Answers in this category:
Disclaimers and legal stuff
What is the purpose of this FAQ?
Who maintains this FAQ?
How do I submit a question for the FAQ?
How do I obtain a copy of the FAQ?

Answers in this category:
Where is Veritas’ web site?
How can I keep up with notifications and updates for NetBackup and Media Manager?
How can I keep up with all of Veritas' product lines?
What about documention? What's available?
Is there an user's group mailing list for Veritas NetBackup?
Can I search the archive of this mailing list?
Is there a newsgroup for NetBackup?
Is there a knowledge base that I can browse?
Where can I get patches for NetBackup?
How can I get special patches for NetBackup extensions?
Can I get technical support from Veritas via email?
How do I get technical support from Veritas?

There's actually quite a bit:

On UNIX, there are actually three directories with manpages under /usr/openv  (/usr/openv/man, /usr/openv/netbackup/man, /usr/openv/volmgr/man).  If you set your MANPATH to include these directories, there's quite a bit of help there.

On NT, there is the standard NT help, of course.

On all platforms, there are PDF versions of all the manuals available.  These files are shipped with the software.  If you don't have them, you should contact your sales rep.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Yes.  Therre is a search engine for the archive at http://mailman.eng.auburn.edu/search.html.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Answers in this category:
What version(s) of NetBackup server am I running?
What version(s) of NetBackup client am I running?</
What patches have been installed for NetBackup?
I am thinking about upgrading to 3.4. Must I upgrade my clients to 3.4 also?
What version of Media Manager am I running?
How can I tell what the firmware level of my tape drive is?

1. Upgrade all your 3.2 clients to the latest 3.2 jumbo patch before upgrading your server to 3.4.
2. Upgrade all your 3.2 clients to 3.4.

In either case, you just need to load the software and the latest patch, and issue the command update_clients (for UNIX clients).  For NT Clients, you will need to upgrade them from an NT server with the proper permissions.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Answers in this category:
Is NetBackup licensed?
How do I get my licenses?
What type of licenses are there?
How does upgrading work?

Answers in this category:
What is the difference between a backup and an archive?
What is the difference between a backup and a "True Image" backup?
What is the difference between multiplexing and multistreaming?
What is the difference between a cumulative incremental backup and a differential incremental backup?
What is busy file processing?

Answers in this category:
How can I learn more about Veritas NetBackup?
Where can I go for training on Veritas NetBackup?
Are there any good books on backup and recovery?

The PDF version of the NetBackup Product docs can also be obtained at:
http://support.veritas.com/
 -> Product Listing
   -> NetBackup Products
     -> NetBackup DataCenter
       -> Online Documentation

How do you configure the OTM on NT/2000?

Answers in this category:
What is NetBackup’s main configuration file?
Why is there a bp.conf file on both the server and the client?
What are some of the more common configuration options for bp.conf and what are they used for (version 3.2)?
How can I receive email notification from NetBackup?
How do I set the maximum number of client connections per client?
How do I set the maximum number of streams for multiplexing?
What is Media Manager’s main configuration file?
How do I exclude files and directories from regular backups?
How do I exclude files and directories from NDMP backups?
How do I properly backup my Oracle/Sybase/Informix/NetApp, etc?
How do I get Oracle/Sybase/Informix/NetApp/etc. to back up over a private network?
How do I get <some script> to run before a backup job starts?
How do I backup through a firewall?
What registry settings should I make to NT servers in order for the OTM to work correctly?
What is the archive bit in NT, and how does it affect me?
Is there an easy way to make registry changes on multiple NT clients?
I have a server with multiple interfaces. How do I get NetBackup traffic to go through both interfaces?
How do you configure NetBackup to work with Network Appliances?
Is there VMS support within a VERITAS NetBackup v3.4 environment?
How do I disable OTM for NetWare?
How do I improve slow backup times for NetWare clients?
One way to backup data outside a firewall.
How do I back up a Netbackup Catalog Database that is too large to fit on a single tape?
What are the "client type" values, as used in bpbackup -t <client type>?
What's with all of these different "max jobs" settings?
How do I use an alternate media server for a restore?
Why did the chicken cross the road?

Answers in this category:
OTM - Status code 11
How do you config OTM in NT4/Windows 2000 to fix error 11?

Caution! The information in this TechNote is based upon certain assumptions, including product, operating system and platform versions. You can review this information in the TechNote Summary portion of this document. This document (233254) is provided subject to the disclaimer at the end of this document.

----------------------------------------------------------------

Symptom:
Backup jobs end with an Exit Status Code 1 or Exit Status Code 11 when using NetBackup with the Open Transaction Manager (OTM) Feature.

Exact Error Message:
exit status code 1: the requested operation was partially successful; or exit status code 11: system call failed
Solution:
This TechNote provides some general guidelines and recommendations for configuring Open Transaction Manager (OTM).
When utilizing OTM, the following two error conditions have been seen in some environments:
Exit Status Code 1 The Backup is started but the required quiescent period is not achieved. OTM is not enabled during the course of the Backup. The backup job will end with an exit status code 1 as open or active files are not included in the back up.
Exit Status code 11 The backup is started and OTM is successfully initialized (quiescent period achieved and snapshot taken). When OTM reaches this point but experiences a resource limitation within the system, the backup job will end in an exit status code 11.

Guidelines:
Disk Cache settings are an integral component for OTM operations. The documented defaults for the cache settings are "initial = 10MB" and "max = 50MB". However, best results have been achieved with the OTM cache settings as "initial=100MB" and "max=0MB". Because of the diversity of environments, these settings will vary. Manually configured values may be needed due to low free disk space or the amount of data to be backed up. Below are some recommendations to allow OTM to work as dynamically as possible. Sufficient disk space is required for OTM to be operational. Insufficient system resources may prevent implementation of OTM in the operating environment. Also, OTM may not be a viable open file manager solution on NT servers that are extremely busy. OTM needs a "quiescent" or inactive period of time (5 seconds by default) to create a snapshot of a disk partition. Backup jobs that run when the level of disk activity is high, or a "quiescent", or inactive period of time cannot be achieved, will end with an "exit status code: 1". This indicates the backup was only partially successful.

