2015-05-19 08:48 AM
I'm trying to solve a problem with SMVI, dedup, & vmware.
Filers: FAS3210, 7-mode 8.1.3
LUN was created with provisioning & cloning utility.
My problem is that we had a surge in VM usage that required me to deploy another VMware LUN to our environment. After 6 or so months, our datastore usage declined to the point that the lun was no longer needed. I've migrated all the VMs off of the LUN, but due to the still valid SMVI snapshots living on the volume, I'm going to leave the lun empty, but mapped, until those snapshots age out(our retention for monthlies is 3 months.)
Now, after a week or so, I check the storage efficiency chart in system manager. it is indeed showing a decline in datastore usage: the LUN is provisioned at 500GB, but the used data space portion of the graph shows 218GB used. But there isn't 218Gb of data on the datastore: it's sitting at about 5MB used(the .vSphere-HA folder.)
So my question is, why hasn't dedup squashed that down farther? It's obviously not ignoring the thin provisioning of the lun, or it would be at 500GB used: the % used field in the LUN screen shows 100%, as does the "before" graph under storage efficiency.
I'm hoping I can find an answer here, that I can then apply to our other LUNs: I recently found & deleted a large amount of orphaned VMDKs in our environment, & want to make sure the space I cleared is being given back to the aggregate, as it is showing quite full at the moment.
2015-05-19 11:57 AM - edited 2015-05-19 11:58 AM
The graph seperates data usage from snapshot usage. It's the green part of the graph that I think should be smaller. edit: the DARK green part, that is.