Backup fails completely when no more media available

[NHL]

Hi Folks

Netbackup 6.5.5 on RedHat 5.2 x86_64.

We have a backup that occasionally fails because we run out of media in the pool.
Fair enough, but can we make netbackup keep the current backed up data instead of simply failing the backup ?

Thanks
Nick .

[Anonymous]

Keeping the current backed up data, whether on tape or disk, is down to how the policy is configured and your retention period. This defines how long netbackup keeps your backup image data for.

The backup will continue to fail until there is expired media (ACTIVE) in the pool.

Of course add more media into the pool or amend frequency/type of backup based on class: Production vs Development data for example.

[NHL]

The backup policy is set to keep the backups for 3 weeks. There are 3 tapes per week for this particular backup. This backup run just once a week. However if all 3 tapes get used and the backup needs more tapes then backup is useless, the data it has already put to tape is no longer retrievable from tape.

So can I configure netbackup to save what it can, and sort of soft error if no media is available to do a complete backup ?

[Andy_Welburn · Accepted solution]

the data that is saved so far is useless unless the backup finishes successfully irrespective of the reason for the error.

There may be other ways to extract this data, but not via NB.

[NHL]

Ok I understand that then, shame really

Thanks for the help.

[Andy_Welburn]

Are they always empty or can you ensure that they are?

It seems strange that sometimes it works sometimes it doesn't - unless the volume of data saved each time does vary markedly?

[NHL]

Yep it fails sometimes because the data to be backed up can vary quite a bit.

Thanks for your help.

[alazanowski]

If the tapes are NOT multiplexed, you can mount the tar files manually and inspect the files inside. However, if this is a database backup, you might as well not even try.But this again, is outside of netbackup and not necessarily a practical method.

Also, if you enable checkpoints on the policy, you should be able to recover from an incomplete state when new media is available.  This should then start from the last check point that it marked at (depends what you have configured but default is 15 minutes); unless its a database backup.