When restoring snapshot it consumes a lot of space FAS 2020

[#1]

Hello,

I have a FAS 2020 with DATAOnTAP 7.2.6.1

My problem is as follows:

I have a 180GB Volume and 130GB lun on it.

I created a snapshot in SnapDrive 6.2 for Windows it was about 5MB, then I created second snapshot.

I want to revert to first snapshot so in SnapDrive window I choose that snapshot, right click on it and select "Restore DISK" 

Restoring begun, I noticed that during restoring operatrion space available for data was rapidly decreasing. It almost consumed all space on the volume.

Currently it uses the snapshot reserve and almost all data space.

Is this behavior normal, or it is something wrong?

Best regards,

Sławomir Dąbrowski

[#2]

SD GUI will basically duplicate the whole LUN from snapshot, so yes, observed behavior is normal.

If you want to do full volume revert to previous snapshot you can do it using SD command line utility (sdcli.exe). Do not forget that reverting volume will irrecoverably delete all snapshots after the one you revert to.

[#3]

So you say when I want to restore lun to a specific snapashot i should use SnapDrive Command line ?

Ok I will check that and replay here.

Thanks!

[#4]

LUN restore via GUI basically does LUN cloning in the background (instead of volume-level snap restore), so existing snapshots are not removed (they would be by snap restore).

Hence existing snapshots may grow, hence available space may be consumed.

[BTW - cool nick ]

Regards,

Radek

[#5]

I was trying to revert to a snapshot from a command line by executing the command:

sdcli snap restore -d z -s SnapShotName

but the situation remains the same

Here is some Syslog:

Severity        Event        Triggerred        Message
info         wafl.volume.snap.autoDelete        Tue Apr 26 13:32:48 CEST        Deleting snapshot '26-04-2011_HOUR_13_00' in volume 'HyperV_SQL' to recover storage
info         wafl.volume.snap.autoDelete        Tue Apr 26 13:32:33 CEST        Deleting snapshot '26-04-2011_HOUR_12_39' in volume 'HyperV_SQL' to recover storage
info   app.log.info        Tue Apr 26 13:30:27 CEST        R710NODE1: SnapDrive 6.2: (198) Snapshot Copy event: For the description of this event please check the application event log on the host system.
info         lun.clone.split.started        Tue Apr 26 13:30:20 CEST        Clone split was started on LUN /vol/HyperV_SQL/HyperV_SQL.
notice         lun.snaprestore.notice        Tue Apr 26 13:30:10 CEST        [74577978, 3369100, 61491606] snap restore: Clone/Restore could not set space reservations on restored LUN
notice         wafl.vol.full        Tue Apr 26 13:30:10 CEST        file system on volume HyperV_SQL is full
notice         lun.snaprestore.notice        Tue Apr 26 13:30:10 CEST        [74577978, 3369100, 61491606] snap restore: started

A also posted screenshot from my system manager representing this lun configuration (Screen3.jpg)

Why is that one snapshot is so big ?

What command should I use to avoid such situation ?

Regards

Sławomir Dąbrowski

screen3.jpg ‏214 KB

[#6]

If you want to revert volume to snapshot you need to use “sdcli snap restore_volume”, not “sdcli snap restore”.

[#7]

Assumming of course that I have one Lun per Volume. But what if I would have several luns on the same volume and I want to recover only one lun not entire volume, then should I use "sdcli snap restore"?

[#8]

That’s correct. But then it makes no difference to GUI – both do exactly the same operation.