Recommendations:
These settings will allow the most scalability with OTM. These settings are not guaranteed to work on all NT servers. Depending on free disk space or available memory, these settings may have to be manually configured.
1. SNAPSHOT TYPE: Set OTM snap shot type to "Individual" unless multiple data streams is enabled. The "Individual" setting will allow OTM to create a snapshot per drive and will allow the cache file to be minimum size as a snapshot of one drive at a time will occur. This setting is recommended for servers with low disk space and when Multiple Data Stream backups are not enabled.
2. INITIAL CACHE SIZE: Set Initial cache size to 100MB. OTM will create the initial cache file before the backup starts. An initial cache setting of 100MB will allow the backup to start faster, but may slow performance slighty if the cache file needs to grow during the course of the backup due to open file activity. Setting this value higher will create a longer delay at the start of the backup until the cache file has been created. Creating a larger Initial cache file may increase performance during backups if frequent disk writes occur during the backup. The time for building the initial cache file will primarily depend on CPU speed and server activity.
3. MAX CACHE SIZE: Set Max cache size to 0MB. With the Max Cache Size set to 0MB, OTM will calculate 10 percent of the used disk space as its maximum.
- When "Global" is set for the snap shot type, OTM will calculate 10 percent of all used disk space for all disks included in the file list to be backed up.
- When "All Available" is set for the snap shot type, OTM will calculate 10 percent of all used disk space for all local disks on the server including drives that are not specified in the file list of files to be backed up. This value may need to be manually configured on some NT server that have over 35GB to be backed up (See below in Caveats).
4. CACHE FILE FIELD: Ensure the "Cache file" field is not configured or left blank. The Cache file field is used to allow the user to define the location of the Cache File. If this field is blank, OTM will dynamically and automatically place the cache file on a drive that has the most free space. OTM will attempt to place the cache file on a drive that is not being backed up. Due to available disk space, the cache file may end up on the drive that is active during the backup. When this occurs, the Initial cache file setting is not used and OTM will create an Initial cache file the same size as the maximum cache file size.
5. OTM ERROR CONTROL: Set "OTM Error Control" to "Retry OTM". This will allow NetBackup to pause the backup when the cache file fills up. OTM will then destroy the cache file, create another snapshot of the drive(s), and continue the backup.
6. CACHED FILES: It is recommended the minimal value for this field is set to 100. When this field is set to 0, the cache file will not grow beyond the size of the configured "initial cache file" size. This applies even if a maximum cache file size set. Setting this field to 100 allows the cache file to flush the changed files after every 100 open files have been backed up. Additionally, a setting in the cached file field allows the cache file to expand beyond the initial cache file setting. If the system contains a large number of small files (file size less than 100K), it is recommended to increase this setting.
CAVEATS
OTM cannot be used over a mapped drive or on a shared drive of a cluster if the node does not have control of the drive. Blue Screens are a common symptom of mapping drives on nodes that do not have control of the drive.
Multiple Data Streams Multiple Data Streams is set for improvements in performance. However, this setting also requires additional system resources to be available when configured.
If the class configuration in NetBackup is set for multiple data streams, it is necessary to set the following parameters for this feature to operate: - Snapshot type must be set to "Use all available drives". - On the Master server, within the NetBackup Global Attributes, the "maximum jobs per client" must be equal to, or more than the number of streams being used.
Please refer to the following link for Microsoft limitations while utilizing NetBackup's OTM feature:
http://support.microsoft.com/support/kb/articles/Q126/4/02.asp?LN=EN-US&SD=gn&FR=0&qry=Q126402&rnk=3&src=DHCS_MSPSS_gn_SRCH&SPR=MSALL
WIN2K Specific Caveats:
File Systems < 30GB

- Max Cache size can be set to 0MB, this setting will allow autocache configuration of 3000 MB (3GB or 10% of 30GB).

File System > 30GB
- The error control setting must be set to "Retry OTM". - Max cache size must be manually configured to 3000MB or less. - This setting will avoid the MS limitation listed above.

OTM uses a "Kernel" mode driver. The most memory kernel mode drivers in WIN2K can allocate is roughly 380MB of memory and is shared by all kernel mode drivers in the system. OTM uses 40KB of available memory per MB of OTM cache file. Depending on the number of kernel mode drivers on the server, the maximum cache file size on any WIN2K server is approximately 3GB. On servers with > 30GB of data, it may be necessary to manually configure the Maximum Cache file size downward. This is dependent on the memory being utilized by other kernel mode drivers on the server. To specifically validate if OTM has reached the available kernel driver memory limitation or have reached a physical memory limitation, look for the following in the bpbkar log "OTM: not enabled (err = 0x000005aa)" from the client. This indicates the kernel mode driver memory or available physical memory has reached its limit. The backup will continue but will not utilize OTM and the backup job will end with an exit code of "Status 1" - not all files backed up. A smaller Maximum cache file size will be needed if it has been determined that there is enough available physical RAM in the server.

NT 4.0 Specific Caveats:
File System < 20GB
- Max Cache size can be set to 0MB. - This setting will allow autocache configuration of 2000 MB (2GB or 10%). - This setting will avoid the MS limitation listed above.

File System > 20GB
- The error control setting must be set to "Retry OTM". - Max cache size must be manually configured to 2000MB or less. - This setting will avoid the MS limitation listed above.

OTM uses a "Kernel" mode driver. The most memory kernel mode drivers in NT 4.0 systems can allocate is roughly 192MB of memory and is shared by all kernel mode drivers in the system. OTM uses 40KB of available memory per MB of OTM cache file. Depending on the number of kernel mode drivers on the server, the maximum cache file size on any NT 4.0 Servers is approximately 2GB. On servers with > 20GB of data, it may be necessary to manually configure the Maximum Cache file size downward. This is dependent on the memory being utilized by other kernel mode drivers on the server. To specifically validate if OTM has reached the available kernel driver memory limitation or have reached a physical memory limitation, look for the following in the bpbkar log "OTM: not enabled (err = 0x000005aa)" from the client. This indicates the kernel mode driver memory or available physical memory has reached its limit. The backup will continue but will not utilize OTM and the backup job will end with an exit code of "Status 1" - not all files backed up. A smaller Maximum cache file size will be needed if it has been determined that there is enough available physical RAM in the server.

For further information, refer to the VERITAS NetBackup DataCenter for Windows NT/2000 System Administrator's Guide.

Acknowledgments: Microsoft

-------------------------------------------------------------------------------- TechNote Summary: TechNote Title: Backup jobs end with an Exit Status Code 1 or Exit Status Code 11 when using NetBackup with the Open Transaction Manager (OTM) Feature. TechNote ID: 233254 Last Updated: December 25 2000 02:42 AM GMT Related Documents: TechNote: 233219 - Special considerations to be made when using Open Transaction Manager (OTM) to backup and restore databases. TechAlert: 233307 - Backing up databases using Open Transaction Manager (OTM) can produce an inconsistent image of the database, leading to potential restore problems. Included in this Alert are links to general guidelines when utilizing the OTM feature.

VERITAS Software, 1600 Plymouth Street, Mountain View, California 94043 World Wide Web: http://www.veritas.com/ Tech Support Web: http://support.veritas.com/ E-Mail for Classic VERITAS Products: support@veritas.com E-Mail for Classic Seagate Software Products: helpdesk@support.veritas.com FTP:ftp://ftp.support.veritas.com/ or http://ftp.support.veritas.com/

THE INFORMATION PROVIDED IN THE VERITAS SOFTWARE KNOWLEDGE BASE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. VERITAS SOFTWARE DISCLAIMS ALL WARRANTIES, EITHER EXPRESS OR IMPLIED, INCLUDING THE WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT SHALL VERITAS SOFTWARE OR ITS SUPPLIERS BE LIABLE FOR ANY DAMAGES WHATSOEVER INCLUDING DIRECT, INDIRECT, INCIDENTAL, CONSEQUENTIAL, LOSS OF BUSINESS PROFITS OR SPECIAL DAMAGES, EVEN IF VERITAS SOFTWARE OR ITS SUPPLIERS HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. SOME STATES DO NOT ALLOW THE EXCLUSION OR LIMITATION OF LIABILITY FOR CONSEQUENTIAL OR INCIDENTAL DAMAGES SO THE FOREGOING LIMITATION MAY NOT APPLY.

mailto:kky@hotmail.com?subject=NetBackup Faq-O-Matic, mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

