Video Screencast Help

parent_notify_script issue (NetBackup, Solaris 10 master server)

Created: 10 Dec 2012 • Updated: 13 Jan 2013 | 2 comments
PWW's picture
This issue has been solved. See solution.

We have two Exchange 2007 Server clusters in two different locations (two separate policies).  Our disk storage is tight in both locations so I created a script for each location that deletes an unexpired backup image when I need to free up more space for a new backup. I call my image_deletion script in the NetBackup parent_start_notify script for these jobs so I can avoid  Status 129 (disk space full) errors.  In one location this setup is working flawlessly, backups always successful, images deleted as needed.  

In the other location sometimes the backup fails due to disk space being used by other backup polices that use the same disk space.  When that happens we get a Staus 129 error and the job retries again and again.  But even though the retry attempts show up as new jobs in the Activity Monitor, the parent_notify_script does not seem to run again, so my image_deletion script does not get called to free up the required free space.

Isn't it normal for the parent_notify_script to be run each time a backup is begun, even if the job is a retry of a previously unsuccessful backup job?

Or am I misunderstanding how NetBackup works?

Comments 2 CommentsJump to latest comment

CRZ's picture

The parent_start_notify script will ONLY run once, when the parent job starts.

parent_start_notify on UNIX/Linux
NetBackup calls the parent_start_notify script each time a parent job starts.

Symantec NetBackup 7.1 Administrator's Guide for UNIX and Linux, Volume II (pages 166-7)

Your new jobs are new siblings of the same parent job, but that parent job doesn't start again - it's already started.

I'm probably explaining it poorly.  Does this help?

Also:  Is there a reason you're not using bpstart_notify (or bpstart_notify.policy or bpstart_notify.policy.schedule) instead?

[See pages 153-5 of the guide linked above for more info on bpstart_notify.] | APPLBN | 761LBN

PWW's picture

I did not realize retries of failed policy/schedule backup attempts are under the same parent.  They show in the Activity Monitor with a dash ("-") in the schedule column just as a new parent backup job does.  

As far as using a version of bpstart_notify, I was concerned that my image_deletion script can take a minute or two to complete and that was optimally started with the parent job rather than the client job.