Disk storage unit full in a group

[Tulkas]

Hello all,

Here is my issue.

I have a Disk Storage unit group which contains 6  Disk storage units.

The first DSU is 100% full, but Netbackup saw "66% Full" on the Storage Units view ... and all next jobs are trying to write on this full disk ... error 129 disk full.

Master is netbackup 6.5.4, the DSU are on a media server 6.0.MP4, with BasicDisk. The High water marks is 80%

Any ideas ?

[Jim_Horalek]

Basic Disk units are stupid. You need lots of free space to operate effectivily.  Your only making it worse with your DSUs on 6.0 and your master is 6.5.4. Netbackup does not recover well from a disk full. You need to avoid it at all costs. Netbackup will continue to write blindly to the same DSU if you let it. each time it will try to expire images(to reduce the size)  but if your backup is still to big you will never win. You can't span DSU's!

Use bpexpdate to expire your oldest data if you can.

[Marianne]

What is the Failover Policy for your STUG?

****EDIT *****

If I'm not mistaken - Failover policies is only applicable to 6.5.x STU's...

Hopefully you did not mix 6.5 and 6.0 STU's in STUG config??

[Will_Restore]

• Choose storage units based on the order specified above  (default)
• Choose least recently selected storage within the list

• Failover (Use the first storage unit in the list. Use subsequent for failover only.)

[Marianne]

Thanks for that correction Bill. 6.0 seems like a bad memory from long-long ago...

i'd still like to know what the selection criteria is for the STUG...

[Marianne]

To get back to the original post:

"The first DSU is 100% full, but Netbackup saw "66% Full" "

NBU possibly knows about the space occupied by valid images. Are there perhaps expired images on the disk that was not removed by image expiration/cleanup? Or has someone perhaps copied 'other stuff' to the file system?

You could possibly try the 'cleanup' command :

/usr/openv/netbackup/bin/admincmd/nbdelete -allvolumes -force

(This won't help if there's 'other stuff'...)

[AbdulRasheed]

nbdelete does not help for 6.0 disk media servers. nbdelete works by querying EMM image tables for images which are in 'to be deleted' status. This concept is new in 6.5, not populated by 6.0 media servers.

The reality is that 6.0 does not have nbrmms deamon which is resposible for reporting High Water Mark events proactively.With 6.0, this processing was always reactive, bptm needs to detect the HWM or disk full condition in order to start the cleanup. This is why NetBackup is always trying to use the same storage unit. If you really need to address this proactively, you need to upgrade your media servers to at least 6.5

Warm regards,

Rasheed

[Tulkas]

The Storage unit selection is Failover.

But the 3 first STU are over the high water mark (the %full is false ... 86% Full on the Storage unit view 3% on the Lun ... see attached file)

I'll try first to upgrade every media servers on 6.5.4 from the 6.0mp4 version.

[Will_Restore]

wrong post, sorry