• ALLOW_MEDIA_OVERWRITE should be used on the server if you wish to overwrite previously used non-NetBackup media without prompting.
• ALLOW_MULTIPLE_RETENTIONS_PER_MEDIA should be used if you have a limited supply of media and a large number of retention periods.
• ALLOW_NON_RESERVED_PORTS will allow ports 1025 through 5000 to be used for data streams instead of 512 through 1024.
• Use BPBACKUP_CLASS to set the default class used for client initiated backups.
• Use BPBACKUP_SCHED to set the default schedule used for client initiated backups.
• Use BPEND_TIMEOUT to increase the amount of time bpend scripts have to finish (don’t forget CLIENT_READ_TIMEOUT must be greater than or equal to BPEND_TIMEOUT).
• Use BPSTART_TIMEOUT to increase the amount of time bpstart scripts have to finish (don’t forget CLIENT_READ_TIMEOUT must be greater than or equal to BPSTART_TIMEOUT).
• Use BUSY_FILE_ACTION to send email notification, try again, or ignore files that cannot be accessed.
• BUSY_FILE_DIRECTORY sets the working temp directory when using busy file processing.
• BUSY_FILE_NOTIFY_USER tells NetBackup whom to notify when busy file email is sent.
• BUSY_FILE_PROCESSING turns on busy file processing.
• CLIENT_NAME specifies the exact name that the client is known as to the NetBackup server(s) it is served by.
• Use CLIENT_READ_TIMEOUT to give bpstart and bpend scripts enough time to finish.
• Use DISALLOW_CLIENT_LIST_RESTORE to prevent clients from listing and restoring files backed up on their systems.
• Use DISALLOW_CLIENT_RESTORE to prevent clients from restoring files backed up on their systems.
• Use DISALLOW_SERVER_FILE_WRITES to prevent server initiated restores or server initiated updates to bp.conf.
• DO_NOT_RESET_FILE_ACCESS_TIME can be used if you do not care about atime but do not want your ctime messed with.
• INITIAL_BROWSE_SEARCH_LIMIT is used to set the default number of days that NetBackup will search to find files for restore.
• Use KEEP_LOGS_DAYS to specify the number of days to keep client logs.
• Use LIMIT_BANDWIDTH to throttle network saturation to a specified number of KB per client.
• Use LOCKED_FILE_ACTION to skip files that have mandatory locking.
• Use REQUIRED_INTERFACE to override the operating systems choice of network interface for server to client communications.
• Use SERVER to specify the NetBackup server(s) the client should use.
• Use the VERBOSE option to log additional information in NetBackup’s logs.
• USEMAIL will specify the user to notify for NetBackup events.
• USE_CTIME_FOR_INCREMENTALS will cause NetBackup to check both the ctime and the mtime when determining which files should be backed up for incrementals. You MUST also enable DO_NOT_RESET_FILE_ACCESS_TIME when using this option.

Here are some comments about excluding files from Larry Kingery <larry.kingery@veritas.com>

/proc and cdrom filesystems are recognized as such on the more common unix systems, and their contents are automatically excluded (see all log entries).  If you do wish/have to specify them, you might want to say something like "/proc/*" so that the /proc mount point exists in case of full system restore (Solaris for one will get mad if it doesn't).

Personally, I feel that excluding things like core files probably adds more in the preprocessing than it's worth (that's one more strcmp on every file).

I really hope that people will NOT exclude the nbu db during regular backups.  (Note from Curtis: I agree.) Let's say that just before leaving on Friday afternoon, you accidentally delete half your images directory.  Come Monday morning, you realize what you did, but by this time it's too late, both of your db tapes have been used, and now you're scanning tapes. Of course, depending upon your needs and time since last backup of the server, you may end up scanning anyway, but at least you're closer (if you're lucky and the master wasn't one of the clients you've just blown away).
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

The way to exclude files on NetApp when using NetBackup with NDMP is to use an entry in the file list that sets the EXCLUDE NDMP environmental variable.

For example, if you wanted to exclude a directory named /vol/vol1/A, but wanted to dump the rest of the volume, your NetBackup class file list would look like this:

SET EXCLUDE = A
/vol/vol1

To exclude multiple files or directories, you would use commas to delimit the field:

SET EXCLUDE = A,B,C
/vol/vol1

The rules are stated as being the same as they are for native dump commands.

The exclude list is specified relative to the dump path.  If a directory name or file name contains a space, you probably need to enclose the string in quotes:

SET EXCLUDE = 'file with space in name'
/vol/volx

Take a look at the ontap manuals for DUMP for more information.

Q: What files and directories should I exclude from a backup?

A: On UNIX systems, you should exclude /proc, which is your system’s memory represented as a filesystem. You may also want to exclude files such as core files, he cdrom partition, and the /tmp partition. Remember that the /usr/openv/netbackup/logs/ directory can grow at an enormous rate if bpdbm is logging.  On both NT and Unix systems, you should also exclude database files that you are backing up by other means (e.g. Oracle, SQL Server, etc.).
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Use the REQUIRED_INTERFACE setting on the Oracle client to tell it what NIC to use.

This seems to help.  It's in the NBU System Admin Guide under bp.conf settings.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Michael Fenton Veritas Software - Technical Support Engineer Direct Phone: 651-604-3230 Email: Michael.Fenton@veritas.com Hours: 9 AM - 6PM CST Mon-Fri Dispatch 1-800-342-0652 for immediate assistance
mailto:g_sperano@hotmail.com?subject=NetBackup Faq-O-Matic

     * add ALLOW_NON_RESERVED_PORTS to the bp.conf on master server

     * add ALLOW_NON_RESERVED_PORTS to the bp.conf on client

     * run on your master server as root:
               bpclient -client $CLIENT -add -connect_nr_port 1

Just for posterity's sake, here's an example configuration:
Client's bp.conf includes the following:
ALLOW_NON_RESERVED_PORTS
CLIENT_PORT_WINDOW = 13740 13750
RANDOM_PORTS = NO
master server's bp.conf includes the following:
ALLOW_NON_RESERVED_PORTS
SERVER_PORT_WINDOW = 13740 13750
RANDOM_PORTS = NO
Ports that need to be open on the firewall: TCP on port 13782, 13720 & 13740 through 13750

mailto:rob@worman.org?subject=NetBackup Faq-O-Matic

[HKEY_LOCAL_MACHINE\SOFTWARE\Veritas\NetBackup\CurrentVersion\Config]
"Use_Archive_Bit"="NO"
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

There's a way that I've been working on to do this with NetBackup, but I've never actually done it.  Here are the steps:

1. Create a class that will backup some file that will be on every NT client (e.g. C:\Winnt\system32\drivers\etc\hosts).  Let's call the class "UPDATE."
2. Write a batch file that contains the command below, and call it bpstart_notify.UPDATE.bat.
    start /wait c:\winnt\regedit.exe c:\temp\update.reg
    del c:\temp\update.reg
3. Create a file that contains the registry key you want to edit and call it update.reg.  For example:

[HKEY_LOCAL_MACHINE\SOFTWARE\Veritas\NetBackup\CurrentVersion\Config]
"Use_Archive_Bit"="NO"

Here's another way, From: Scott Kendall <Scott.Kendall@ln.ssw.abbott.com>

For changes like this, I usually run a "for" loop against my NT servers using the NT Resource Kit utility "reg.exe".  This utility lets you remotely edit/add/delete Registry keys.

The command below from an NT box would add the value of the second column for each server listed in the first column of each line in fileA, assuming you
have the reskit and it is in your path (note: "add" does not replace if it already exists.  If so , you must either "update" or "delete" then "add").

fileA contents:
server1     server1
server2     somethingelse
server3     server3

command:
C:\>for /f "tokens=1-2" %i in (fileA) do reg add
HKLM\Software\Veritas\NetBackup\CurrentVersion\Config\REQUIRED_INTERFACE=%j
REG_SZ \\%i
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

He had three clients, each with a different inter

In the master's bp.conf file:

SERVER=master-neta
SERVER=master-netb
SERVER=master-netc

One entry for each NIC in the box.

For the clients on NETA - SERVER=master-neta
For the clients on NETB - SERVER=master-netb
For the clients on NETC - SERVER=master-netc
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Lets say the hostname of your Network Appliance box is "toaster" and the hostname of your Master Server is "dumpster"

Do the following:

1) Login to dumpster as root, and install the NDMP packages (SUNWnbdmp). If you are not aware, you have to purchase the NDMP option from Veritas for NetBackup. You get the NDMPD package, documentation, and paper license at least with 3.2.

2) Set your NDMP authorization:

dumpster# /usr/openv/volmgr/bin/set_ndmp_attr -auth toaster root

It will ask you for a Password, and enter toaster's password. 

3) Put the following line in /usr/openv/netbackup/bp.conf:

ALLOW_NDMP

4) Connect toaster to one of the drives in your Jukebox, and reboot it so it can recognize the drive. Unfortunately, NetApps dont have drvconfig or alike.. 

