCP/IP configuration parameter that is set to
7,200,000 (two hours, in milliseconds) by default. VERITAS recommends that you set
this parameter to 15 minutes (900,000 milliseconds) or the next lowest, configurable
value that is greater than 15 minutes. (This parameter is called KeepAliveTime on
Microsoft Windows and tcp_keepalive_interval on Solaris systems. You can check the
man page on your system to identify what this parameter is called. You can find more
information about the KeepAliveTime and other associated TCP/IP configuration
parameters on Microsoft Windows in the following Microsoft Knowledge Base
articles: Q140325 and Q120642 or using by man tcp on UNIX systems.)
14. Files Written count (Current, Previous) in Activity Monitor does not include extra
directory entries used for restore file browsing.
When viewing the Files Written count for a job in the Activity Monitor, the value for
Current Files Written does not include the extra directory entries (which are used
during restore file browsing). The value displayed for Estimated Files Written does
include those extra directory entries, and are slightly larger.
1. Updates and enhancements have been made to the nbemmcmd command/man page
that supersedes what is documented with the release of the NetBackup 6.0
documentation set. Because this is a new and important command for our
NetBackup 6.0 customers, an update of this command has been documented in a
TechNote that can be viewed from the VERITAS support web site. The TechNote
number is 278592. Here is a link to this document,
http://support.veritas.com/docs/278592.
Note Other command enhancements that were developed after the official release of
the NetBackup documentation may also be included in this TechNote.
Note Stop bpdbm and bprd before using this command. Also, ensure that bpcd is
running on any system that is being recovered.
This note is not exactly correct. The note should state that these two processes should
be stopped when recovering an offline backup only.
3. There are a few references in the NetBackup System Administrator’s Guide for UNIX
and Linux that tell the user to install VxSS from the VxSS installation CD. This is
misleading. The documentation should tell the user to install VxSS from the ICS CD
for NetBackup.
4. The following errors appear in the NetBackup for SAP System Administrator’s Guide for
UNIX and Linux:
◆ On page 113, the following text is incorrect:
The switch_list parameter must specify a file path in the following location:
$SAPDATA_HOME/sapbackup/.switch.lis
It should read as follows:
The switch_list parameter must specify a file path in the following location:
$ORACLE_HOME /sapbackup/.switch.lis
◆ On page 114, the following text is incorrect:
The switch_log parameter must specify a file path in the following location:
$SAPDATA_HOME/sapbackup/.switch.log
It should read as follows:
The switch_log parameter must specify a file path in the following location:
$ORACLE_HOME/sapbackup/.switch.log
◆ On page 114, the following text is incorrect:
The switch_sem parameter must specify a file path in the following location:
$SAPDATA_HOME/sapbackup/.switch.sem
It should read as follows:
The switch_sem parameter must specify a file path in the following location:
$ORACLE_HOME /sapbackup/.switch.sem
5. The following errors appear in the NetBackup for SAP System Administrator’s Guide for
Windows:
◆ On page 101, the following text is incorrect:
The switch_list parameter must specify a file path in the following location:
%SAPDATA_HOME%\sapbackup\.switch.lis
It should read as follows:
The switch_list parameter must specify a file path in the following location:
%SAPBACKUP%\sapbackup\.switch.lis
◆ On page 102, the following text is incorrect:
The switch_log parameter must specify a file path in the following location:
%SAPDATA_HOME%\sapbackup\.switch.log
It should read as follows:
The switch_log parameter must specify a file path in the following location:
%SAPBACKUP%\sapbackup\.switch.log
◆ On page 102, the following text is incorrect:
The switch_sem parameter must specify a file path in the following location:
%SAPDATA_HOME%\sapbackup\.switch.sem
It should read as follows:
The switch_sem parameter must specify a file path in the following location:
%SAPBACKUP%\sapbackup\.switch.sem
2. If you are leaving any clustered media servers at NetBackup 5.x, you must run the
following nbpushdata command before running, nbpushdata
-modify_5x_hosts.
For each node of a clustered 5.x media server, run:
nbpushdata -add_5x_inactivenode host
Although it is not necessary to run this command for the node that was active when
nbpushdata -add was run on the Global Device Database host, running it for all of
the nodes does not require you to know which node was active.
Since nbpushdata -add_5x_inactivenode host will not overwrite an existing
device configuration for a node in the EMM database, you will receive an error when
the command is run for the node that was active. This is expected and it is not a
problem.
3. Before running nbpushdata, if you have storage units associated with an application
in a cluster environment (also known as an application cluster type in NetBackup 6.0),
the application cluster type must exist in the EMM database.
a. Run the following command to add a host with the virtual name of the
application cluster:
nbemmcmd -addhost -machinename virtual_name -machinetype
app_cluster
where virtual_name is the virtual name of the application.
b. After nbpushdata is complete, run the following command to associate the host
with each node on which the application runs in the cluster:
nbemmcmd -updatehost -add_server_to_app_cluster -machinename
node_name -machinetype media -clustername virtual_name
-netbackupversion 6.0 -masterserver master_name
4. If you have a Tru64 cluster that is being upgraded to NetBackup 6.0, the procedure to
run nbpushdata is different than for the other types of clusters. Normally, it is only
necessary to run nbpushdata once in a cluster, but Tru64 clusters are different. For
Tru64 clusters only, the following procedure must be followed:
a. Run nbpushdata -add on the active node, just as on the other cluster types.
b. Failover to each of the inactive nodes and run nbpushdata -add -nocheck.
5. When installing on an HP-UX system, an error may occur while creating the
NetBackup database during the installation process. This error, shown below, can be
corrected by increasing the number of semaphores in the HP-UX kernel. For a detailed
explanation on how to change HP-UX kernel parameters, refer to TechNote number
243461 on the VERITAS support web site.
SQL error (-832) -- Connection error: Insufficient system resources - failed to allocate a
SYSV semaphore
6. If you have a pre-NetBackup 6.0 version (with NBAR or GDM) installed on your
server, you must uninstall GDM and NBAR from your server completely before you
upgrade to NetBackup 6.0.
7. After you have successfully installed NetBackup 6.0 server on an HP-UX machine,
you should issue the following commands to determine if you need to update your
HP-UX patches:
# cd /usr/openv/netbackup/bin/admincmd
# ./bpstsinfo -lsuinfo -serverprefix STSBasicDisk:
-isremotecall
bpstsinfo should run for a few seconds and issue this message:
EXIT STATUS = 0
If the bpstsinfo command hangs, you should update your HP-UX patches. Refer
again to the “NetBackup Product Dependencies” chapter for the latest HP-UX
patches.
9. If you have changed the application association of the cmd files from the default in the
Windows NT Command Script on your machine, the VERITAS Private Branch
Exchange SCM service will not start and you may receive the following error:
Error 1920: Service VERITAS Private Branch Exchange (VRTSpbx)
failed to start. Verify that you have sufficient privileges to
start system services
To avoid this issue, refer to TechNote number 276504 on the VERITAS support web
site for a detailed procedure on how to work around this problem.
