Volume Space Used Larger Than Volume Size?

[#1]

Here's the odd question of the day.

I've got this one volume that is showing some weird stats where the used space is greater than the volume size.

CS220 is running NOS 2.1.4.

Nimble volume - 70 GB

ESXi DataStore - 70 GB

Windows volume size created on datastore - 60 GB

The volume is simply a file store on an SFTP server so there's a lot of data churn going on over time.

Volume Info:

Performance policy: VMware ESX

Block size (bytes): 4096

Reserve: 0.00%

Warn level: 90.00%

Quota: 100.00%

Snapshot reserve: 0.00%

Snapshot warn level: N/A

Snapshot quota: unlimited

Snapshot count: 2

Volume usage (MB): 109350

Volume compression: 1.02X

Volume space saved (MB): 2659

Snapshot usage (MB): 12748

Snapshot compression: 1.03X

Snapshot space reduction: 14.45X

Snapshot space saved (MB): 171507

How can a volume sized ~ 70 GB and showing a usage size of ~106 GB be saving 2.6 GB of space?  Is there any way to 'fix' this oddity other than create another volume and just move the data?

[tmoore106]

Randy are you seeing the volume usage stats in the web gui or Infosight or both?

[#3]

I see the same stats in the Web GUI, InfoSight, and CLI.

The downstream replication partner has more normal stats, showing 58 GB used on the 70 GB volume.

[tmoore106]

Are there any orphaned snapshots?

[#5]

No orphaned snapshots.  I only snap once a week and keep 2.

[tmoore106 · Accepted solution]

I would do two things.... 1. Run the iSCSI UNMAP on the volume from an ESX host, see if anything changed, then if not, 2. call support so they can look closer.

[#7]

Todd, thanks for the tips.  I've run through the esxcli storage vmfs unmap command earlier with no luck.  I guess I'll see if I can 'stump the support tech' next.

-Randy

[mkieran59]

Hey Randy,

Were you able to get your issue fully resolved?

Thanks, Michael 

[#9]

I ended up opening a support ticket to solve this one.  Thank you Josh@Nimble.

Cause - not exactly sure.  Somewhere something got out of sync and locked a bunch of blocks on the volume that can only be cleared by deleting the volume.

Solution - Storage vMotion the volume to another datastore.  Unmounted and detached the offending volume following VMware best practices and then deleted it from Nimble.  Recreated the volume on my Nimble, created the datastore and then vMotioned the vmdk back.

- Randy