Check to make sure the drive is recognized after reboot:

toaster % sysconfig -t

This will show you the drive, and all device files you can use with it. I normally use the norewind device nrst0a. (or b.. whatever comes up in sysconfig's output)

5) On toaster, start the ndmpd daemon. ndmpd daemon comes with DataONTAP so it should be there (atleast in recent versions). To start ndmpd, do

toaster % ndmpd on 

To see the usage of ndmpd, just enter ndmpd. 

6) Come back to your master server (dumpster), and add the NDMP drive:

Pull up xdevadm, select DRIVES -> ADD DRIVE. This will pull up the ADD DRIVE window. In that window, select/provide the following information:

DRIVE TYPE: DLT (or what ever type your drive is)
DRIVE INDEX: 0 (or any number of your choice)
DRIVE NAME: toaster_jukeboxname_drive# (or what ever you like)
NO REWIND DEVICE: toaster:nrst0a
DRIVE STATUS: UP
CLEANING FREQUENCY: 300 (or what ever you like)
ROBOTIC DRIVE: YES
ROBOT TYPE: TLD (or what ever type your Jukebox is)
ROBOT NUMBER: <your robot's number>
ROBOT DRIVE: <drive number of the drive thats connected to toaster>

At this point, you're ready to test NDMP backups. Use xbpadm to create a class of type NDMP, and include toaster as client, and a sample directory under file list. Create a schedule "manual_backup" dont put any regular dates on it, and start a manual backup of that NDMP class for toaster and see how it goes. 

You do not have to install any software on toaster. All you need to do is start ndmpd. You want to put that in its rc file so its started every time its rebooted.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

NEW_STREAM
/SYS
/SSI
NEW_STREAM
/VOL1
NEW_STREAM
/VOL2
NEW_STREAM
/VOL3

3) Set the schedule's to match MPX with the number of streams; in this case 4.
This change improved backups for some large NetWare fileserver from 9.7 Gig an hour to 20.3 Gig an hour.
mailto:ryan_anderson@udlp.com?subject=NetBackup Faq-O-Matic, mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

NEW_STREAM
/SYS
/SSI
NEW_STREAM
/VOL1
NEW_STREAM
/VOL2
NEW_STREAM
/VOL3

3) Set the schedule's to match MPX with the number of streams; in this case 4.
This change improved backups for some large NetWare fileserver from 9.7 Gig an hour to 20.3 Gig an hour.
mailto:ryan_anderson@udlp.com?subject=NetBackup Faq-O-Matic, mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

http://www.backupcentral.com/cgi-bin/netbackup-fom?_highlightWords=firewall&file=70
mailto:rob@worman.org?subject=NetBackup Faq-O-Matic

For a long time, this was an "unsupported, but we'll tell you how to do it" 
tactic from Veritas support.  With the release of 3.4, it became the official 
way to do it and it's documented on page 157 of the NetBackup Admin Guide.

But if you're running NBU 3.2 you might wonder how to do this.  Here it is, 
as excerpted from the 3.4 documentation:

Multiple-Tape Backups 
===================== 
The basic concept behind the protection of large NetBackup catalogs is
to split the catalog-backup process into two steps. The first step backs
up the majority of the data from the images subdirectory on the master
server, while the second step backs up a small sub-set of the images
subdirectory, together with the remainder of the NetBackup and Media
Manager catalog files and directories from the master and media servers.

Since the first backup contains the majority of the data, it must be
able to span tapes. This is achieved by using a normal NetBackup job to
back up the data. As a result of this normal backup, an entry is placed
in the images subdirectory tree for the master server. This catalog
entry allows the user to browse the catalog for files during a restore
operation.

The second backup must back up the portion of the images subdirectory
that contains the catalog entries for the master server, together with
the other parts of the NetBackup and Media Manager catalogs. Since this
is a relatively small amount of data, it fits onto a single tape. It
must also be possible to recover this backup without the NetBackup
catalogs being available. This is achieved by using the normal NetBackup
catalog-backup mechanism to perform the backups.

Setting up Multiple-Tape NetBackup Catalog Backups 
================================================== 
In order to configure NetBackup to perform multiple-tape backups of its
catalogs, you must define both a normal NetBackup class and make changes
to the NetBackup catalog-backup configuration. In addition, you must
create a shell script or executable file that will initiate the
multiple-tape catalog backups. These steps are detailed below.

Define a NetBackup Class 
------------------------ 
1. Use the NetBackup Administration interface to create a new class. 
Set the class attributes as follows: 
--Set the Class type to Standard if the master server is a UNIX machine 
or MS-Windows-NT if the master server is a WindowsNT/2000 machine. 
--Do not choose Cross mount points if the master server is a UNIX or
Windows 2000 machine.
--Do not choose Follow NFS for UNIX or Backup network drives for Windows 
NT. 
--Pick a suitable storage unit and volume pool. 
--Set Limit jobs per class to 1. 
--Do not choose Compression. 
--Set Job priority to 0.

2. Add the master server to the client list

3. Put the following path in the file list: /usr/openv/netbackup/db/images

Note On UNIX, if /usr/openv/netbackup/db/images is a symbolic link to
another filesystem, you MUST specify the true location of the images
directory here. Symbolic links do not apply to Windows NT/2000.

4. Set up schedules to meet your requirements. VERITAS recommends that
the class contain only a full backup schedule, since this will minimize
tape mounting and positioning during restores.

Do not set any backup windows for the schedules that you define. This
ensures that the backup class is never initiated automatically by the
NetBackup scheduler. Instead, you must initiate the backup job manually.

5. Save your changes.


Configure the NetBackup Catalog Backups 
======================================= 
Use the NetBackup Administration interface to set up the configuration
of the NetBackup catalog backups. Make the following changes:

1. Ensure that the Media server setting is for the required backup server. 
2. Specify the following for Absolute pathname.

  masterserver:/usr/openv/netbackup/db/[A-Za-hj-z]* 
  masterserver:/usr/openv/netbackup/db/images/masterserver 
  masterserver:/usr/openv/var 
  mediaserver1:/usr/openv/netbackup/db/media 
  mediaserver1:/usr/openv/volmgr/database (repeat for all media servers) 
  
3. Change the schedule to Only when manually initiated. This stops the
NetBackup catalog backups from running automatically and allows you to
control when they run manually.

4. Select appropriate media types, densities, and IDs for the two
catalog-backup media. 

5. Save your changes. 


Create a Shell Script to Initiate the Backups 
=============================================
It is also important that the second-stage backup of the NetBackup
catalogs occurs directly after the first-stage backup. A good way to
ensure this is to write a script that initiates both backups, one after
the other. The following is an example catalog-backup script:

#!/bin/sh
#
# catalog_backup script
#
# Performs a two-stage backup of the NetBackup catalogs
#
CLASS=nbu_cat_backup # Change to the name of the correct class
SCHED=full_backup # Change to the name of the correct schedule
LOGDIR=/usr/openv/netbackup/logs/catalog_backup
if [ -d $LOGDIR ]; then
exec >> $LOGDIR/log.`date “+%m%d%y”` 2>&1
else
exec > /dev/null 2>&1
fi
echo “Running first stage catalog backup”
/usr/openv/netbackup/bin/bpbackup –w –i –c $CLASS –s $SCHED
EXIT_STAT=$?
if [ $EXIT_STAT –ne 0 ]; then
echo “First stage catalog backup failed ($EXIT_STAT)”
exit 1;
fi
echo “Running second stage catalog backup”
/usr/openv/netbackup/bin/admincmd/bpbackupdb
EXIT_STAT=$?
if [ $EXIT_STAT –ne 0 ]; then
echo “Second stage catalog backup failed ($EXIT_STAT)”
exit 1;
fi
exit 0;
## end of script

How To Initiate a Multiple-Tape Catalog Backup 
==============================================
Similar to how the automatic-catalog backup works, it is important to
ensure that no other NetBackup operations that modify the NetBackup
catalogs are in progress while the two catalog backups are performed.
Such operations include:
--Backups and archives 
--Catalog compression 
--TIR record expiration or retrieval (during a restore operation) 
--Catalog image record expiration 
--Image imports 
--Image duplication 

Performing the catalog backups when any of these operations are in
progress can cause an inconsistent catalog backup. Since both image
import and image duplication operations must be initiated manually by
the NetBackup administrator, it is relatively easy to ensure that these
are not in progress during the catalog backup. However, it is more
difficult to ensure that no backups or restores are running, since both
the NetBackup scheduler and other users can initiate these.