10. If you are upgrading from a previous version of NetBackup and currently have a
volume pool named CatalogBackup, there will be a volume pool name conflict. Prior to
the upgrade, users should do the following in the previous version of NetBackup;
b. Move the media that is currently in the CatalogBackup volume pool into the new
volume pool.
11. When upgrading from a NetBackup Server evaluation license key to a permanent
NetBackup Enterprise Server product license key on Windows 2000 or greater, restart
the NetBackup services. Restarting the services will ensure that the Help > About
NetBackup menu displays the proper licensing information.
12. The following files remain after the uninstall of NetBackup 6.0. (The files listed below
are shown in their default installation directories.)
C:\Program Files\Common Files\VERITAS Shared\VxFI\Bin\firtl.dll
C:\WINDOWS\system32\MSVCP71.dll - 488KB (499,712 bytes) - v.7.10.3077.0
C:\WINDOWS\system32\MSVCR71.dll - 340KB (348,160 bytes) - v.7.10.3052.4
13. The following files, installed by NetBackup 6.0, are missing file information, such as;
company name, product name, and/or file version.
drvwddm.dll - 12/28/01 03:56 AM - 116KB (118,784 bytes) - v.2.56.18a
drvwenu.dll - 04/26/01 12:01 AM - 48KB (49,152 bytes) - v.1.01.04c
drvwunin.exe - 04/26/01 12:01 AM - 24KB (24,576 bytes) - v.1.01.04c
14. The following are special instructions that provide clarification for installing
NetBackup client software in non-global zones.
On a server whose non-global zone is not using a read-only loopback device for the
/usr directory, the client push-install or local install is done normally. Please see the
appropriate installation guide.
On a server that:
◆ Does not have a master/media server on the global zone
◆ Has a sparse non-global zone using a read-only loopback device for the /usr
directory
The procedure is:
a. In the global zone, create /usr/openv as a sym-link to the location in which you
will be installing the software to in the non-global. This will need to be done even
if the global zone does NOT have that directory.
For eaxmple:
# ln -s /nonglobdir/openv /usr/openv
# ls /nonglobdir/openv
/nonglobdir/openv: No such file or directory
# ls -al /usr/openv
rwxrwxrwx 1 root root 10 Aug 23 15:13
/usr/openv -> /nonglobdir/openv
c. In the non-global zone, make sure that the directory linked exists and is writable.
For eaxmple:
# ls -al /nonglobdir/openv
total 32
drwxr-xr-x 9 root bin 512 Aug 18 15:23 ./
d. Client push-install or local install can now be done normally according to guides.
On a server that:
◆ Has a master/media server installed on the global zone
◆ Has a sparse non-global zone using a read-only loopback device for the /usr
directory
The procedure is:
b. In the non-global zone, create a writable directory where the /usr/openv points
to.
For eaxmple:
# mkdir /opt/openv
c. Client push-install or local install can now be done normally according to guides.
NetBackup has supported mixed patch-level versions for many releases. Mixed
patch-level support enables NetBackup servers to run a mixture of NetBackup major
releases and patches in the same environment. For example, with this release of
NetBackup a customer can have the following in their environment and have all managed
by one master server:
◆ A master server running NetBackup 6.0. (The EMM server must also run
NetBackup 6.0.)
◆ All remaining master servers must also run NetBackup 6.0.
◆ A media server running NetBackup 6.0.
◆ Any other media server running either NetBackup 5.1 GA, 5.1 MPx, or 5.0 MP4 or
greater.
◆ Any number of clients running NetBackup 5.x or 6.0.
For more information about upgrading NetBackup from 5.x to 6.0, refer to the NetBackup
Installation Guide for your operating system. The upgrade sections contain various
mixed-version scenarios and how they must be upgraded to this version of NetBackup.
However, the general rules for a mixed-server environment are as follows:
◆ In a mixed environment such as this, the master server must be running the highest
version of NetBackup in use in that configuration.
◆ A master server can inter-operate with a media server that is running a level of
NetBackup that is one major release lower.
◆ Backlevel servers (those servers that are running NetBackup 5.x) cannot be used as a
Volume Database Host, global device database host, or vmd/DA host. These
capabilities are all on the EMM database host as part of the EMM server. Always
upgrade NetBackup 5.x servers performing these roles first. Only then can media
servers be upgraded and finally client machines. Again, refer to the NetBackup
Installation Guides for detailed upgrade information.
◆ Media server backwards compatibility extends only one major release behind the
master server. No media server may have a numerically higher version than the
master server. (The media servers must be running equal or lower levels of
NetBackup.)
◆ Client backwards compatibility extends only one major release behind the master
server. This includes database clients.
◆ Applying a NetBackup maintenance pack (MP) does not alter the version of the
release for NetBackup 5.1. For example, NetBackup 5.1 GA and NetBackup 5.1
Maintenance Pack 1 (MP1) are both treated as NetBackup 5.1 for compatibility
purposes and have no co-existence issues.
However, for NetBackup 5.0 media servers, this does not apply. In NetBackup
versions 5.0 GA, MP1, MP2, and MP3, if the media server and the client were the same
machine, you would not be able to restore on the client. A fix was introduced for this
issue in NetBackup version 5.0 MP4. Therefore, any media server that you want to
coexist in a NetBackup 6.0 configuration must be at a version level of 5.0 MP4 or
greater. (For more information about this fix, please refer to TechNote 278153, on the
VERITAS support web site. The following is a link to this document,
http://support.veritas.com/docs/278153.)
For customers who are currently at 5.0GA and do not want to upgrade to5.0 MP4, you
can perform the following work-around on your machine.
Create the file /usr/openv/netbackup/NOSHM on a UNIX machine and
install_path\NetBackup\NOSHM on a Windows machine.
By creating this touchfile, you will likely see a performance degradation for backups
and restores.
VERITAS recommends that you remove this file after you have upgrades to 5.0 MP4
or greater or upgrade to NetBackup 6.0.
◆ All components (master server, media server, client, console, and agent) on an
individual system must be at the same version.
◆ Backup images created under an older version of NetBackup will always be
recoverable with a newer version of NetBackup.
Cluster Limitations
1. When installing NetBackup in a Sun Cluster all nodes must have NetBackup installed.
2. The VERITAS Private Branch Exchange (PBX) software cannot be installed in a UNIX
cluster using the push installation method provided with the infrastructure Core
Services installer (installics). You must run the installics installer on every
node in the cluster on which you are installing NetBackup.
4. During cluster_config you must specify the same name used during the
installation of NetBackup (install_bp) as the server name. If the name specified to
install was fully qualified, the name supplied to cluster_config must also be fully
qualified and vise versa. This is because install_bp populates the databases
server.conf file and cluster_config passes the -server option to create_nbdb. If
the names differ, the creation of the Sybase ASA database for NetBackup will fail. If
you have encountered this issue, perform the following steps to recover.
c. Modify /usr/var/global/server.conf,
/usr/openv/netbackup/bin/cluster/NBU_RSP, and
/usr/openv/netbackup/bp.conf and make sure all of the references to the
virtual name are consistent.
f. Run /usr/openv/netbackup/bin/nbnos
g. Run /usr/openv/netbackup/bin/nbemm
h. Run /usr/openv/volmgr/bin/tpext
If this does not work, call support for more information about the following error:
EMM interface initialization failed, status = 77
k. If you have any add-ons to install, please following the procedures documented
in the NetBackup High Availablility System Administrator’s Guide.
