This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

transfer failed problem with vss subsystem.

I know this has been addressed on this forum before so i apologize for re-posting but I've been down the rabbit hole for some time now.  I have RR 6.1.3.100 running and the agent is on about a dozen VM's in vmware.  Yesterday my hourly backups began failing with.

 

DosnapshotSet failed with error "they system was unable to flush I/O writes."   "problem with vss subsystem"    We've seen this before in other posts.

 

If i reboot, the first transfer is successful but not after that.  The application log has a vss error id 12297, 12289.  12297 states the volume shadow copy service error:  I/O writes cannot be flushed during shadow copy creation on volume (its the system reserve partition).

 

I've run chkdsk on both C:\ and system restore.

defrag on both.

gone as far as to reset the system restore so size isn't an issue, 90% free.    Same happens first one after reboot is good then fail.

I/O can't be a problem since the one i'm testing all changes on isn't in use period.

no other backup software in use, and shadow copies within windows is off. 

 

Thoughts or idea's?

 

EDIT: additionally if someone can help me understand how windows 'shadow copies' plays into how RR is working.

  • I have seen this several times and the issue has always been with the ESX server or config. I am not able to look up my notes currently but check vmware knowle base.

    I think the issues I have seen are:

    - vmware tools not installed or not updated

    - The disk IO on the esx disks being to high. That causes VSS to not be able to the disk during the client snapshot. The fact the client itself is not disk IO intensive does not matter in this case.

    - Check vmware jobs. I will often see vmware clone running and this creates a ton of IO

    On a related note, check to see if VSS is being called by any other process. I have seen vmware use vss backups and the job will fail and leave vss in a failed state. So then our backups fail (since vss is still in a failed state)
  • Shadow Copies. This could be a bit dicey as the term maybe a bit flexible. I typically think of Shadow Copies as OS backups and they have nothing to do with RR.

    RR uses MS VSS to create snapshots (just like shadow copies) but the shadow copy is deleted after it is done. So RR should not leave any shadow copies around

    But when RR calls VSS to take a snapshot, many of the shadow copy settings are used. So you may have to change shadow copy settings for certain RR / VSS backups to run. For example:

    support.quest.com/.../137398

    If anyone has better info, feel free to add or correct me