Using DPM 2012 SP1 in Hyper-V Failover Cluster Environment RRS feed

  • General discussion

  • We have a 3-node Server 2012 Hyper-V Failover cluster using 4 CSV drives attached to an EqualLogic iSCSI SAN.  We have DPM 2012 SP1 configured and backing up each VM through the Cluster.  Everything works ok outside of the occasional Replica Inconsistent errors.  When a Recovery Point is initiated we can watch the HIT Kit take a snapshot at the Equallogic level, attach the snapshot to the Hyper-V host and backup over the LAN to the DPM server.   My question is when it comes to restore we have to recover the VHDX file and copy it back to another Hyper-V host.  Hyper-V is installed on our DPM server, is there a way to mount the VHDX directly from the DPM server?  If not, would it be better to just create Hyper-V Replicas from each VM back to my backup host?  Am I missing some configuration on how to setup DPM and integrating Hyper-V Replicas?

    • Changed type BScholl Monday, October 21, 2013 4:42 PM
    Tuesday, October 15, 2013 6:57 PM

All replies

  • I am not sure if I understand your question correctly. When you go to "Recovery" tap on DPM and select the VM you want to restore, in the navigation menu if you just select the VM you are able to restore the "whole" VM but if you go on level down you are able to select the VHDX in the right hand window. This VHDX can be recovered to any Hyper-V Hosts or Hyper-V Cluster as long as it has at least the same version and has a DPM Agent installed. So what requirements do you have that you want to use another way? If your target Hyper-V Host or cluster doesn't have a DPM Agent installed you could use the "network share" option to restore the VHDX to your host/cluster.

    "Mounting" the VM within DPM lets you move the file level within this VHDX.

    I am not aware that there is an integration between Hyper-V Replica and DPM. There is a short blog here:

    Wednesday, October 16, 2013 9:18 AM
  • Everything you’ve mentioned works correct and as expected.  In the event of a site failure or let’s say the entire Cluster Failed, we would need another Hyper-V host to startup all failed VMs.  Instead of using DPM why not just use Hyper-V replicas from each VM to the Stand-by host.  Then in case of a failure we could have a “hot-site” and start the VMs instantly without having to perform recovery.   Our machine purchased for DPM has enough memory and local storage for all production Servers to startup. 

    Wednesday, October 16, 2013 12:02 PM