More difficult still are operations that are started automatically by
NetBackup, such as catalog compression, TIR record expiration or
retrieval, and image record expiration. Due to the way the NetBackup
scheduler interlocks processes, do not start the two-step backup script
with the /usr/openv/netbackup/bin/session_notify script.

We suggest using another scheduler (such as cron on UNIX) to start the
two-step backup script or execute it manually when the above operations
are not occurring.


Multiple-Tape Restores
======================
A restore of the NetBackup catalogs is also achieved in two steps. The
first step is to use the most recent NetBackup catalog backup to recover
the portion of the image catalog containing information about the
backups taken from the master server, together with the other parts of
the NetBackup and Media Manager catalogs on the master and, if
configured, the media servers.

Once this information has been recovered, NetBackup can be started and
one of the user interfaces can be used to browse the files backed up
from the master. These will include the files and directories that
constitute the NetBackup images catalog, which were backed up using the
first step of the catalog backup described above. Using the normal
restore process, these files and directories are restored, completing
the operation. You must ensure the option Overwrite Existing Files is
not selected, since this replaces the files previously recovered in
stage 1.

Here are the ones that I know...

standard = 0
apollo = 3
auspex = 12
afs = 22
msnt = 13
netware = 10
os2 = 14
SAP = 17
MS-SQL = 15
SQL-Backtrack = 11
Sybase = 7
DB2 = 18
Oracle = 4
MS-Exchange = 16
Informix = 6

(*) this configuration option is barely documented so it's not well known. If specified, it will override the global "max jobs per client" value. To specify a client-level "max jobs" parameter:
bpclient -client $CLIENTNAME -add -max_jobs $MAXJOBS

===== These topics are discussed in the NetBackup SysAdmin Guide as part of the discussion of the "Allow Multiple DataStreams" and "Multiplexing" features.
mailto:rob@worman.org?subject=NetBackup Faq-O-Matic, mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Answers in this category:
How can I tell what processes are running under NetBackup?
How can I tell what processes are running under Media Manager?
How do I make sure that all NetBackup processes are completely shut down?
How do I make sure that all NetBackup processes are completely shut down?
How can I check the status of my tapes?
Why do my tapes keep ejecting after every backup?
Is there a way that I can find out in advance what backups are going to be scheduled?
I have tapes in the library, and yet NetBackup isn't using any of them. I'm getting error 96s.
What to do with jobs that APPEAR to be active, but have hung?

ltid -nsu
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

bpschedreq -predict mm/dd/yyyy
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

The only thing I can think of off hand would be to look at your retention periods. NetBackup will not mix retention periods on a tape unless you force it (which I wouldn't recommend). If all the tapes that have space don't match the retention period you're going after, and there are no other tapes available in the storage unit, you might get the "No Available Media" message.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Answers in this category:
How often should I backup my indexes if I decide to use bpbackupdb?
Should I make duplicates?
How can I do tape duplication and off-site rotation with NetBackup?
How many copies can I make of a given image?
What is a good generic backup schedule?
When should I do my full and cumulative incremental backups?
How long should I keep tape archives?
Can I backup a live database with NetBackup?
How often should I send tapes off-site?
Is there a way to check the consistency of the database?
Is there a way to backup the NBU Databases onto a specific tape besides one in the NetBackup Pool?

A: NetBackup currently limits you to two copies per image, one of which is the original.  You can expire one of the copies with bpexpdate -copy n, allowing you to make another copy, but the copy that you expired will no longer be tracked by NetBackup.

This command does not actually fix the inconsistency.

At least two strings that can be searched for to find problems:
"Bad image header"
"does not exist"

I've also found it useful to filter out lines containing the strings "checking image file" and "checking files file". What's left will be a list of machines followed by errors.

Generally your only recourse to correct the problem is to delete the offending images. Delete the image header, the image '.f' file, and the image indexes. The indexes for the image are located in /usr/openv/netbackup/db/images/client-name/INDEX and will have the same file name as the header file plus suffixes that match '.i?'.
mailto:vandevegt@yahoo.com?subject=NetBackup Faq-O-Matic, mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Answers in this category:
How do I get additional logging to occur?
Where does NetBackup put its logs?
Where does NetBackup place logs for user initiated activity?
Where can I find tape drive errors?
Where do the processes log information?

Master Server Logs

• bpadm - Admin
• bprd – request daemon
• bpdbm – database manager
• bpsched – backup scheduler
• xbpadm – X administrator interface
• xbpmon – job monitor

• bpbrm – Backup/Restore Manager
• bpcd – Client Daemon
• bptm – Media manager for tape
• bpdm – Disk Manager

• bp – Menu user interface (curses)
• bparchive – Client archive program
• bpbackup – client backup program
• bplist – lists backup and archive images
• bprestore – client restore program
• bpbkar – manages backups and archives
• bpcd – client daemon
• bphdb – hot database backup program
• db_log - backup log for databases

Answers in this category:
What do I need to know before I do a recovery of a whole system?
How do I recover a filesystem, directory, or file?
How can I tell which tapes will be needed for a particular restore?
How do I restore from NDMP tapes, especially without NetBackup?

There is a product available for NetBackup users called Intelligent Disaster Recovery (IDR) that completely automates this process.  IDR is available for NT in 3.4, and Veritas intends to have support for popular Unix platforms in 4.1 (3Q 2001).

This is referred to as a bare metal recovery.  This topic is covered in some detail in Unix Backup & Recovery.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

1. If you are using the Java GUI, start the restore, and click "Preview Media Required".
2. You can also perform the restore (either via the GUI or command line), and have messages written to a log file.  The beginning of the log will contain a list of tapes needed for the restore.

Let's assume you want to restore from a backup of CLIENTA that happened on August 1st.

(these commands would be run on your NBU master server)
### First you need to find the exact start time of the backup you want:

 nbumaster# bpimagelist -U -client CLIENTA -d 08/01/2001 -e /08/01/2001

 Backed Up         Expires       Files       KB  C  Sched Type   Class
 ----------------  ---------- -------- --------  -  ------------ ------------
 08/01/2001 05:05  08/15/2001   729846 52624096  N  Full Backup  CLIENTA_OS
 08/01/2001 02:05  11/02/2001    58233  5029312  N  Full Backup  CLIENTA_DATA

 nbumaster# bpimagelist -media -U -client CLIENTA -d 08/01/2001 05:05 -e 08/01/2001 05:05

 Media ID   Last Written       Server
 --------   ----------------   ----------
 602595     08/01/2001 05:05   nbumaster
 603084     08/01/2001 05:05   nbumaster 

We have read backup tapes created on the NetApp using NDMP manually using the following two methods:

1) Veritas via NDMP just uses the NetApp "dump" command to creat the backups.  This format is compatible with ufsdump, thus the tape can be read on a Solaris system using ufsrestore (though some information maybe lost if using NT ACL's).

2) the following is cut and paste from my notes. empress is the Sun where Veritas is installed and odin is the NetApp...

to restore files on odin without the aid of netbackup:

from empress, have a tape mounted:
tpreq -ev def100 -p odin-vol0-full /tmp/tp

now do look at /tmp/tp to find out what tape drive has the tape
cat /tmp/tp
odin /ndmp/nrst0a

from odin, we can manipulate the tape using mt.

