10

Last night I took a snapshot of a VM before adding a new disk and running an offline defrag of Exchange (using the newly added disk as the temporary path).

When I woke up today the server was offline, and when I opened vSphere I saw the message "Configuration Issues - Virtual machine disks consolidation is needed". I answered the question and chose Abort/Cancel. The VM then booted. The Exchange database is now ~160GB (down from about 240GB), which is about the size I expected it to be after the defrag. That made me suspect the defrag may have completed, but alas it wouldn't mount.

At this stage my preference would be to revert to the snapshot. Has anyone been in this situation before? I've done lots of reading but have only found articles explaining how to fix the "needs consolidation" error by consolidating. I can't find anything about deleting the unwanted delta disks to revert to the original.

Is there a supported way to delete these unwanted delta disks without consolidating them?

Could I simply move the diskname-000001.vmdk files elsewhere then boot the VM? (I also have some files with the .REDO_nSGail extension that I assume I would need to move too). The original .vmdk files will be in read-only mode though - is there a way I can set them back to read-write?

Thanks to anyone that can offer some advice.

Other info: - ESXi 6.0 - VM is a Windows SBS box (which has both Exchange and AD). It also has about 1TB of data and not enough free space on the data store to restore it all. I'd have to move off all the VM files first, then restore a backup which would take many hours so I'd rather avoid that if possible.

2
  • 2
    Damn. That's a terrible mess. And you don't feel like you can fix the present issue with the database not mounting?
    – ewwhite
    Dec 27, 2015 at 0:07
  • 1
    I've had to recover VMs that have snapshots manually before. You will need to point the VMDK file to the flat file instead of the delta disk, thus bypassing it. More info:link
    – McKenning
    Dec 27, 2015 at 3:38

1 Answer 1

1

To revert to the original, you simply pick "revert to current snapshot" from the right click menu, assuming you only have the one snapshot.

In VMware, you can do two things with a snapshot (I'm simplifying somewhat);

1) go back to "prior state" (the way it was BEFORE the snapshot), which is the "revert to current snapshot". In this approach the term "current snapshot" is kind of misleading but think of it as the state before the snapshot.

2) Delete the snapshot which accepts all of the changes you have made since taking the snapshot and rolls them into the running VM. Sounds like you do not want to do this.

"Consolidate" really has nothing to do with which choice you want to pursue; consolidate simply consolidates redundant logs or .vmdks from failed past processes and has nothing to do with the disk changes. For example, if a snap fails, you an have redundant logs and perhaps .vmdk. "Consolidate" would clean them up, probably by removing the redundant ones. Consolidate won't change the number of snapshots you have. It simply cleans up partial or duplicate logs and/or .VMDK from failed prior operations.

So, in theory, if you want to go back to "prior state" you would "revert". It will discard the delta changes and go back to that pre-snapshot state. "Consolidate: Merges the hierarchy of redo logs. This is available in vSphere 5.0 and later." from VMware website.

Of course, considering your situation, all of the normal caveats would apply, have backups, etc., etc.

Might read this for more information:

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .