Vmware Server Console crashes when removing snapshot

Tonight while doing maintenance on one of my virtual machines I decided it was time to remove a snapshot I created a few days ago while upgrading some software. Shortly after hitting the “remove snapshot” selection my vmware console crashed with a lost pipe error.

Attempting to reconnect yeilded:

There was a problem connecting:
511 Error connecting to /usr/bin/vmware-serverd process

I was almost ready to reboot the machine when painful memories of my last vmware snapshot gone bad experience entered my mind. I investigated further and found that there were .WRITELOCK files in place in my virtual machine folder. After checking my processes on my vmware box I noticed a file called “sdhelper” using a bit of cpu time.

After a brief google search I stumbled across this article which stated that in the event of removing large snapshots it is not unusual for the console to loose connectivity. It is important to simply wait it out and it will eventually finish recombining your disks and restore your connectivity.

I know from prior experience that snapshots are very finicky creatures and rebooting your machine at a time like this could easily lead to data corruption and the loss of your virtual machine.

March 6, 2008 | Filed Under Uncategorized 

Comments

4 Responses to “Vmware Server Console crashes when removing snapshot”

  1. Mark Moreno on April 16th, 2008 1:47 pm

    Thanks for the reference to my article. Glad I could help.

  2. Brian Boos on August 20th, 2008 8:18 am

    VirtualCenter has a hardcoded timeout of 15 minutes for service console operations. For large snapshots or slow machines, this can easily be exceeded. Fortunately, because this type of operation runs on the ESX host, it will continue to run. If you connect your VC client directly to the host (as opposed to the VC server), you’ll be able to see the task still running (and probably sitting at 95%). If you get worried and are wondering what the host is doing, you can keep an eye on the VM’s folder in the VMFS file system. You’ll see the modification dates of the VMDKs change as the snapshots are being committed.

  3. Duane on May 20th, 2009 6:43 pm

    Exactly. I got the same pipe error and would have been more alarmed if not for reading this post.

    I waited it out. Took a long time. Completed successfully. Good tip on seeing the VMDKs change (drop out of the file system) as they are processed.

  4. Brendan D on December 3rd, 2009 5:09 pm

    Jeremy,

    Thanks to your post i avoided rebooting my vmware server and spending 12 hours rebuilding the corrupted VM. Thank you, Thank you, Thank you :D . After waiting 15-20min sdhelper disappeared and everything finished and i was able to connect normally using the console again.

Leave a Reply