PureDisk DataPool usage

[AlexeyF]

Hi all

I need to understand what takes space on the PureDisk DataPools because it seems that we consume more volume than expected for the amount of Data that we have.

So I've used the commands adviced in this topic Solved: need to fetch report for all images on a Data doma... - VOX (veritas.com) but that doen't help a lot.

Sure, using this command I can list all the images and their sizes but as far as I unserstand that is "the virtual image size", not physical. I don't see the real space consumed on the data pool by them.

So is there any other solution to determine how much physical space all these images consume?

Thanks

[davidmoline]

Hi @AlexeyF 

Use the post you provided gives the "protected" space. This is what NetBackup thinks is stored in the pool as per the NetBackup catalog. To get the sum of the storage (in KB) use this modified version of the command (after a while it will spit out a single figure):

bpimmedia -dp DiskPoolName -stype PureDisk -l | awk '{ if ( $1 == "IMAGE" ) print $3 }' | xargs -n 1 bpimagelist -l -backupid | awk '{ sum += $19 } END { print sum }'

For the pool use the crstats command to show both the consume pool usage and what the pool considers it is protecting. If the protected figures are way out, then you probably have orphaned images in the pool - which is best solved by opening a support case.

This is the type of output from the crcontrol command:

# /usr/openv/pdde/pdcr/bin/crstats --convert-size 2> /dev/null
Storage Pool Raw Size=4.0TiB
Storage Pool Reserved Space=164.1GiB
Storage Pool Size=3.8TiB
Storage Pool Used Space=1.7TiB
Storage Pool Available Space=2.1TiB
Catalog Logical Size=43.8TiB
Catalog files Count=235909
Space Allocated For Containers=1.7TiB
Deduplication Ratio=26.2
#

The things to note - the Catalog Logical Size should match (reasonably closely to the sum of the output from the bpimmedia output. The Storage Pool Used Space is what is being used in the pool (this will match the file system usage), 

The crstats output can also be obtained by looking at the storage server in the GUI, and examining the properties.

Cheers
David

[AlexeyF]

thank you for you answer @davidmoline 

I have 
Catalog Logical Size=1.5PB
and the sum of all the images gives 1.3PB

not sure how significant is the mismatch.

If there is nothing else to check, we'll open a support case with Veritas.

[davidmoline]

Hi @AlexeyF 

That is a significant difference worth further investigation. If the pool/media server is busy, it is possible housekeeping is not keeping up - and there are ways to process the queue to look to remedy this - the instructions are contained in the Deduplication guide here: https://www.veritas.com/support/en_US/doc/25074086-149019166-0/v35805606-149019166

This may help, but given you are talking about 200 TB of storage more in the MSDP pool, I would suggest opening a case to have support help you identify if there are orphaned images. This process basically reads the MSDP catalog for the pool and compares to the NetBackup catalog and the output will be a list of backup images that are no longer referenced by NetBackup - make sure this is the case before you hit the delete button from the pool. 

Cheers
David