5. Due to Microsoft Cluster polling of shared disks, restore of Flashbackup images will
fail. Therefore, VERITAS recommends that you not use Flashbackup for shared disks
in a Microsoft Cluster.
2. NetBackup Java cannot display Chinese characters on Linux Red Hat AS4.0 and
AS3.0, and on HP-UX IA64, 11.23 systems because of the JRE1.4(.2_04) font
configuration. This problem will be resolved when NetBackup Java begins using a
later version of JRE.
3. NetBackup menu user interfaces (MUI) can not input and modify multi-byte
characters and are not localized to any language. The MUI includes bp, bpadm,
tpconfig menu, vmadm, and vltadm.
5. For Windows installations, the install path must not contain multi-byte characters.
8. File names and other text are not displayed correctly across different platforms, such
as, Windows and UNIX. This occurs because the same locale on different operating
systems does not use the same encoding for the various locales with default character
sets. These character sets may have characters beyond the ASCII-defined character
set, such as, on Windows the encoding for the Latin-1 character set is different than it
is on Solaris. The Solaris Latin-1 character set contains additional characters. The
following examples demonstrate how encoding for various locales can effect how
multi-byte file path characters or the file names of backups are displayed. These are
operating system limitations and cannot be resolved through NetBackup.
◆ For multi-byte file path characters, characters that are specified on one locale may
not display correctly in another locale. For instance, if a Solaris client specifies a
file name in the ja locale (EUC) and then attempts to display the file with the
Backup, Archive and Restore GUI on HP server in the ja_JP.SJIS locale (SJIS), the
file name does not display correctly.
◆ A backup of file names with Portuguese characters resulted in the characters
looking corrupt when displayed on a UNIX Master Server, however, these files
displayed correctly when they were backed up from a Windows client. Once
again, varying encoding being used on different platforms caused this displaying
issue.
10. At the time of this release, the following database agents have restricted support of
localized environments:
◆ NetBackup for DB2
◆ NetBackup for Informix
◆ NetBackup for Oracle
◆ NetBackup for SAP
◆ NetBackup for SharePoint
◆ NetBackup for SQL Server with Advanced Client
◆ NetBackup for Sybase ASE
When using any of these agents, the use of localized characters (for example,
non-ASCII, multi-byte characters) is not supported in:
◆ Any database object names, like instances, databases, tablespaces, filegroups,
datafiles, portals, and so forth.
◆ Path names of database files, directories, transaction logs, or other database
storage locations
◆ Path names specified in the policy “Backup Selections”, like script, template, or
batch file locations
This applies to all supported platforms, including the use of previous versions of
those NetBackup database agents with NetBackup 6.0 servers.
11. The following NetBackup user-defined strings must not contain Multi-byte
characters:
◆ Host Name (Master Server, Media Server, EMM server, Volume Database Host,
Media Host, Client)
◆ Policy Name
◆ Policy KEYWORD (Windows Only)
◆ Backup, Archive and Restore KEYWORD (Windows Only)
◆ Storage Unit Name
◆ Storage Unit Disk pathname (Windows Only)
◆ Robot Name
◆ Device Name
◆ Schedule Name
◆ Media ID
12. Viewing log files with non-ASCII characters in some text editors may make the
non-ASCII characters appear as gibberish. While there are some editors that will
display the log file correctly, you can use the Backup, Archive and Restore GUI (Task
Progress tab on UNIX, or the View Status display on Windows) to view these files.
13. In the Frozen Image Client Configuration dialog box, the following strings are being
displayed in English:
Keep frozen image after backup
Resynchronize mirror in background
14. An issue exists if you attempt to install an English version of NetBackup into a
Chinese-clustered system. If the clustered system's public network connection name
contains any Chinese characters, then the wrong cluster information is entered into
Registry, causing the cluster configuration to subsequently fail.
Caution An NBU_Security Admin member can delete all NBAC groups, causing all
NBAC-enabled operations to stop and making re-initialization of NBAC
security objects necessary.
1. In NetBackup 6.0, limitations with the new connection protocols exist with Enhanced
Authorization using a non-root (UNIX) or non-Administrator (Windows) user to
administer NetBackup. These limitations also apply to root, or an Administrator, on
hosts not in the NetBackup server list. Most notably, administration of storage units,
devices, and media will be affected.
VERITAS recommends NetBackup Access Control (NBAC) for authentication and
authorization for non-root and non-Administrator NetBackup administration.
4. Do not modify settings in the Access Control properties pane in Host Properties.
If you have erroneously modified settings in the Access Control pane and have been
denied access to the NetBackup Administration Console, you can perform the
appropriate procedure below to regain access to the NetBackup Administration
Console.
◆ For UNIX-based systems, you must modify a line in the bp.conf file on the master
server. Perform the following steps:
◆ Open the file /usr/openv/netbackup/bp.conf for editing.
◆ In the bp.conf file you will see one of the following lines:
USE_VXSS = AUTOMATIC
or
USE_VXSS = REQUIRED
Change whichever line you have in the bp.conf file to the following:
USE_VXSS = PROHIBITED
◆ Save this file.
◆ Stop and then start the NetBackup daemons.
◆ For Windows-based systems, you must edit the registry to modify a specific
resource in the Registry Keys.
◆ Select Start > Run...
◆ Enter regedt32 (for Windows 2000 and greater systems) or regedit (for
Windows NT systems) in the Run window and select OK.
◆ Select HKEY_Local_Machine\Software\VERITAS\NetBackup\
CurrentVersion\Config
◆ Double-click USE_VXSS
◆ Change the value to PROHIBITED and click OK.
◆ Exit the Registry.
◆ Stop and then start the NetBackup services.
5. On Windows machines, VERITAS recommends that the user re-boots (the install
target) after installing the Authentication (AT) server software.
6. On Windows machines, VERITAS recommends that the user re-boots (the install
target) after installing the Authorization (AZ) server software.
7. The NBAC security credential generated upon login using either the NetBackup-Java
Administration Console or NetBackup Console (MFC) GUI’s, will by default, expire
twenty-four (24) hours after the login has been accomplished, thus disabling further
operations. To avoid this issue, use the VxSS client interface (CLI) or the VERITAS
Security Services Java GUI to increase the default expiration time.
10. The use of NetBackup Database Agents in conjunction with NBAC was not
implemented in the NetBackup 6.0 release.
11. The NBU_Operator group alone does not possess the necessary permissions to
inventory a robot.
12. Currently, anyone who belongs to the NetBackup Administration group has the
ability to modify the Host Properties, consequently they can change or disable NBAC.
13. On HP-UX, the debug log may show a NON FATAL Unresolved external symbol.
This is expected behavior that is caused by mixing C and C++ code on an HP-UX
system. This issue does not effect normal NetBackup operation.
