Nimble automatic snapshot while volume extension

[rashid0709]

Dear Experts

I have extended nimble volume presented to a windows server. After volume extension, I have noted that Nimble array has created an automatic snapshot to avoid any crash during the volume extension.

And I have noted that snapshot usage space is increasing day by day. Suppose if  I remove the snapshot now, what happens in background ?

Where will save the data written from snapshot creation date to snapshot removal date?

How Nimble snapshot works in background once it is created as part of volume extension? 

Where will commit the new write ?

is it similar like VMware snapshot ?

Snapshot reserve: 0%
Snapshot warn level: N/A
Snapshot quota: unlimited
Snapshot count: 1
Volume usage (MiB): 2544191
Volume compression: 1.56X
Volume space saved (MiB): 1431045
Snapshot usage including pending deletes (MiB): 2933589
Snapshot compression: 1.56X
Snapshot space reduction: 3.05X
Snapshot space saved (MiB): 6003224
Read only: No
Multi-initiator: Yes
Thinly-provisioned: Yes

[Sheldon Smith]

A snapshot points to the volume data locations at the time of the snapshot. All new writes are to the parent volume.
If you remove the snapshot, old data is dereferenced. When a data location's count reaches zero, the garbage collector cleans up.

I still don't understand why VMware snapshots work the way they do, but basically, no, it is not like VMware snapshots.

[rashid0709]

I didn't understand what do you mean by dereferenced.

As per the Nimble Admin Guide, my understandings are given below.

1. If you have 100GB volume and 10GB data is written on the volume and if we take a snapshot, the snapshot usage space will be Zero snapshot usage.
2. After first snapshot if you have written new block of 20 GB data, still snapshot usage space will be Zero
3. But incase if you have over written the initial 10GB data continuosuly on 5 days (for example daily data change rate= 5GB) then after 5 days snapshot usage space will be 25GB = 5days X 5GB
4. And if you delete/remove any NSs snapshot, this will not effect any production data overwritten to the volume in between snapshot creation date to snapshot removal date. The production data will be safe.
5. If you look at my first post, snapshot usage space is very high because data is overwritten daily. So if i delete the snapshot, my new data overwritten will persist on the volume.

If my understandings are not correct, then please correct me.

[Sheldon Smith]

It reads like you are still thinking in terms of logical volumes, LUNs, being tied to specific physical disks. Volume X is on RAID set 1, Volume Y is on RAID set 2 and so on.
As new data is written to a Nimble array, it goes to the next sequential locations and the logical volumes simply point to the new locations.

There are a few good videos that may help clear things up. Start with these two:

1. Storage Snapshots - HPE Nimble Intro & Admin Training
   https://www.youtube.com/watch?v=-SGFt632Jos
2. HPE Nimble Storage – Snapshots, Clones & Copy Data Management Lightboard
   https://www.youtube.com/watch?v=-GVhPJ0DRbM

[rashid0709]

---

@Sheldon Smith wrote:

It reads like you are still thinking in terms of logical volumes, LUNs, being tied to specific physical disks. Volume X is on RAID set 1, Volume Y is on RAID set 2 and so on.
As new data is written to a Nimble array, it goes to the next sequential locations and the logical volumes simply point to the new locations.

There are a few good videos that may help clear things up. Start with these two:

1. Storage Snapshots - HPE Nimble Intro & Admin Training
   https://www.youtube.com/watch?v=-SGFt632Jos
2. HPE Nimble Storage – Snapshots, Clones & Copy Data Management Lightboard
   https://www.youtube.com/watch?v=-GVhPJ0DRbM

---

Dear @Sheldon Smith  

You have mentioned that As new data is written to a Nimble array, it goes to the next sequential locations and the logical volumes simply- I have a question at here, then why snapshot usage space is increasing day by day ?

[Sheldon Smith]

The snapshots provide a point-in-time (T0) image of the volume. As data is changed and new data is written to the volume, any data from time T0 must still be preserved until the snapshot is deleted. The volume doesn't need it anymore, it's got new and improved data. The "obsolete" T0 data is assigned to the snapshot. Therefore the snapshot slowly grows.

[Muru123]

HPE will keep your case 5356084072 open for next 6 hours. We shall go ahead and close the case if there is no response from your side.