MSDP has no images yet 94% full

[phscott]

I encountered a situation where the NBU 5220 I was replicating to via AIR was 94% full, even though I had expired all of the images from it.  For expediency, I just blew away the disk pool and re-created it to get the desired results.  My question is: what do I look for in the future because I know this will more than likely happen again?

I did do the 'nbdelete -allvolumes -force' to reclaim the space after the images were expired, but to no avail.

Thoughts?

[sdo]

Don't use -force.   This trims NetBackup catalog without ensuring storage is released, and then never again tries to remove fragments from storage.  i.e. using -force can cause your problem.

What version of appliance was it running?

Did you run the usual garbage collection manually?

I'm trying to chase down a similar thing in a v2.5.2 N5220.  It shouldn;t have anything on it, but it's 73% full, and refuses to relinquish space, even though the PureDisk object counts are very low.

[phscott]

I don't do garbage collection since this is the first I've heard of it.

Good advice on -force.  I'll make sure it doesn't get used in the future.

I am running 2.6.0.4.

Luckily, this is the DR side, so blowing the entire MSDP away and recreating it wasn't a monumental event.

[watsons]

I would investigate why the DR site disk pool becomes full first - whether there is a need to increase the storage size, or the image expiration simply did not work. Were there any orphaned images? Is queue processing running without issue or has been disabled? 

Checking the last few image cleanup job should give you the idea of how much spaces had been cleared, and if the same is reflected in the actual storage consumption. Make sure replicated images are not all set to INFINITY in target site, and there is no other mechanism set to avoid expiring data (such as NOexpire touch file).