14. When adding someone who is not an administrator to the NBU_Admin group, you
must set the permissions of the following directory to word readable/writable.
/user/openv/netbackup/logs/user_ops (UNIX)
<install path>\NetBackup\logs\user_ops (Windows)
Set the permissions of this directory to [0777] for UNIX systems and FULL CONTROL
for the user group on a Windows system.
2. For Linux and AIX client data configured in VxVM volumes on supported EMC or
Hitachi disk arrays, when using the TimeFinder snapshot method on EMC or the
ShadowImage method on Hitachi, the VxVM names for the disks underlying the
volumes must not use the enclosure based naming scheme. Use of enclosure based
naming will cause the backup to fail.
Instead, use the older VxVM disk naming scheme (called hdisk# naming format on AIX
and old format or OS native naming format on Linux). Use the vxdiskadm utility to
change the disk naming scheme.
3. Snapshots of Windows 2003 clients using Advanced Client with the VSS snapshot
method will fail with a job status of 156 if the Volume Shadow Copy Service is not
started. Since the service starts only on an as-needed basis and NetBackup does not
currently have a means of requesting it, the service may not always be running when
an Advanced Client VSS snapshot backup is attempted. If the service is not running, it
must be manually started through the Windows Services Administrative Tool before
the Advanced Client VSS snapshot backup can succeed. Windows Open File Backup
using the VSS snapshot method is not affected by this problem.
4. The VxVM, FlashSnap, and VVR snapshot methods do not support VxVM Volume
Sets when used in conjunction with the Instant Recovery feature. The backup fails and
the following message appears in the client's bpfis log:
<32> onlfi_freeze_fim_fs: FTL - pfi_get_fmr_att:
FastResync not enabled for volume: fs_vset
6. For Point in Time Rollback from a snapshot on the SnapVault secondary system, the
entire subvolume (qtree) is restored. Note that the restore always goes to a new
subvolume on the primary host, never overwriting an existing subvolume. The
format of the new subvolume name is the following:
mountpointname_restore.time_stamp
For example: subvol1_restore.2005.05.19.10h49m04s
To use the restored, new subvolume, you can do one of the following (see the
documentation provided by your NAS vendor for command details):
Unmount the original subvolume on the client, then on the filer rename the original
subvolume and rename the new subvolume with the name of the original, making
sure the new subvolume is still shared, and then mount the new subvolume on the
client.
Or
Unmount the original subvolume on the client, then on the filer share the new
subvolume, retaining its current name, and then mount the new subvolume on the
client.
a. Go to http://support.veritas.com.
b. In the search field in the upper, right corner of the web page, enter the following:
Advanced client configuration
c. Click “>” search icon. The guide will appear when the web page refreshes and
shows the results of the search.
10. Due to limitations in volume name size in VxVM for Windows 2000 3.1, for backup
volumes targeted by the VxVM and FlashSnap snapshot methods, VxVM volume
names should be limited to 12 characters in order to guarantee that snapshot volume
names, which are derived from parent volume names, are unique.
11. The space optimize snapshot method in VxVM 4.0 does not work with alternate client
backups.
12. Advanced Client VxVM and FlashSnap snapshot methods on Windows 2000
platforms are supported by VERITAS Volume Manager for Windows 2000. To
successfully employ these snapshot methods, VERITAS Volume Manager for
Windows 2000 version 3.1 or greater must be installed.
For VERITAS Volume Manager for Windows 2000 version 3.1, Service Pack SP1
should be installed, hot fix HF03 should be installed, and the Volume Manager build
level should be 5.31.56.0 or greater. After installing VERITAS Volume Manager for
Windows 2000, the build level is reflected in the Windows Registry under the subkey
HKEY_LOCAL_MACHINE\SOFTWARE\VERITAS\Volume Manager 3.1\Version.
13. The system command, fsck, crashes when the file system device name is comprised
of more than 63 characters for VxVS 3.x, VxVS 4.0, and VxVS 4.1. Because of this issue,
Advanced Client FlashSnap backups fail if the file system device names that are used
on the alternate client contain more than 63 characters. To avoid this issue, the user
must install a VxVS patch that includes the Itool Incident 154093, and is labeled, “fsck
coredumped in recover_qlog() due to a long device name”.
◆ NetBackup 6.0 master server, NetBackup 5.x source client, and NetBackup 5.x
(different host than source) or NetBackup 6.0 destination client, Instant Recovery
is not supported.
◆ NetBackup 6.0 master server, NetBackup 6.0 source client, and NetBackup 5.x
destination client, Instant Recovery is not supported.
bpadm Interface
1. When displaying menus or other output, the data does not appear correct - usually
missing the first character. This issue seems to be caused by the interaction between
the terminal type implemented by Windows telnet and the remote UNIX terminal
output. Changing the default terminal type in the Windows client from 'ansi' to 'vt100'
seems to correct this display problem.
To avoid this issue, you can change the terminal type implemented by the client and
then connect to the remote system (where it will detect that terminal type) or change
the terminal type after you have logged in.
To change the terminal type implemented by the client, you can do the following:
a. On Windows, start the telnet command with a command line option that specifies
the terminal type.
FlashBackup
1. FlashBackup-Windows policies do not back up Windows system-protected files (the
System State, such as the Registry and Active Directory).
2. FlashBackup-Windows policies do not support the backup of file names that are in the
8-character plus 3-character format. The 8.3 character format is retained, but the name
of the restored file may be different from the original name.
5. When restoring a reparse point for a FlashBackup Windows policy, the restore will not
replace the reparse point or data files under the reparse point unless the reparse point
or files are first deleted. Examples of this problem would be:
If a backup is done of a Mount Point Directory, and then the volume is unmounted, a
restore of the directory on which the volume was mounted will not actually remount
the volume (for example, restore of the reparse point is not actually done). If, the
directory for the mount is first deleted, then the restore of the reparse point will result
in the directory being restored with the volume mounted.
If a file is restored that exists on a Mount Point Directory, the file will not actually be
restored if it still exists. It is necessary to first delete any files that are to be restored.
This problem exists for all reparse points including RSM.
2. The NearStore storage unit does not support backups based on the following policy
configurations:
◆ Multistreamed backups using NetBackup for Microsoft SQL Server.
◆ Multistreamed backups using NetBackup for Sybase.
◆ Multistreamed NetBackup for Oracle Proxy Block Level Incremental (BLI)
backups.
3. The following Data Base agents: Oracle, DB2, Informix, and SAP are not permitted to
run as multiple concurrent jobs if these agents are run in multi-streaming mode. The
NetBackup job scheduler can only run one job stream at a time. It will queue
subsequent job streams.
2. BMR will delay its support of the BMR restore for Solaris NetBackup media servers
that back up their own data until after the release of NetBackup 6.0. (In the meantime,
you can protect a Solaris NetBackup media server as a regular BMR client.)
3. AIX clients require at least 256 MB of RAM to support BMR restore operations.
4. HP-UX systems can host only one shared resource tree at a time. You can create more
than one SRT at OS levels at or below the OS level of the boot server, but the newly
created SRT will invalidate any existing SRTs at other OS levels.
