2010-08-02 01:20 PM
I have a question about restoring a VM from the snapshots share on the FAS2020. We initated a restore of a single VM folder with a single 35gb VMDK file within it as the largest file. The restore time stated it would take 140 minutes to restore. It’s a snapshot restore from within the same FAS2020. I believe the restore (snapshot time) is from a few hours ago. The restore was initiated the restore from a workstation using the fas2020\.snapshots share.
Is that length of time normal for a restore?
We were a little surprised it would take 140 minutes (or at least state that it would, I don’t know how long it actually took yet.) to restore a single 35gb vmdk file from within the same FAS2020.
Maybe that is normal? Looking for an experienced persons perspective.
I asked someone else about this and they were a bit vague:
I would have mounted the volume from a Linux or unix system rather that using a windows system. This way there is an actual windows file copy operation taking place "outside" the datacenter. When doing file restores that is how I would say to restore but for VM restores I would do it a little different. That is going to handle the restore from the workstation API and is going to take some time unfortunately.
Could someone please help clarify?
2010-08-02 10:31 PM
If it was a single file restore (snap restore –t file) – yes, it takes long time. Full volume restore (snap restore –t vol) is almost instant.
If you have flexclone license you could rather try file clone (see clone command). Like volume clone, destination is available immediately and any long running tasks are started in background.
2013-09-26 05:18 PM
Single File restore takes longer because it has to check all active maps across the snapshot. After restoring single-file the system has to look at all snapshots to see if it can free the blocks in the file.