Dealing with VM UUID changes

[rowl]

I just encountered this message and was wondering what the definitive best way is to deal with this.

ERROR ANS1715E A filespace already exists for VM (vm001), but with a different VM uuid (421a3b7a-d496-23c9-da89-b6b01c010de2) than the current one (564d2bba-6f40-0465-8f2a-d1d179669fb1).

I have found several recommendations online
- Rename the file space: This gets rid of the error, but now I don't see how I can restore data from the original file space. When I go into the client to mount the -OLD virtual machine name I am prompted that "No snapshots are available for the selected virtual machine. Or, the authentication node is not authorized to restore this virtual machine." Not sure what to make of this, it is still a file space under the same node name as it was before, it just had -OLD appended to the original name during the rename process.
- Delete the file space: This also gets rid of the error, but deleting all the backups is not a viable option.
- Enter testflag VMBACKUP_UPDATE_UUID into the dsm.opt file: I tried this, the error messages didn't go away but it looks like the backups ran.

Obviously I would prefer that the backups run successfully and we don't get any errors. Our internal VMWare support team told me that I should expect to see UUID changes from time to time for various reasons.

Thanks,
-Rowl

 

[fr4g3r]

Hi,

I only know the following reasons that lead to a UUID change of the VM:

- The VM was moved with the option "I copied the VM" and not with "I moved the VM"
- The VM was restored under a new name

Is one of the above reason leading to the UUID change?

We had this problem too when someone restored a server under a new name.
He then wants the restored server to be the active one, deleted the old one and renamed the restored server to the original name.
The best way to deal with that is to shutdown the vm and change the UUID of the VM to the old/original ID.
If you use IFINCR backup, then a new IFFULL is needed.