First, fast forward past the volume header
mt -f nrst0a fsf 1

Second, fast forward past the partition header
mt -f nrst0a fsr 2

cd to the directory where you want files restored to

to list dumpfile contents:
restore tf nrst0a

(If you issue the restore command above, you will need to rewind the tape and reissue the fsf 1 and fsr 2 commands before you issue the command below.)

to restore the files:
restore xf nrst0a ./dir/path

when finished, unload tape:
tpunmount /tmp/tp
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Answers in this category:
Can non-root or non-Administrator personnel administer NetBackup?
How is the NetBackup product installed and configured?
How do I start the administrative GUI?
How do I add a robot to a Media Server?
How do I add a tape drive to a Media Server?
How do I bring a drive up or down?
What volume pools do I want to setup?
How do I add a volume pool?
How do I delete a volume pool?
What is a volume group?
How do I delete a volume group?
How do I inventory my library?
How do I compare my library inventory to the media database?
My library inventory and volume configuration don’t match. What should I do?
How do I label media?
When do I replace media?
How do I replace media in my library?
How do I recycle (or deassign) Netbackup database (catalog) backup tapes?
Why do I want to recycle media?
How do I recycle media?
How do I change the attributes of media?
How do I add a volume to the media database?
What does a class influence?
How do I add a class?
How do I delete a class?
How does scheduling work?
How do I add a schedule?
How do I modify a schedule?
How do I delete a schedule?
How do I add a client?
How do I delete a client?
What is the difference between Cross Mount Points and ALL_LOCAL_DRIVES?
How can I securely push a UNIX NetBackup client?
How do I clear the client list in the NetBackup Windows client?
How do I empty my tape silo's inport (CAP) from the command line?

On UNIX:

Use the suggestions in the NetBackup Admin Guide labelled "Allowing nonroot users to administer NetBackup."  They amount to using chmod to change the permissions of the executables, and possible making a special group for non-root administrators.

On NT:

Just use the NT Administrative Client:

From Bob Bakh <bbakh@veritas.com>:

"It slices it dices and does all you ask for and more.  Just set up the users pc as a server in the bp.conf, restart the daemons or on NT sacrifice a lamb, or is that add their PC to the server list and restart services. The Administrative client can be found on any NetBackup Server CD, run the server install and select Server install, you get 3 choices, Master, Media, and Admin Client."

     - it needs to be re-enabled after any patching or upgrading

     - it will NOT work when you are using the java GUI (aka jnbSA).  The
     first thing jnbSA asks for is the root password.

     - it will NOT work when you are using the java GUI (aka jnbSA).  The
     first thing jnbSA asks for is the root password.

If using xdevadm, select Add Robot from the Robot dialog menu. 

If using the Java or NT GUIs, start Media and Device Management and click on the Media Manager that you wish to add the robot to and select New | Robot from the Actions dialog menu.

Select the appropriate robot type by consulting the Release Notes document – Supported Peripherals (release_notes.pdf). Specify a unique robot number that will not be used within the rest of the organization, and specify the volume database host that will store the client indexes (it is recommended that you always specify your master server). If your robot is being controlled by another system, specify the host that is controlling it. If your robot is locally controlled, specify the specify the robot control file. Consult the Media Manager Device Configuration Guide for additional details. 

The JAVA and NT GUIs will then prompt you to restart the media manager daemon, and do it for you automatically.  If you are using xdevadm, you must restart the media manager daemon (ltid) when you are ready for NetBackup to see the robot. To do this, use the command /usr/openv/volmgr/bin/stopltid followed by /usr/openv/volmgr/bin/ltid to restart it. Any backups that are in progress at that time may fail. 
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

If using xdevadm, select Add Drive from the Drives dialog menu. 

If using the Java or NT GUIs, start Media and Device Management and click on the Media Manager to which you’re adding the drive.  Select New | Drive from the Actions dialog menu. 

Select the appropriate drive type and give it a drive index number that will be unique for the media server the drive is connected to as well as a drive name which should be unique for the organization. If on Unix, specify the no-rewind device file for the device using the Berkley style close and compression (if available). If using Solaris, verify that the type of drive you are using is in the st.conf file. Consult the Media Manager Device Configuration Guide for details. If the drive and OS do not have native tape alert cleaning features (generally Exabyte 8900 Mammoths, Quantum and Tandberg DLT2/4/7000s, Sony AIT SDX-300Cs, STK ½ inch cartridge, and STK 9840s) select a frequency based cleaning routine (in hours). Set the robot type and robot number that is feeding the drive media.

The JAVA and NT GUIs will then prompt you to restart the media manager daemon, and do it for you automatically.  If you are using xdevadm, you must restart the media manager daemon (ltid) when you are ready for NetBackup to see the robot. To do this, use the command /usr/openv/volmgr/bin/stopltid followed by /usr/openv/volmgr/bin/ltid to restart it. Any backups that are in progress at that time may fail.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Also important to note is that a "DEFAULT" barcode rule is a good thing to have when you are using a scratch pool. This will insure that volumes that are automatically added to your media database will be assigned to the scratch pool.
mailto:rob@colltech.com?subject=NetBackup Faq-O-Matic

Q: How do I add a volume group?

A: You can add a volume group when you are adding new volumes. When adding a single volume and specifying the volume group, simply type in the name of the new volume group you wish to create.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

vmquery -deassignbyid media_id n 0

media_id is the tape media id
n  is the pool number assigned to NBDB pool.  You can determine the pool number with:

vmquery -m media_id

This will deassign it unless it is frozen.  The vmquery command is in
\Veritas\Netbackup\Volmgr\bin
(or /user/openv/netbackup/volmgr/bin for UNIX)
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Q: How do I copy a class?

A: If using the Java or NT GUIs, start Backup Policy Management, click on the Master Server you wish to manage, then right-click on the Class you wish to copy.  Select Copy then specify a name for the copy in the New class box.  Edit any of the class characteristics you wish to change, then click OK.  If desired, modify the schedules, files, or clients as necessary.

Q: How do I modify a class?

A: If using the Java or NT GUIs, start Backup Policy Management.  To change basic class attributes (such as Multiplexing, Class Storage Unit, etc.), click on the Master Server you wish to manage, then select the class you wish to modify, right click, and choose Change.  Select the attribute(s) you wish to change and apply.   If you wish to change schedules, files, or clients, click on the “+” symbol next to the desired class.  Then click on the area below the class you wish to change (Schedules, Files, or Clients).  The existing items will display in the right window.  If you wish to change/delete an existing item, right-click on the specific item you wish to change/delete, and select Change or Delete as appropriate.  Change the desired attribute(s) and select OK.  If using xbpadm, select the class you wish to modify, right click and choose Change Class. Select the attribute you wish to change and apply.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

If using the Java or NT GUIs, use Backup Policy Management click on the Master Server you wish to manage, then click on the “+” sign for the class you wish to work on, and right-click Schedules and select New.  Give the schedule a meaningful name and specify the frequency period and what level of multiplexing (if any) to allow.  Next specify the type of backup this schedule is for, how long to retain the backups, and specify whether or not you wish to override the default storage unit or volume pool.  Finally, you should specify what days of the week you wish to do backups, specifying the start time and the duration of the backup window for starting backup jobs in the class.  When satisfied with the settings, click OK.  If you wish to copy an existing schedule, locate the class that contains the schedule you wish to copy.  Click the “+” symbol next to the class, then click on Schedules below that class. The individual schedules will display in the right window.  Right-click on the desired schedule and select Copy.  Select the name for the new schedule and what class it should be added to.  Make any changes desired in the schedule and click OK. 

If using xbpadm, select the class you wish to work on, then select the schedule tab/button. Select New, and either copy the schedule from an existing class or give it a name and select okay. At that point you will have to specify whether or not you wish to override the default storage unit or volume pool. You will need to specify what type of this schedule is for as well as the retention period, frequency, and whether or not to multiplex with this schedule. Finally, you should specify what days of the week you wish to do backups, specifying the start time and the duration of the backup window for starting backup jobs in the class.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Q: How do I merge two media servers?