5. To function as a boot server, an HP-UX system must contain at least one volume
group that is controlled by HP-UX Logical Volume Manager (LVM). BMR does not
support multiple stripe widths per volume when a Solaris system that uses Solaris
Volume Manager is restored. For Solaris Volume Manager disk stripes, all stripes on a
volume on the restored system will be the width of the first stripe on that volume on
the protected system. If a Windows BMR boot server is also an active directory server,
restores may fail because DOS cannot copy files from the shared resource table (SRT).
To avoid this issue, do the following:
In Administrative Tools > Local Security Policy > Local Policies > Security Options,
change the following two options:
◆ "Microsoft network server: Digitally sign communications (always)" to "Disable"
◆ "Microsoft network server: Digitally sign communications (if client agrees)" to
"Enable"
After changing these settings, reboot the system.
6. BMR point in time restore does not work when the policy that backs up the client has
compression enabled and the BMR master server resides on a system that is different
than the client (for example, Windows client and UNIX master server). (BMR point in
time restore is different than a regular NetBackup point in time restore.)
7. You cannot use the NetBackup Administration Console to manage Bare Metal Restore
if you use the Change Server option to change to a NetBackup master server that is
licensed for BMR. To use the Administration Console to manage Bare Metal Restore,
you must:
◆ Log into the NetBackup master server that is licensed for BMR each time you start
the Administration Console (UNIX and Linux systems).
◆ Start the NetBackup Administration Console on the NetBackup master server that
is licensed for BMR (Windows systems).
After you log into or start the Administration Console on a NetBackup master server
that is licensed for BMR, you can change to another server and then change back to
the server on which you logged in or started and continue to manage BMR.
8. Restores from user initiated backups are not supported. BMR requires true image
restore information, and true image restore information is not collected during a user
initiated backup.
14. NetBackup database agents are not supported with BMR. BMR does not restore
databases that require use of database agents. After BMR restores the operating
system, the user can restore these databases manually.
15. Block-level incremental, available with VERITAS Oracle Edition, is not supported
with BMR.
NetBackup Catalog
1. Manually initiating an offline backup of the NetBackup catalog can fail if backups are
in progress.
Backing up the NetBackup catalog can fail if it is manually initiated and backups are
in progress. The workaround for this is to initiate these cold backups when no other
NetBackup activity is occurring, or to schedule the cold backup of the catalog to occur
automatically. If catalog backups are required while NetBackup activities are
occurring, then you should use a hot catalog backup.
2. NetBackup 6.0 offers a new method to perform catalog backups of a Master server.
The System Administrators guide goes into details on how to configure the new Hot
Catalog backup method. However it doesn't mention that the NetBackup Jobs
database will not be included in the Hot Catalog backup, nor will it be restored during
a Hot Catalog recovery. After a Hot Catalog Recovery of the Master server the Jobs
database will be reset so all new jobs begin at JobID 1. This is the correct behavior for
NetBackup 6.0 due to the design of the Hot Catalog backup recovery. For more detail
on this issue, search for TechNote number 278610 on the VERITAS support web site.
3. Catalog backup jobs started with -host option to run on the media servers cannot be
killed from the Activity Monitor. If the catalog backup is started on a media server by
a vault job, canceling the vault job from the Activity Monitor will not terminate the
catalog backup.
DB2 Agent
1. The 6.0 NetBackup for DB2 database agent enables you to restore a DB2 database to a
different/same instance on a different client or to a different/same instance on the
database's original client. Within an instance, all databases must have unique names.
For more information on performing alternate restores, see the NetBackup for DB2
System Administrator's Guide for UNIX or the NetBackup for DB2 System Administrator's
Guide for Windows.
Note If you are restoring a database to an instance that already has an existing
database named identically to the one you are restoring, the existing database
will be overwritten. The IBM DB2 restore command enables you to rename the
database at the time of the restore to avoid overwriting the original database.
2. The NetBackup Client GUI does note recognize DB2 version 7, 64-bit databases. DB2
version 7 does not support 32-bit application access of 64-bit databases.
3. For 64-bit, DB2 databases on AIX, you must clear LIBPATH before executing
bpdb2proxy or bpdbsbdb2 from the command line.
2. You must perform an advanced client backup after making any configuration changes
to the database.
3. There is no API available for exclusive database access in DB2 version 7. Users have to
be manually disconnected before performing an advanced client restore.
Informix Agent
1. You may encounter issues with the onsmsync utility with various versions of
Informix. Please contact Informix Support for onsmsync-related patches.
2. Master initiated Lotus Notes, Point-In-Time restores may fail with a status 12. If the
<install_path> of the NetBackup master server and the NetBackup client is
different, the automatic restore of Lotus transaction log extents during the recovery of
the Lotus database will not be able to locate the path to the progress log and will fail.
3. The use of exclude lists for Lotus backups on NetBackup Windows clients has been
disabled.
2. Compression is not supported when using NetBackup for Exchange. The compression
attribute should not be enabled on Exchange policies. This capability is planned for a
future release.
4. When an Exchange backup is launched from the NetBackup Client GUI and uses the
Advanced Client offhost backup capability, the progress log window does not display
the usual progress messages evident when a scheduled backup is executed. The
backup operation is not impacted by this lack of progress logging. If detailed progress
is desired, use the NetBackup Administrator’s GUI to launch a Manual Backup
operation on an Exchange policy. See the Testing Configurations Settings section in
the NetBackup for Exchange System Administrator’s Guide for instructions regarding
a manual backup operation.
5. When restoring data in an Exchange cluster environment, one must set the destination
client value to be the virtual server name. If restoring an Exchange database using a
NetBackup client-only installation on a cluster, it may not be possible to change the
destination client value to match the virtual server name. In that case, use a
NetBackup Client GUI on a NetBackup server to change the destination client value
to the virtual server name.
1. When using the NetBackup Client Backup interface or the bpbackup command to
initiate an Exchange offhost alternate client backup, the backup will fail and report a
status of 200. Instead, use the NetBackup Administration Console to initiate a manual
backup for that Exchange policy.
Oracle Agent
1. The NetBackup Client interface is unable to connect to Oracle 8.1.6 databases that are
on HP-UX 11.00 systems.
NetBackup explicitly loads Oracle database shared libraries during program
execution. For systems running HP-UX 11.00, this involves a call to the shl_load()
method to explicitly load the shared library. The shl_load() method does not support
the use of a library that contains thread local storage (TLS). The Oracle 8.1.6 client
library, libclntsh.sl, is dependent upon the Java Runtime Library, libjava.sl, (a runtime
library that is part of JRE 1.1.8 and supplied with Oracle 8.1.6). The Java Runtime
Library libjava.sl contains TLS references that prevent the Oracle library from being
loaded. This problem is specific to Oracle 8.1.6.
To avoid this issue and use Oracle 8.1.6 with HP-UX 11.00, you must remove the
dependency of the Oracle 8.1.6 client library on the libjava.sl Java Runtime Library by
performing the following procedure:
b. Search for, and remove, references to libjava.sl from the ldflags file in
$ORACLE_HOME/lib.
d. Use the following command to confirm that the libjava.sl link has been removed
from libclntsh.sl file:
chatr linclntsh.sl
The chatr command lists all of the libraries that the specified one links to. The
runtime library, libjava.sl, should not appear on the list.
