Volumes Reporting Over 100% Space Utilization

[ryanmus124]

Hello -

I'm experiencing an issue where some volumes that have a lot of churn end up reporting over 100% space utilization.  One example is a drive that we are using to store SQL backups.  We're running Windows 2008R2, SQL 2008, and Nimble CS260G 1.4.7.0-45626-opt.

Every night we drop copies of our databases onto the T: drive, which is a 100GB fully thin-provisioned volume.  This process also deletes any backups that are older than 2 days.  When I look at the volume in Windows, we are using 60GB of space.  When looking in the Nimble Array, it reports that we are using 172.31 of 100 GB.

I understand that there's no way for 2008R2 to natively give space back to the array, but I don't understand how it can take up more than 100% of the space that it has been given.  Has anyone else run into something like this?  It's currently not an issue for us but I could see this becoming a problem months/years down the road.

[wen35]

Hi Ryan, This seems a bit odd -AFAIK, we do not allow over values over 100% for quota.  Have you filed a support case with us to see what might be wrong?

Thanks.

-wen

[ryanmus124]

I don't believe we've opened a support case yet, I'll go down that route and update this thread with the findings.  Thanks Wen!

[marktheblue45]

White space? Thin provisioned?

[marktheblue45]

It will be interesting to see what happens if you clear out the white space using space reclaim action with exi cli or another utility. Granted it shouldn't show above 100GB.

[marktheblue45]

Have a look at the "Reclaim Space Quickly" post. I appreciate this isn't the direct issue but could have some bearing on the diagnostic action a subsequent solution.

[ryanmus124]

Hey Mark -

I did go down that route trying to clear up space.  It works as expected on test volumes, but on the volumes reporting over 100% utilization there is only minimal change, it will usually clear up a couple percent the first time that it runs.

I did open a support case as Wen suggested, it turns out this is being actively investigated by the Nimble engineering team.  I'll update the thread when I hear back.

[marktheblue45]

Cheers Ryan, an interesting problem.

[ryanmus124 · Accepted solution]

Update on this.  Sounds like this is a known bug.  This happens when promoting a replica or snapshot and setting its Quota to 100%, it will act as if it has an unlimited quota.  The bug number is 46136 and should be resolved (but not retroactively) in a coming update.

So if you are running into this issue, it sounds like the supported resolution is to create a new volume, migrate the data to it and switch volumes on the host.

I would highly recommend using the 'reclaim space quickly' method (Re: Reclaim space quickly (sdelete alternative)) or sdelete before going down this route, sounds like it's a pretty edge-case scenario.

I'd be happy to provide more details if anyone has requests - thanks all for the input!