Hi, I had a VM backup strangely last night. It seemed to snap and backup the VM correctly. But then it cleaned up the snapshots. Waited until another slot was open and then snapped the machine again and completed the indexing portion. It was successful the 2nd time. And it was a quick process. 

This is my log of when it failed doing the indexing copy. I wonder if anyone has seen this or knows what might cause this? On the 2nd attempt the 2nd "IDXCONNECT" line says that it succeeded. 

5944  23c0  05/15 19:58:54 1012569 Copying file[VolHdr(50053ef9-4f23-a764-15a9-72282966693f).imf]...

5944  23c0  05/15 19:58:54 1012569 All files transfered successfully.

5944  23c0  05/15 19:58:54 1012569 Disconnecting before returning.

5944  23c0  05/15 19:58:54 1012569 IDXCONNECT Files transferred to indexCache directory

5944  23c0  05/15 19:59:27 1012569 IDXCONNECT IndexingService failed to process dirchange/optional files with errorCode 87270544

5944  23c0  05/15 19:59:27 1012569 VSBkpController::SendVMMetadataFiles() - Failed to send index files for VM [50053ef9-4f23-a764-15a9-72282966693f]

5944  23c0  05/15 19:59:27 1012569 CVMWareInfo::_UnmountVM_VCB4() - Unmounting Disks for VM [cs99ma58]

Hi jhottes,

It looks like there are two issues here:

1. The job was interrupted or the backup failed for that specific VM.

A snapshot will only exist on the VM while we are backing it up. If the snapshot was taken, cleaned up and then performed on the next attempt, then there must of been an issue at the time we were backing up that VM. Otherwise we wouldn't have seen the second snapshot appear if the VM backup completed on the first attempt.

2. There was an issue processing the index files on the MediaAgent.

At the end of the backup job, the VSA will collect all the index files generated for the job and then push them to the MediaAgent Index Cache. For some reason there was an issue with the Indexing Service on the MediaAgent when handling these files.

Had this operation completed successfully, we should see the line "IDXCONNECT [uploadFilesToIndexCacheDir] succeeded" outputted to the logs.

Unfortunately there isn't too much more we can tell from the log snippets provided. We would need the vsbkp logs from the VSA proxy, and the indexing logs from the MA to get a better idea as to what went wrong.

If you could open up a case with us and upload the logs for Job ID 1012569, we'd be happy to review and provide an RCA on the issue for you.

Any questions, let me know.

Cheers,

Peter Kalinic - CommVault/Sydney
Customer Support Engineer - Client Group

Thanks Peter. I did get the IDXCONNECT [uploadFilesToIndexCacheDir] succeeded the 2nd time. I was just curious of what else to check as to why the media agent could complete the backup yet not properly transfer the indexes. I looked through some other logs and didn't see anything noteable.

I am not sure I need a ticket yet as this didnt impact our backups just a hiccup. I might if I ever get some free time (haha) to mess with this.

Thanks Peter. I did get the IDXCONNECT [uploadFilesToIndexCacheDir] succeeded the 2nd time. I was just curious of what else to check as to why the media agent could complete the backup yet not properly transfer the indexes. I looked through some other logs and didn't see anything noteable.

I am not sure I need a ticket yet as this didnt impact our backups just a hiccup. I might if I ever get some free time (haha) to mess with this.