SAP Agent
1. SAPDB backups fail if the SAPDB database is installed in C:\Program Files or the
directory name contains spaces in it. The default path for SAPDB installation is
C:\Program Files. If the default location is chosen for installation, then backups
will not work.
Users should install SAPDB under a directory that does not contain spaces in the
directory name.
2. Backup of files that contain Japanese or Chinese characters in their names will not
work. This is a regression from NetBackup 5.1 that is scheduled to be fixed in the first
NetBackup 6.0 Maintenance Pack release.
SharePoint Agent
1. When attempting to open multiple NetBackup Client Backup interfaces to browse
SharePoint resources, the interfaces unexpectedly terminate. To avoid this issue, use a
single client interface to browse SharePoint resources for backup.
2. When the SharePoint Search Index service is running, it sometimes contends with
NetBackup for access to the portal index database. This can result in a failed backup
or restore operation of that database. VERITAS recommends that you try the backup
or restore repeatedly until successful.
5. In a SharePoint Farm that contains multiple SQL Server hosts, attempting to restore a
SharePoint Portal database to SQL Server host B may fail when the SharePoint Farm
Configuration database resides on SQL Server host A.
To avoid this issue, change the NetBackup Client Service account, on the SQL Server
host B (where the portal database(s) resides, to a domain account other than
LOCALSYSTEM. This account must have network access to the SQL server host A
(where the configuration database resides).
Note There are special batch files that run during a backup that allow the backup to
be written to a mapped drive. These batch files map the drive in for the bpbkar
process and then un-map the drive after the process is done.
2. If the NetBackup installation directory or the VxUL logging directory for NetBackup
fills up, such as with logging files, NetBackup may become unresponsive. For
example, NetBackup jobs may remain queued for long periods, even though all
NetBackup processes and services are running. For more information on how to avoid
this issue, refer to, “Resolving Problems Related to a Full Disk”, in the NetBackup
Troubleshooting Guide for UNIX, Windows, and Linux.
3. If a customer uses Windows FAT file systems for their disk storage units (DSUs) on
media-servers, then the size of an image fragment is limited to 4GB, even though
NetBackup defaults to something much greater. The result is that this limit causes
backups to fail with the message, EC_diskfull (129). The real issue is that the
fragment is too large, and Windows reports that error as the disk is full. VERITAS
recommends that users configure DSUs on FAT file systems to have a maximum
fragment size of 4000MB instead of using the NetBackup default.
NetBackup Interfaces
c. Click New.
d. Enter a path and click on a different path that already exists. (You can enter
additional paths by repeating steps c and d until all new paths have been
entered.)
e. Click OK.
f. Go back to the NetBackup Catalog Backup window and verify that the correct
paths are present.
DEBUG_PROPERTIES=$DEBUG_PROPERTIES -Xms$INITIAL_MEMORY
-Xmx$MAX_MEMORY vrts.nbe.
AdminConsole >>$LOG_FILE 2>&1
Another exception has been detected while handling the previous error. The
following is the failed backup information about this error :
ERROR REPORT FILE = (N/A)
PC = 0x0x554d0084
SIGNAL = 11
FUNCTION NAME = (N/A)
OFFSET = 0x554D0084
LIBRARY NAME = (N/A)
Please check ERROR REPORT FILE for further information, if
there is any.
Good bye.
Use the information in the following list to avoid this issue in the future:
◆ Install another processor (the issue does not present itself on dual processor
machines).
◆ Reconfigure the kernel to be non-SMP.
◆ Edit the file /etc/grub.conf and add the following lines to disable the
no-execute function.
(NX) page protection
nexec=off
◆ Reinstall update 2 of this version of the operating system.
◆ Run the NetBackup-Java Administration Console on another host. If you need to
administer this type of configured host, specify it in the login dialog of the
NetBackup-Java Administration Console (running on another host).
4. A problem exists from the way Tru64 reports the timezone to the Java Timezone class
causing the time in the Tasks Performed window to be one hour off from the actual
time. To correct this problem, the user can enter the correct settings in to the Adjust
Application Timezone dialog, and save this as the default. A persistent file is saved in
the user’s directory, and is used on subsequent program operations to enable the
NetBackup-Java Administration Console to use the correct default timezone.
5. Paint problems in tables with large sets of data viewing with Exceed
There is a Java problem that causes table painting problems when the NetBackup-Java
Administration Console is viewed through Exceed for large data sets, for example, the
job table in the Activity Monitor. This bug may cause the interface to deteriorate to the
extent where it is unusable. Resizing the window causes the NetBackup-Java
Administration Console to paint correctly. Another workaround involves adding the
following to the Java command in the jnbSA or jbpSA startup scripts.
-Dsun.java2d.pmoffscreen=false
This workaround will greatly degrade the performance of the GUI.
Other options to consider are:
◆ Install the NetBackup-Java Windows Display Console and use it on the Windows
desktop. You can administer any NetBackup server by specifying the server in the
login dialog.
◆ Install and use the Windows Remote Administration Station to remotely
administer your UNIX NetBackup servers.
6. If a non-modal dialog is visible and behind the parent frame and subsequently a
modal dialog is made visible, Java moves the original non-modal dialog in front of the
parent frame, but behind the newest dialog (a modal dialog requires a response and
blocks input to all other dialogs for the application until the response is received).
When the modal dialog is hidden, the non-modal dialog is moved back to its original
position behind the parent frame.
7. If you enter text into a dialog’s text field while either NetBackup Java is busy
performing other operations, or while your system is busy, the characters you enter
may be transposed in order when displayed in the text field. For example, if you have
started the Backup Policy Management Install Client Software operation and it is
currently busy installing software, if you then try to perform other operations in
Backup Policy Management which require entering text, the characters you enter into
the text field may be displayed in the field out of order. You may also notice that
response time is slow.
8. On Tru64 platforms, the menu and button mnemonics do not work properly.
9. The mouse control may lock up when deleting a large number of jobs.
If the mouse seems to lock up when deleting a large number of jobs from Activity
Monitor, be patient, it may take a while for the operation to complete. The mouse will
work eventually in this situation.
11. In the NetBackup Java Administration Console, the NetBackup patch version is not
displayed for hosts in the Media and Device Management > Devices > Hosts view
and the Bare Metal Restore Management > Hosts > Bare Metal Restore Clients view.
(The patch version is displayed in the Administration Console title bar for the host
being administered and in the hosts listed in NetBackup Management > Host
Properties views (Master Servers, Media Servers, and Clients).)
12. Remote display of the NetBackup Java Administration Console in multi-byte locale
environments is not supported.
Media
The following items are known limitations with features associated with the Media and
Device Management portion of the NetBackup Administration Console.
