I am stressing a hyper-v host at the moment and have noticed a new issue using DPM ti back up the VMs from via host backups. All the VMs are on one host volume, at is a a stand-alone server with just a few internal hard disks. DPM is DPM 2012 SP1 and Hyper-V is the role in Server 2012.
now that the memory usage on the host has become high I am seeing about once every two weeks a random VM end up in "Paused" some time soon after a DPM backup as been initiated. (At least, I say the host ram is high, which it is, maybe this is a ram in the VM issue). Anyway, I will find that the backups run successfully but with just a note in the event logs, one of the VMs will end up as "Paused" with the following note in Hyper-V Manager: Disk(s) running out of space
neither the host or the vhds comprising the VM are actually anywhere near being full. If I unpause the VM it resumes running normally. From my experience with DPM I would assume that this is a condition from VSS having some issue with the load on the host server and some timeout of the VSS snapshot or something.
Since it only occurs randomly I was going to look into destressing the host server but was wondering if there was anything I could log or report to help make the issue either go away with some dpm or windows patches or to have DPM note the problem and report it proactively. I see no event log entries of any consequence in the DPM server (note, I have no SCOM integration turned on as I have no SCPM to integrate tom so that might be why)
Is there any useful technet article noting this issue to explain in detail what could bring about this erroneous disk space / paused condition.