Forum Discussion

Dan_Vargas's avatar
Dan_Vargas
Level 4
10 years ago

Checkpoints and calendar based backups with big execution windows

Hi all,

 

I'm having the same issue as the guy over here, who couldn't have it resolved: https://www.veritas.com/community/forums/netbackup-checkpoint-calendar-schedule-failes-netbackup-7x

Basically, a very big and slow full backup takes 35-40h to complete. Last time it failed after 36h and it restarted from scratch, not using any checkpoints. It started on a Saturday early morning and failes on a Sunday afternoon.

According to the admin guide, checkpoints are ignored in the following scenarios

 

1. If a new job is due to run, or, for calendar-based scheduling, another run day
has arrived.

2. If the time indicated by the Clean-up property, Move backup job from
incomplete state to done state, has passed.

 

This policy is MS-Windows and thefragment size for the robot (this is a backup to tape) is set to 50GB.

I believe 1. is my case. I changed 2. to 6h (from 3) and the backup job is retried 2 times every 3 hours.

2 Replies

  • Oops, hit the publish button thinking it was the save one.

     

    I was saying that I changed (2), but I wonder if there is any way around (1)? Maybe configuring it to run every x weeks instead? I'm not a fan of that as it's easier to remember that 3rd fridays of the month are full backup fridays, but if it is what it takes I'll just suck it up.

  • Hello,

    from my experience, using calendar-based scheduling is more rare then frequency based scheduling. If you would convert to frequency based schedule (with frequency larger > 2 days at least here), you would not encounter this issue.

    But I admit there could be situations where calendar-based scheduling is important. Consider if "third Sunday in a month" is important is some way in your company for full backups, or if full backup can be simply taken "every 4 weeks" regardless a certain week. Depends on operations on source data.

    Regards

    Michal