1. It is recommended that you move only one volume or volume group at a time.
1. Pending mount requests to "DOWN" robots for storage unit requests: NetBackup will
not pend mount requests to "DOWN" robots (robots whose drives are in AVR mode)
for storage unit related requests. NetBackup will try and pick another storage unit
that has a robot that is not "DOWN". If NetBackup cannot find a storage unit for the
job, NetBackup will queue the job and a "Queued" state will be displayed in the
Activity Monitor (refer to the section on job queuing for the reasons why a job can be
in a "Queued" state). This is different than pre NetBackup 6.0, where NetBackup
would pend mount requests to "DOWN" robots, even when the job could have run on
other robots. If for some reason, the robot's hardware is not functional, and the robot
needs to be operated in a manual mode (where NetBackup will not be loading media
in the drives of the robot), the robot can be made to operate in the "pend if robot
down" (PIRD) mode. The PIRD mode is set on a robot and media server connection,
and tells NetBackup to pend mount requests to the robot even if it is "DOWN" (refer
to the tpconfig usage for details on how to set the PIRD state). These pending
requests will appear in the Device Monitor, and it will be possible to manually assign
these mount requests to drives. NetBackup will not autoassign, and it will not be
possible to manually assign a mount request targeted to one storage unit to drives of
another storage unit. This is different than pre-NetBackup 6.0 where any mount
request could be autoassigned and manually assigned to any compatible drive.
Note The above section applies only to storage unit specific mount requests
(including backups, write side of duplicates, and so on). Media specific mount
requests (including restores, imports, verify, tpreq mount requests, and so on)
are not storage unit related and will pend even if the robot is "DOWN"
(irrespective of the PIRD state). Such mount requests can be autoassigned by
NetBackup and manually assigned to any compatible drive.
2. After media servers have been upgraded to NetBackup 6.0 and their Remote NDMP
drives converted to Media Manager drives, the nbemmcmd -listhosts command
will continue to show an NDMP host for each of the media servers on which Remote
NDMP was configured. There are no known issues with having this unnecessary
NDMP host in the list. NetBackup will down all the paths on all the media servers for
a drive if the TapeAlert information that the drive reports indicates that the drive
needs to be downed.
3. It is not possible to mount media that are not configured in EMM using NetBackup.
Prior to NetBackup 6.0, it was possible to mount media that were not configured in
the Media Manager volume database.
4. MDS will not select media with conflicts for write jobs (including backups, write side
of duplicates, tpreq mount requests for write, and so on). Refer to media conflicts
section for details on media conflicts and how to remove them.
5. MDS will unload specific drives on media unload delay on NetBackup 6.0 media
servers when necessary.
lead to the device being unusable (for example, the tape drive may be downed). This
can happen when ltid performs automatic path correction, or when the run-time
Plug-n-Play code (Windows only) performs serial number checks. In such cases, you
can do the following:
a. Update the serial number or reconfigure the device causing the new serial
number to be stored in the EMM database.
2. On UNIX, ejecting tapes through the media access port of the ATL P3000 or Sun
StorEdge L11000 robotic library can lead to the error unable to sense robotic
device (applies only to NetBackup Enterprise Server).
The workaround is to wait until the robot has finished initializing again (checking for
media) after the media access port door has been closed, before responding (in the
NetBackup Administration Console) that the tapes have been removed. During
testing, the library initialization period for an empty port lasted for 2 minutes, 10
seconds.
3. The mount/unmount feature makes a system call to run its callout scripts.The system
call returned the wrong value on a few platforms, when using the
drive_unmount_notify script; causing NetBackup to down drive(s). If you
encounter this type of issue with this feature and you are not using the scripts, then
the easiest solution is to rename the scripts or move them from the
/usr/openv/volmgr/bin directory.
To fix this issue, the shell script drive_unmount_notify has been altered and
placed in /usr/openv/volmgr/bin/goodies/ directory along with the
drive_mount_notify script. If you have changed this script in any way and would
like to retrieve your changes, you must merge your original script with the newly
installed script in the /usr/openv/volmgr/bin/goodies directory. After you
complete the merge, place the script in /usr/openv/volmgr/bin directory.
4. Some libraries implement different functionality for their media access ports. For
example, some libraries have front-panel inject/eject features that conflict with
NetBackup's use of the media access port (for example, Spectra Logic Bullfrog). Other
libraries require front-panel interaction when using the media access port (such as,
Spectra Logic Gator). Make sure you read each library’s operator manual thoroughly
to gain an understanding of its media access port functionality, or libraries such as
these may not work properly with NetBackup’s inject/eject features. The Breece Hill
Q47 (TLD) is not supported with NetBackup multiple eject and inject.
6. In environments where network traffic or machine load is high, restarting the Media
Management device daemon (UNIX) or NetBackup Device Manager service (on
supported Windows servers) can result in individual robotic daemons not being
active. This situation occurs when the daemon/service start occurs too soon after the
shutdown, such that the full shutdown of all daemons/services does not complete
before the daemons/services are started. This in turn leads to a newly-started
daemon/service being terminated by a delayed stop signal.
How do you know if this situation occurs? In the Volume Configuration wizard, at the
robot inventory stage, the inventory fails with a message similar to, cannot
connect to tldcd on host <hostname>. If you see this message, VERITAS
recommends that you stop the NetBackup Media Management device daemon (on
UNIX) or Device Manager service (on supported Windows servers), wait at least 20 to
30 seconds, and then restart it.
Note For detailed information regarding the use of vmadm or tpformat, refer to the
NetBackup Media Manager System Administrator’s Guide.
8. Robot inventory may add media with a media type that is not compatible with the
drive type (applies only to NetBackup Enterprise Server).
This limitation only applies to TLD or TL8 robots that have no drives configured on
the robot control host or are using a multimedia robot (for example, DLT7000 and
DLT8000 in the same library).
Example 1: If your TLD robot has DLT2 drives on a host that is not the robot control
host and the robot control host has no configured drives in that robot then you must
specify DLT2 as the media type on the inventory robot update.
Example 2: If you are using a multimedia robot, you must use barcode rules (see the
NetBackup Media Manager System Administrator’s Guide).
If media are configured with the wrong media type, the media will be unmountable,
possibly causing backups to fail.
9. Inventory during installation results in duplicate media IDs when a robot contains
media with the barcodes that are more than 6 characters and the last 6 characters are
not unique. This is a known operational issue, and as a workaround, the user should
choose to use the standard inventory, which includes Media ID Generation Rules, by
clicking on the Advanced button in the wizard.
10. When using NetBackup media management on Windows, a robot may be selected for
robotic inventory when listed with the wrong robot control host, leading to a robotic
inventory that fails because it tried connecting to a robotic control daemon or process
on the wrong host.
11. The VERITAS implementation using SCSI reserve/release has the following
limitations:
◆ SCSI reserve/release is not applicable for NDMP configurations (no reserve
command is available).
◆ Third party copy configurations must be configured correctly. To retain
reservations of a tape device when doing a third-party copy backup, refer to the
description of the mover.conf file in the NetBackup Advanced Client System
Administrator’s Guide for UNIX and Windows.
12. Cluster environments or multi-path environments with dynamic path sharing (Tru64
for example) will cause backup and restore failures if the path changes during these
operations. If path sharing cannot be eliminated, then the NetBackup SCSI
reserve/release function must be disabled.