Find out all the tapes assigned to the old slave server, run bpmedia -movedb on all of them to move them to the new media server.  You will need both the old server and the new server up running during this command.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Single Stream
ALL_LOCAL_DRIVES no Allow Multiple Data Streams
/ + Cross Mount Points Allow Multiple Data Streams doesn't matter

Multiple Streams
ALL_LOCAL_DRIVES With Allow Multiple Data Streams.
mailto:bbakh@veritas.com?subject=NetBackup Faq-O-Matic, mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

A supported method for pushing a NetBackup client to a UNIX sever uses rcp and
rsh. This method is insecure, since each client must trust the NetBackup server
(via an entry in .rhosts). To make pushing a client more secure, it would be
nice to use ssh and scp in place of the "r" utilities. The modification
described here does just that.

Requirements:

1. The clients must be running sshd, and sshd must permit (a) root logins, and
(b) allow RSA (or DSA for ssh v2) authentication.

2. The NetBackup server which is to push the software

3. There must exist an ssh passkey, with the private key on the server, and
the public key in root's authorized_hosts (authorized_hosts2 for ssh2) file 
on the clients.

The Modification:

Edit the install_client script appropriate to your client's arch/OS. For
example, for Solaris 7, this would be:
/usr/openv/netbackup/client/Solaris/Solaris7/install_client
Rewrite the section defining rcp and rsh to point to your local versions of
scp and ssh, as follows:

---begin original---                                                            

# Find the rsh and rcp commands                                                 
if [ -x /usr/net/rsh ] ; then                                                   
        RSH=/usr/net/rsh                                                        
elif [ -x /usr/bin/remsh ] ; then                                               
        RSH=/usr/bin/remsh
... and so on ...
else
        /bin/echo Cannot find rsh command
        exit 1
fi

if [ -x /usr/ucb/rcp ] ; then
        RCP=/usr/ucb/rcp
elif [ -x /usr/bsd/rcp ] ; then
        RCP=/usr/bsd/rcp
elif [ -x /usr/bin/rcp ] ; then
        RCP=/usr/bin/rcp
else
        /bin/echo Cannot find rcp command
        exit 1
fi

---end original---

If ssh and scp are installed in /usr/local/bin, modify the above sections to
read as follows:

---begin modified section---

# Modified to use ssh and scp instead of rsh and rcp;                           
# unmodified file is install_client.orig -jon                                   
#                                                                               
# Find the ssh and scp commands                                                 
if [ -x /usr/local/bin/ssh ] ; then                                             
        RSH=/usr/local/bin/ssh                                                  
else                                                                            
        /bin/echo Cannot find ssh command                                       
        exit 1                                                                  
fi                                                                              
                                                                                
if [ -x /usr/local/bin/scp ] ; then                                             
        RCP=/usr/local/bin/scp                                                  
else                                                                            
        /bin/echo Cannot find scp command                                       
        exit 1                                                                  
fi                                                                              
# End modified section    
       
---end modified section---

Example:

Once you've made that change to any client you want to push, you can            
install clients via scp/ssh easily enough:                                      
                                                                                
(add clients to NetBackup, but don't let it push the clients)
(put public key on all clients, private key on NB server)
# ssh-agent bash
bash # ssh-add backup-key      # if "backup-key" is the name of your key...
Need passphrase for backup-key
Enter passphrase for jon@tweety.main.gnac.com
Identity added: backup-key (jon@tweety.main.gnac.com)
bash # /usr/openv/netbackup/bin/install_client_files rsh ALL  # really ssh!
       192.168.2.1 ...
Client 192.168.2.1 -- Solaris hardware running Solaris2.6
Installing NetBackup software on 192.168.2.1
...

One final note: it is possible to push a client using ssh with passwords, but
you'll have to type in the password a few dozen times for each client. If you
only need to push one client, it's marginally bearable.

Jonathan Tourtellot
Global Networking and Computing (GNAC)

Answers in this category:
What directory structure does NetBackup and Media Manager use?

Answers in this category:
How do I start the Backup and Restore GUI?
How do I do a basic user initiated backup?
How do I do a basic user initiated archive?
How do I do a basic user initiated restore from a backup?
How can I get a list of my backups and archives?

If using the Java or NT GUI, select the Backup and Restore GUI and click on the Backup Files tab.  Verify/set your server.  In the left browse window, expand the directory tree until you get to the directory that you want to backup (or the directory that contains the file(s) you want to backup).  Click on the small dot next to the directory to backup the whole directory.  If you wish to backup certain files within the directory, click on the directory name in the left window then select the small dot next to the file(s) you want to backup in the right browse window.  When all files have been selected, click Backup in the lower-right corner.  If you desire to have a keyword phrase, type it into the window that pops up, then click Start Backup. .  You may check progress of the backup by clicking on the Task Progress tab. 

If using the xbp GUI, go to File | Configuration and verify/set your server and client. Go to File | Browse file system (backup or archive). Set the "Directory to Search" to be the base directory of what you are interested in backing up to media. Set an appropriate "Directory Depth" so that you can see the directory tree you are interested in. Select the files and directories that you want to backup. Select Backup | Backup Selected Files and Directories. Verify the files and directories you are going to backup. Select a phrase or mnemonic that will identify the backup when you wish to find it later. Select a location for writing the progress log. Select OK, and the backup will be initiated. You may check progress of the log file by using tail –f or by using the GUI.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

If using the Java or NT GUIs, select Backup and Restore GUI and click on the Backup Files tab.  Verify/set your server.  In the left browse window, expand the directory tree until you get to the directory that you want to backup (or the directory that contains the file(s) you want to backup).  Click on the small dot next to the directory to backup the whole directory.  If you wish to backup certain files within the directory, click on the directory name in the left window then select the small dot next to the file(s) you want to backup in the right browse window.  When all files have been selected, click Backup in the lower-right corner.  Select the Archive files box.  If you desire to have a keyword phrase, type it into the window that pops up, then click Start Backup.   You may check progress of the archive by clicking on the Task Progress tab. 

If using the xbp GUI, go to File | Configuration and verify/set your server and client. Go to File | Browse file system (backup or archive). Set the "Directory to Search" to be the base directory of what you are interested in archiving to media. Set an appropriate "Directory Depth" so that you can see the directory tree you are interested in. Select the files and directories that you want to archive. Select Backup | Archive Selected Files and Directories. Verify the files and directories you are going to archive. Select a phrase or mnemonic that will identify the archive if you wish to find it later. Select a location for writing the progress log. Select OK, and the archive will be initiated. You may check progress of the log file by using "tail –f" or by using the GUI. 

Your archived files will be deleted from disk after they are written to media. 
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

If using the JAVA or NT GUIs, select Backup and Restore.  Click on the Restore Files tab.  Verify/set the server, source client, type of client, and destination client.  Note that restoring to a different client is usually best done from the Master Server rather than from a client.  Set the source of the backup and then choose the date range you wish to search.  If the backup used a keyword phrase, complete the Keyword phrase to search for box.  Set the Browse directory to the base directory of the file(s) you’re interested in restoring.  Click on the Update File List button.  This should result in a directory structure that meets your search criteria showing in the lower left window.  Expand the directory structure as necessary to locate the directory you wish to restore from.  If you want to restore the whole directory, click the small circle next to the directory name.  If you wish to restore files within the directory, click on the directory name and a list of files will appear in the right window.  Click on the small circle next to the files you wish to restore.  When all the files you wish to restore are selected, click on the Restore button.  If you wish to restore the files to an alternate path, choose the appropriate selection (Restore everything to a different location or Restore individual directories and files to different locations) and set the source and destination paths.  Set whether or not you want the restore to overwrite existing files and whether or not the restore should cross mount points.  Once the selections are correct, click on the Start Restore button.  You may check progress of the archive by clicking on the Task Progress tab.

