Hi,
We have a deployment of 3 x Server 2012 Physical Hosts, each Physical host is running 2 VM's (variety of guest OS 2003, 2008, 2012) and hyper-v replica is enabled on all of the guests.
One of the guests is a 2012 server standard which is being used as a file server and is hosting around 2.5TB of data. With both a 3rd party backup software and WSB backing up to an external USB drive (4TB) the backup gets to just over 1.3TB of data backed
up then users complain they are unable to access files on the file server.
We had been running the 3rd party software on all guests with no issues, but the backups are scheduled to run out of hours. Obviously with the amount of data on the file server we expected it to over-run but also hoped that while the backup was running the
server should also have been operational! We decided last night to use WSB to see if it was the 3rd party software that was the cause and have found out it isn't!
Just to provide a bit more info, the Servers have plenty of CPU and RAM available for the host and guest. The Host OS runs off 2 x SATA drives in RAID-1 and this particular guest runs off 4 x SATA drives in RAID 10.
While WSB was in progress (at the time of checking when users were complaining) the state of the VM was shown as 'running-critical'. When we stopped the backup it did so quite quickly but the hyper-v manager stopped showing any data on the guests, we were
then forced to reboot the host, which also hung on shutdown causing us to have to hard power-cycle the unit.
All the other hosts are fine, and operation of this host is fine, also backup of any of the other guests on this host cause no problems, other than the fact they are small enough to complete their backups out of hours. But surely with VSS backing up during
the day shouldnt be a problem (we use WSB to backup physical servers during the day at other sites with no issue)
It now means we only have a full backup of their main file server from the weekend, and its now looking like we will only be able to run another backup this weekend. Obviously there is the replica to go to in the event of a failover, but we need this problem
sorting urgently, any help here would be greatly appreciated!