15. A long erase should be avoided on SGI IRIX servers, because system performance
may become degraded during this operation. If you want to test a long erase,
VERITAS recommends that you perform a test of this function during an idle time on
the server.
Note The SGI IRIX media server is not supported in a pure NetBackup 6.0
environment.
16. On AIX systems, that are using ovpass to control robots, the robot will appear to be
missing if the robot control daemon has the robot open at the time a remote system is
running auto-discovery. The remote system takes control away from the AIX system
leaving the EMM database inconsistent.
To avoid this issue, you should perform your configurations when no backups are
running or about to run.
17. Device discovery attempts may cause system reboots when Tru64 5.1b systems are
scanned, if non-current operating system patches and HBA driver firmware are used.
This can cause the NetBackup Device Configuration Wizard running on a different
host to hang. Users of Tru64 5.1b systems should keep current with the latest
operating system patch and HBA firmware levels.
Note For NetBackup ACS Robot Auto Configuration to work, the ACSLS Server must be
able to report the serial numbers of the robotic tape drives. ACSLS 6.1 and later
releases report drive serial numbers by using the display command. However,
ACSLS can only report serial numbers when the drives report their serial numbers
to the library and the library reports the drive serial numbers to ACSLS.
The following drives and libraries report drive serial numbers to ACSLS and allow Auto
Configuration:
◆ Drives:
T9x40x, LTO (Gen 1, 2, 3), DLT 7000 and later drives, and SDLT.
◆ Libraries:
SCSI attached and SCSI over Fibre Channel attached STK libraries 97xx, L20/40/80,
L180, L700, L700e, L5500, SL500, and SL8500.
The following drives and libraries do not report serial numbers to ACSLS and must be
configured manually:
◆ Drives:
9490 (TimberLine), SD3 (RedWood) and earlier drives.
◆ Libraries:
9310 (PowderHorn), 9360 (WolfCreek), 4410, and serial-attached 9740.
Note NetBackup ACS Robot Auto Configuration will not work with HSC/LibStation
attached libraries since LibStation does not report drive serial numbers. You must
configure these devices manually.
Storage Migrator
The Storage Migrator Administration utility does not honor the console’s
NBJAVA_CLIENT_PORT_WINDOW option settings. It will use whatever free local port the
operating system provides to connect to its daemon process on the host.
NDMP
1. On UNIX systems and when the NetBackup catalog is in binary format, you must
back up the object of a symbolic link if you want to be able to restore that object. If you
only back up the symbolic link, only the link file will be restored and not the actual
object the link points to.
Also, if you back up both an object and a symbolic link to the object, the symbolic link
must appear before the object itself in the file list. If not, the backup may fail.
Finally, if a symbolic link points to an object that is one level or more higher, the
backup may fail. You should back up the object rather than the symbolic link.
If the NetBackup catalog is in ASCII format, these limitations do not exist.
1. NOM is included with the NetBackup 6.0 Enterprise Server product at no additional
charge. There are no license keys to install or manage.
4. NOM is not a clustered application. It is recommended that you do not install NOM
on a clustered master server.
6. If you are installing NOM on a NetBackup master server, ensure that NetBackup is
idle (not doing any critical operations). The PBX installer will leave dependent
NetBackup services stopped. You must restart these services after NOM is installed.
7. To use the purge command to clean up the NOM database, the DoDelete option must
be specified. File-path includes the file name. NOMAdmin -Purge [Never | Daily |
Weekly | Monthly] File-path DoDelete reload-file-path.
8. An alert policy can be defined for the alert conditions Lost Contact with Media Server
or Mount Request, but no alerts are raised and no alert notifications are sent for either
condition.
9. If the entire NOM domain is selected in the monitoring alerts summary view, alerts
are correctly shown in the summary displayed at the top. But, these alerts are not
reflected in the Group Component Summary table displayed at the bottom.
Also in the monitoring alerts detail view, the total alert count shown in the details
table does not match the total number of alerts in Alert Summary Panel window. This
happens because the details table does not include a row for the count of domain level
alerts.
10. The following localization issue applies only to NOM software installed on Solaris
servers.
If the system locale (where the NOM server is installed) is Japanese or Chinese, alert
policy names should be entered in English when an alert policy is created. Otherwise,
email notification will not be sent.
11. When viewing log files, some files may be truncated by the NOM server if it runs out
of memory. This may occur when processing larger log files.
12. With the default VxUL debug logging level set to 1, NOM produces extensive debug
messages. Set the debug level to 0, if extensive debug messages are not required.
If NOM stops logging messages, examine the nblog.conf file with an editor to set
the log level, making sure that there are no trailing spaces.
13. Logs for jobs of policy type Vault are not supported in this release of NOM. When
Vault logs are viewed, no logs are returned or the NOM server returns an error
exception.
14. You are unable to view all of the log messages from NOM interface when a message
record contains multiples lines. NOM only shows the first line of NetBackup log
messages that contain more than one line.
15. The table component that NOM uses in monitoring and reporting only supports fixed
width tables without horizontal scrolling. Therefore, some tables in NOM will be
unreadable due to shortage of screen real-estate. The remedy is to reduce the number
of selected columns to view using the Table Settings icon or expand the column width
using the mouse (this icon is not available with reports).
16. The Throughput Variance standard report shows incorrect data for job attempt
instances.
17. Media filtering (for example, filtering for cleaning media or frozen media) is not
operational in this release.
18. NOM displays the retention level for media as an integer value instead of showing the
currently associated retention period with the integer.
19. Custom filters created when using Netscape are not seen from Windows Internet
Explorer.
The NOM UI initializes a user’s personal preferences and various settings such as
filter definitions and column settings, from disk on startup and keeps any changes
made during the session in memory and saves them to disk at session logout.
Therefore changes made in one session are private and will not be visible in another
session. Also, changes made in one session may get overwritten by another session.
20. User authentication fails (User authentication failed) when you try to log in with
Unixpasswd user or NIS user when Root+AB is installed on windows.
When NOM is used on a Windows server, users in NIS and Unixpasswd domains can
not be used.
21. The NOM alert policy conditions window in a Netscape browser does not show all
available conditions and descriptions, and does not contain a scroll bar. This window
can be resized to show all conditions by grabbing and dragging the icon at the right
bottom corner of the window, as shown in the following figure.
22. When the task "Enable All Logs" is used from the Managing -> Host Properties view,
the browser will appear to hang for the duration it takes to perform the operation on
the NetBackup server. Please wait for the operation to complete. The wait will depend
on system performance and can be in the order of minutes.
23. NOM truncates Email and SNMP recipient names and master server group names to
60 characters. In this case, use unique names that are less than 60 characters.
NetBackup Vault
1. The command line interface for the parsable output from the Vault reports is not
published.
2. Using the vlt_ejectlist_notify script, users can eject and vault media that are
managed by Media Manager but not created by NetBackup or Vault. This is achieved
by updating the addon_medialist file in the session directory. Note however that
even if Vault fails to eject one or more of such media for whatever reason, it will not
fail the job. The job will succeed if the regular media was ejected successfully.