What is logical-space-in-use for volume ? We see many volumes have this value greater than
"vol-size" ? We are using REST API to get this information.
This is a little clearer in the CLI or GUI where the term for this value is called "Space in Use (VSG)".
Here's a write-up I did last time I was asked about this value...
The “VSG” part stands for “Volume Snapshot Group”, so this refers to the combined space in use for all of the volumes in the snapshots group related to that volume (ie, the volume itself, plus all snapshots taken from that volume).
The “space in use” part refers to the pre-dedup/pre-compression space used by the volume/snapshots. In effect, it’s how much of that volume you’ve actually written data to, or to look at it another way, how much of the volume ISN’T thin provisioned.
An example will probably work best. Say you create a 100GB volume. Initially this will show as a VSG-Space-In-Use of 0GB as you’ve written nothing to it (it’s 100% thin provisioned)
You then write 10GB of data to the volume. VSG-Space-In-Use will then show 10GB, even though this might actually only be using 5GB on disk due to dedup/compression.
If you then take a snapshot of the volume, the VSG-Space-In-Use will still show as 10GB, as although snapshots give you a second virtual copy of the data, then no actual additional space used by them (initially).
If you then write 5GB to the snapshot, and another 5GB to the original volume, then the total VSG-Space-In-Use will now be 20GB – (10+5)GB written to the original volume, plus the additional 5GB written to the snapshots. Again this is before dedup/compression, so that 20GB might still only be using 5GB of actual disk space (eg, if the new data written was actually all duplicates).
If you were then to delete the snapshots, the VSG-Space-In-use would drop back down to 15GB, as the 5GB of data written to the snapshot would be freed up.
This also means that VSG Space In Use can be higher than the volume size, as in effect it's counting across both the volume and all of it's snapshots. For a 100GB volume with 1 snapshot, it could be as high as 200GB (100GB for the volume, and another 100GB for the snapshot). With 2 snapshots it could be as high as 300GB (100GB for the volume, 2x100GB for the snapshots), and so on.
Numbers this high are unlikely, as it would require a 100% change rate between copies - but with enough copies it's not uncommon to see a VSG-Space-In-Use value higher than the volume size.
Thank you for the detailed explanation. I have some data which I am trying to anaylze based on the explanation. From what I understand, it doesn't seem to be possible to get to the used capacity just by that volume, is it?
If the "logical-space-in-use" always gives added up capacity for the volume (be it snapshot or original volume), there is no way to deduct just the snapshot space in use, is that correct?
That is correct - at least for X1 (In X2, with version 6.1, we're changing this a little - but unfortunately those changes won't be available on X1 arrays)
There is one way to tell (roughly) how much data a snapshot is using - and that is to delete it (or refresh it)!!
Whilst that may sound like a bit of a silly solution, odds are at some point you're going to delete or refresh the snapshot anyway - so use that as an opportunity to see how much the VSG-Space-In-Use drops when you do that, and that will tell you how much space it was using (before dedup/compression). Note that depending on the size of the delta between the volume and the snapshot it might take a few minutes for it to cleanup everything, so you might have to wait a little for the space to drop back.
You can even use the reporting on the array to look at this for a previous occasion that you deleted/refreshed a snapshot - create a report looking at the VSG-Space-In-Use for that volume over the time period where you did a delete/refresh, and it should be fairly easy to see the drop in space in the report.
Thank you. You have mentioned that for X2 arrays you are changing a bit. What are the changes? We have X2 arrays as well and was wondering is there is something which we can use to report the LUN used for those without the snapshot capacities.
Yes, the new feature in X2 works for both Snapshot and regular volumes. Have a look in the XtremIO 6.1 Users Guide (available on support.emc.com), specifically the section headed "Calculating a Volume’s Data Reduction Metrics"