If using the xbp user interface. Go to File | Configuration and verify/set your server and client. Go to File | Browse backups (restore). Set the "Directory to Search" to be the base directory of what you are interested in archiving to media. Set the date & time range you are interested in browsing. Specify the phrase or mnemonic used for the backup you are interested in restoring from. Set an appropriate "Directory Depth" so that you can see the directory tree you are interested in. If multiple copies of the files appear, select the backup image that you are interested in. Select the files and directories that you want to restore. If you wish to restore to an alternate path, select Restore | Specify Alternate Path and enter the alternate restore location. Select Restore | Restore Search Directory to restore. Verify the files and directories you are going to restore. Determine whether you want to overwrite existing files and/or restore to an alternate path. Select a location for writing the progress log. Select OK, and the restore will be initiated. You may check progress of the log file by using "tail –f" or by using the GUI.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Answers in this category:
I have tapes whose physical expiration date is shorter than I'd like. How do I change that date?
How do I tell the Media Manager that I have a cleaning tape?
How often should I clean my drives?
What is the expected DLT tape longevity ?
What is the expected 9840 tape longevity?
How do I know if I have a volume that is FROZEN or SUSPENDED?
What's the difference between the media manager expiration date (given by vmquery) and NetBackup's expiration date (given by bpmedialist)?
How do I show how many cleanings are left on the cleaning tape?
How do I show how many cleanings are left on the cleaning tape?

vmchange -M master -h volhost -exp 0 -m mediaid

The command above will tell media manager that the tape should never expire.

vmchange -M master -h volhost -exp mm/dd/yy -m mediaid

This will set it to expire on a specific date.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

When you enter a Cleaning Tape into a library via NetBackup you have to tell it how many cleanings the tape is good for. A new tape is good for 20 cleanings and when that number is exhausted, it will not be mounted again. 

From: "RYAN C. ANDERSON" <RYAN_ANDERSON@udlp.com>

One can make a home-brewed approach to finding out cleanings that are
left by lifting the cleaning data using this command (with DLTs):

vmquery -mt dlt_clean -b
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

From: "Dave Pearson (EML)" <Dave.Pearson@eml.ericsson.se>

We also have an L11000 and haven't been able to get TapeAlert (cleaning when a drive requests it) to work. Sun told us it is a known problem and they have reported it to Veritas.
 
We used to clean our drives every 20 hours until we started getting lots of media write errors. We had all the drives replaced and the problems went away. The engineer said the drives are basically self-cleaning and even if the light does come on it usually goes out when the next tape is loaded. However, he suggested 120 hours as a reasonal interval for scheduled cleaning. So we set it to that and, so far, no cleaning lights and no more errors.
mailto:curtis@backupcentral.com?subject=NetBackup Faq-O-Matic

Tape media longevity can be broken down in to 2 sections:

1) Shelf life; What is the estimated time a tape can retain the data that is 
stored on it.

2) Usage life; How many times can a tape be accessed before it becomes 
unreliable.

Shelf life:
The DLT tape media shelf life is more than 30 years. The environment is 
specified as 18°C to 28°C (68°F to 82°F) with 40% to 60% relative humidity
(non-condensing).
The cartridge should be stored in its protective case to keep out dust and 
contaminants. 

Usage life:
The DLT tape cartridges are specified at one million passes. In laboratory 
tests, the media has been usable even after very high numbers of passes and
the media actually seems to improve with use. At one million passes the tape
should continue to function normally, unless it is being used in an environment
 with heat, humidity or contamination levels beyond those recommended.
A pass refers to a single movement of a section of the tape media across the
head assembly. The number of passes may be determined by the number of times
the cartridge is read or written "end to end". 

In short:
You should be probably more worried regarding if you have still the equipment
 available to read the tapes as opposed to if your tape is still readable.

The training materials I have say that a tape goes into the frozen state when it is first used and NetBackup detects another format already on the tape.

It goes on to say that the frozen state can be "set by command to protect backups from aging out of the images database." I don't know what that means. (Will update here when I find out.)

In my experience I find myself with several tapes in the frozen state that have valid images on them, and I never issued the command to freeze them. (Will update here when I find out more.)
mailto:vandevegt@yahoo.com?subject=NetBackup Faq-O-Matic

Answers in this category:

Answers in this category:
Error 11 with OTM
Status 134 errors with NetBackup 3.4?
Error 14 on NT Clients
Error 24 on Unix clients but WinXX clients work
How fast should my tape drive go? (i.e. what does it mean to "stream" a tape drive?)
Error 41 - bpbkar using 98% cpu
Status 74 on Win2k backup

First off, keep the following conversion in mind:

1 MegaBytes/sec =  3.6 GigaBytes / hour
100  Megabits/sec = 12.5 MegaBytes/sec = 45 GB/hr

If you're not getting at least  "native speed" 
(e.g. 5MB/sec or 16GB/hr from a DLT 7000 drive) performance from your tape 
drives, there is very likely some fixing that needs to be done to your backup
system.

Many tape drives are "streaming" drives, which means that the drive's 
native speed should be considered a MINIMUM incoming data rate.  If one of 
these types of drives is always getting data at a fast enough rate, it's said
 to be "streaming".   If the drive doesn't get data at that speed, it is 
"NOT streaming" and a performance penalty results.

The source of this performance penalty (as I understand it) is basically time 
wasted due to contant head repositioning.  Below is my "quick and dirty" 
explanation of what happens when a streaming-type tape drive doesn't get fed 
fast enough:

     1-incoming data stream begins

     2-tape spins up to write the incoming data

     3-the data buffer is written to tape

     4-look for more data, but it's not arrived yet! 
(the tape is spinning fast enough that we've gotten a bit ahead of ourselves
 at this point)

     5-stop the tape, rewind to the last bit that got written

     6-go back to step 1 (or more likely step2, more data probably arrived and 
piled up during step 4)


the end result is that your tape drive writes even more slowly than the 
incoming stream of data.

The solution:  depends on the problem, but it's probably one or more of the 
following:
    -tweaking the buffer sizes on the media server.  tweaking the netbackup 
data buffers is discussed in this FAQ and on the veritas-bu mailing list.
    -use of multiplexing
    -improvement of network bandwidth into the media server
    -use of FlashBackup (if your backup happens to be slow because of a large 
number of small files)

First off, keep the following conversion in mind:

1 MegaBytes/sec =  3.6 GigaBytes / hour
100  Megabits/sec = 12.5 MegaBytes/sec = 45 GB/hr

If you're not getting at least  "native speed" (e.g. 5MB/sec or 16GB/hr from a
DLT 7000 drive) performance from your tape drives, there is very likely some 
fixing that needs to be done to your backup system.

Many tape drives are "streaming" drives, which means that the drive's native 
speed should be considered a MINIMUM incoming data rate.  If one of these types 
of drives is always getting data at a fast enough rate, it's said to be
"streaming".   If the drive doesn't get data at that speed, it is 
"NOT streaming" and a performance penalty results.

The source of this performance penalty (as I understand it) is basically time
wasted due to contant head repositioning.  Below is my "quick and dirty" 
explanation of what happens when a streaming-type tape drive doesn't get fed 
fast enough:

     1-incoming data stream begins

     2-tape spins up to write the incoming data

     3-the data buffer is written to tape

     4-look for more data, but it's not arrived yet!  (the tape is spinning 
       fast enough that we've gotten a bit ahead of ourselves at this point)

     5-stop the tape, rewind to the last bit that got written

     6-go back to step 1 (or more likely step2, more data probably arrived and
       piled up during step 4)


The end result is that your tape drive writes even more slowly than the 
incoming stream of data.

The solution:  depends on the problem, but it's probably one or more of the 
following:
    -tweaking the buffer sizes on the media server.  tweaking the netbackup 
     data buffers is discussed in this FAQ and on the veritas-bu mailing list.
    -use of multiplexing
    -improvement of network bandwidth into the media server
    -use of FlashBackup (if your backup happens to be slow because of a 
     large number of small files)