We have a Server 2012 R2 cluster that connects to an Equallogic SAN and backs up to a DPM server with a dedicated JBOD array
When we backup a VM using DPM 2012 R2, there is a large amount of network traffic over both the Cluster Heartbeat and Live Migration networks. Both networks on the host running the VM show usage of aprox 500Mbps inbound and both networks on the Host that owns the CSV volume show usage of 500Mbps outbound. That almost seems to me to be CSV Redirected IO but when I run the Powershell command Get-ClusterSharedVolumeState it shows all CSV's in Direct mode.
My 2 question is, why is it using both networks (Cluster and VM) on each host, with almost an identical throughput?
The data transfer to the Management network on the DPM server is about 1/4 of the throughput between the hosts.
Simon Holman
Expeed Technology
Australian Web Hosting