Find Nimble
I was looking at the volume usage in both the array GUI and in Infosight. Both showed wildly different numbers and I found no explanation for the numbers so I logged a support case. Here is the run-down for a VMDK on VMware VVol.
Aug 31, 2019 The Nimble Setup Manager is part of the Nimble Windows Toolkit and can be downloaded from InfoSight. If you do not have an InfoSight login yet, please register as a new user. Note: You will need your array serial number to register successfully. The Nimble Setup Manager is part of the Nimble Windows Toolkit and can be downloaded from InfoSight. If you do not have an InfoSight login yet, please register as a new user. Note: You will need your array serial number to register successfully.
First the Infosight numbers
Volume usage at 6.9 GiB is the logical space used in the VMDK, before dedupe and compression, out of the 80GiB provisioned.
Snapshot usage at 5.6 GiB is the logical space used by the snapshots(ie backups), before dedupe and compression.
Simple enough, on the array GUI.
The names are the same, but the numbers are not!
Volume usage is suddenly 2.7GiB. This is physical space used after compression.
Snapshot usage at 2.3GiB. This is also after compression.
Total usage at 5.0GiB, the sum of volume and snapshot usage. Easy enough.
Compression savings at 7.6GiB. This is what compression saved you of physical disk.
Total usage at 5.0GiB plus compression savings at 7.6GiB is 12.6GiB.
Back to Infosight, the sum of volume and snapshot usage is 12.5GiB.
Hey! What’s up with the 0.1GiB difference?
Infosight is based on the daily logs, so it lags a bit behind. Which might cause the difference.
What about deduplication? It is run before compression, but it is not shown on the volume level, you can only see it on the array level.
Let’s check one more.
Ok, 10.7TiB in Infosight.
9.3TiB + 1.4TiB in the array GUI, so 10.7TiB in total. Perfect! Just the same as in Infosight.
Still asking about deduplication for the volume? Sorry, you just can’t see it.
Find Nimble Serial Number
My 2 cents? I’m not 100% convinced that compression savings in the array GUI is just from compression, even though that’s what support said.
Why? When checking the array level data reduction savings on where this last VM is stored I see a savings due to compression at 937GiB.
The VM we just looked at had a savings from “just compression” at 1.4TiB. It doesn’t make sense that the array itself has huge drop in savings from compression compared to one of the VMs it contains.
It is also a bit suspicious that this last VM in Infosight is at 10.7TiB after savings(both dedupe and compression), while the array GUI claims it is 10.7TiB without dedupe savings. So zero or close to zero savings on this VM for dedupe? A bit odd.
Nimble Find Unmanaged Snapshots
Well, now you have it. What I got from support and my personal comment. Feel free to add your own comments below.