bpstart_notify script execution

[somesh_p]

I need a help. Does bpstart_notify script execution requires a resource (tape drive) in netbackup? I think it requires a resource. Is there anyway that we can execute the bpstart_notify regardless of availability of resources

[Mark_Solutions]

bpstart_notify is used on a client prior to the start of a backup job

So when a job starts it runs  this first - it is not dependant on a resource but of course the backup job itself would fail without one

Hope this helps

[Marianne · Accepted solution]

Please help us to understand what exactly needs to be done with bpstart_notify without needing a backup resource?

I do not think that backup will go to Active state without resources identified and assigned. Client connection and bpstart_notify happens way later in the process flow.

You may be better off executing the client script using the OS scheduler, such as cron.

[somesh_p]

On further debugging, we noticed that it is not bpstart that is occupying the resource. It is occupying resource to "Transfer BMR information to Master server". What is the significance of ticking the below option in policy attribute

"Collect Disaster recovery information for Baremetal Restore"

Our client operating sysem is Solaris 10. We have this policy setup for data file systems (not for ROOT file systems). Can we deselect "Baremetal Restore". What is the impact by doing so ??

[Marianne]

If you do not have BMR configured in your environment, then please de-select the option.

There is no use for this option if you do not intend using the feature.

For more info on BMR, see the BMR Admin Guide: http://www.symantec.com/docs/DOC5163

[somesh_p]

sorry.. I was wrong. bpstart_notify looking for resource regardless of  Baremetal Restore option selected.  Now, I am back to original questions, How can we make bpstart_notify to run regardless of availability of resource ? We have a TCP socket program included with in bpstart_notify script, which monitors the backup preparation on both nodes of HA server. Backup will start only after comlpetion of preparation on both nodes. We are facing an issue if one node gets drive and other couldn't get drive, Backup is indefinitely waiting for preparation on other node

[Marianne]

Please have another look at my post above: 

https://www-secure.symantec.com/connect/forums/bpstartnotify-script-execution#comment-8745261

[somesh_p]

We have an application running in HA environment, which has SAN and NAS based file systems. Usually, application will running on one host (admin1) and sybase database will be running on other host (admin2). SAN file systems are vx mirrored file systems and NAS file systems are NFS based.. Backup preparation will perform mirror detach for vx file systems and mount the detached plex under /Backup/<dg_name>/..... Snapshot will be made on NAS file systems and mounted under /Backup/ombssnap/... To perform backup, policy will be created with both nodes of HA systems and include all such detached/snapshot file systems (/Backup/..) in backup selection. Daily_Incr backup of this policy is setup as frequency based with frequency of once in a day with 12 hrs backup window. Weekly_full backup of this policy is setup as frequency based with frequency of once in a week with 12 hrs backup window. Backup shall start for both nodes of HA at the same time in order to maintain the integrity between sybase database and OSS application. We have TCP socket program which is included in bpstart script, which monitor thru the backup preparation that includes sybase quiesce, detach SAN file systems, snapshot of NAS file systems and mount these file systems under /Backup/... directory. Once this backup preparation is done, control will be given to netbackup to perform backups of each file systems under a separate stream. Upon successful completion of backup, these file systems will be synced back with help of bpend_notify script. Default policy settings will be as follows

Allow multiple data streams is ticked
Follow NFS is ticked
Daily Incr: Frequency based with once in a day scheduling from Monday thru Saturday with 12 hrs backup window
Weekly Full : Frequency based with once in a week scheduling on Sunday with 12 hrs backup window
Enable multiplexing is not be set on storage unit properties

Ask me if you need any further details

[Mark_Solutions]

If you use allow multiple data streams in the policy then the bpstart_notify will run for each stream and so you will need to add extra scripting into it to allow for that and prevent it running mutiple times.

You will also need to make sure that the script does close out otherwise the backup will not start

You also need to make sure it either finishes within 5 minutes (default timeout) or you need to increase the timeout value in the Timesouts section of the Master Servers host properties

For the script you could add a line so that it runs only if the stream number is 1 as it does pass the stream variable to the script (%STREAM_NUMBER%)

Hope this gives some ideas - maybe you just need it to run on the one node but control things on both nodes?

Alternatively you would need it to run on the Master Server to control remotely everything from there using the parent_start_notify script which does only run once during a backup - especially important when snapshots are involved as it can try and do those before it runs the bpstart_notify script

Hope this helps

[Marianne]

Marks is 100% right about multistreaming issue with notify scripts.

Looking at this statement:

We are facing an issue if one node gets drive and other couldn't get drive

Is there any possibility that you can fix this with MPX setting in STU and schedule?

Due to the complexity of your process, I still feel that you will be better off to schedule the entire pre-processing tasks, followed by bpbackup command using an external scheduler.
Even OS cron will be a better fit that NBU scheduling in this instance.

Have you ever thought about NBU snapshot client option for the SAN based backups?