VMware Cloud Community
unsichtbare
Expert
Expert

Long time to delete snapshot (4+hours) plus long snapshot Stun

Hi,

I am working with a VM that is presenting problems with snapshot delete and Stun time. The snapshots are taking 4+ hours to "delete" (I still think the correct word is "commit") and the Stun during this process is enormous. The reason for the snapshots in the first place is backup.

Here's the rub:

  • ESXi is 5.1
  • VM is 900 GB on 2 disks (2 SCSI controllers too)
  • VM has 8 CPU and 12 GB RAM
  • snapshot data has been redirected to its own LUN on the same SAN (only have the one!) [workingDir = "new_path_locationsnapshot.redoNotWithParent.= true]
  • NtfsDisableLastAccessUpdate=1
  • VSS is disabled in VM and VMware Tools
  • VM is a G.I.S. server that scans for updated files/images, compares and then updates files. Its database is elsewhere and it has no problems with snapshots.
  • Unfortunately, RAID 6 is set for the entire volume on the single SAN in the environment

What else can I do to get this to back-up correctly in vSphere?

THX:smileyconfused:

+The Invisible Admin+ If you find me useful, follow my blog: http://johnborhek.com/
0 Kudos
1 Reply
craigcrack
Contributor
Contributor

Just leaving a note for anyone else who is looking for an answer and ends up here like I did.

(I assume unsichtbare has long ago sorted his particular issue.)

I just spent an afternoon on this and found a leftover snapshot that had blown out to half a GB which made the commit time very long, which caused auth issues when the clock jumped forward 15 minutes or so at the end of each commit. (Flexlm hates that FYI)

For me

Takeaway /1

>> In vSphere, Inventory, Datastores and Datastore clusters;

>> select the "Datastore" then check the last tab called "storage views" .. scroll down along the "snapshot storage" column. Anything that's not 0 or 45 kb indicates a snapshot.

Takeaway /2

Check this manually or scripted, however you normally work on some sort of regular basis. ("veeam one" if you have veeam)

Rule-of-thumb is all snapshots should be cleared before backups begin.

While it's vmware function (or just plain IO but the OP and my kit should both be fine), the veeam guys have it covered in this --  Community Forums • View topic - Snapshot removal issues of a large VM   --post. it's 13 pages as of Aug2013, skim backwards I guess.

0 Kudos