VMware Solutions Discussions

VSC 4.2 Backup/Recovery - Last Run Status Failed

justin_smith
7,343 Views

Is there anyway to troubleshoot why some backup jobs fail?

Im running VSC 4.2 and have about 80% of the backup jobs fail now... need to find out why.

7 REPLIES 7

ydk
NetApp Alumni
7,214 Views

Assuming you have configured the backup jobs to send status e-mails, I would start by checking backup status e-mail and see if that provides any clue. If not, here are other options:

1. Go to Home -> Solutions and Applications -> NetApp and then click on backup to see list of jobs. Right click on backup job and select "Run job now" option. You can then check the vSphere tasks panel and see if you see anything obvious there.

2. Try "Backup now" option for a VM (or a datastore) in the backup job. If your backup job contains many VMs (or datastores), you may want to try #1 above first.

3. Collect the VSC log bundle, extract the generated zip file and look smvi.log file. Open the file in text editor and search for "error".

If none of that works, I would recommend opening a case.

justin_smith
7,214 Views

I think it has to do with the wrong UUID. The VM guys formatted some of the DS's with VMFS5 and it changed the UUID. How can I "reset" it so to speak?

  2013-08-06 11:47:31,402 WARN - Could not find the following backup ids as either a datastore or virtual machine: [4e933213-312a0da2-e2b3-002219c7f982]

   2013-08-06 11:47:31,402 ERROR - Did not find any of the following datastores or virtual machines to backup: [4e933213-312a0da2-e2b3-002219c7f982]

   2013-08-06 11:47:31,402 ERROR - FLOW-11019: Failure in VmGetVirtualMachinesToBackup: Did not find any of the following datastores or virtual machines to backup: [4e933213-312a0da2-e2b3-002219c7f982]

   2013-08-06 11:47:31,402 ERROR - FLOW-11010: Operation transitioning to abort due to prior failure.

ydk
NetApp Alumni
7,214 Views

You can create a new backup job and select the datastores and/or VMs from the updated datastores.     

justin_smith
7,214 Views

I have roughly 50 backup jobs... seems like a lot of work.

schrie
7,214 Views

Unfortunately there isn't a way to "automate" this process if the UUID's of the VMFS volumes you've selected (or VM's if you've chosen them) have been changed/modified in the vCenter DB.  VSC Backup and Recovery uses XML files to store your original choices, and enumerates them based on the UUID retrieved from vCenters' SDK at the time it was selected.

To avoid having to entirely recreate each backup job, you can edit each of the backup jobs, and go to the "virtual entities" tab - move any entities from the right hand column, and then readd the "same" one from the new list appearing on the left side.  This will take awhile, but will allow you to keep the rest of the properties of the job and save a significant amount of time compared to completely recreating the jobs.  This also isn't something that will need to be repeated unless/until the UUID changes again, which is not likely to occur on a regular basis.

justin_smith
7,214 Views

That was the next thing I was going to try.... thanks for the info.

Last question though, would I have to remove it from the selected entities, then click OK so SVC knows, then open up the same backup job and add the same DS? Or would a simple remove and re-add without clicking OK do the trick?

Only asking because I want to say I tried doing it in one swift motion and it didnt work. Ill try removing it, clicking OK and re-adding.

Update:

It doesnt work when simply removing it and re-adding it to the selected entities... also, it wont allow me to remove the entity and click OK.

I might have to add a different DS to the job, go back, edit it, and add the right DS..... unless theres a way to script the removal of the entities? I can then just go in and add the right one....

Message was edited by: Justin Smith

schrie
7,214 Views

Ok, lets go ahead and get you a support case opened and we'll help modify them manually for the time being (possibly with powercli to help), and then for a longer-term solution we can bring this up with the product managers/developers and come up with a solution which will re-enumerate each of them in some way or have a sort of "refresh" button.

Public