We have tried making sure its been run from theĬoordinator node with no different, we have tried disabling ODX and TRIM, we have tried with VMQ disabled and also with Recieve Side Scaling disabled but the behaviour is still present. If the job is cancelled the node comes back to life and if the job is started again, exactly the same behaviour is displayed. The copy begins with a transfer speed of around 200 MB/s and gives a estimated time of approximately 11 minutes. With about 5min 30sec remaining the transfer speed drops to about 25MB/s and stays there. We created a 127GB VHDX in one of the CSVs within the cluster, then from one of the Hyper-V nodes we copy the VHDX to the local system drive.
We have a 3 node Hyper-V 2012 R2 cluster with 2 x 1GB iSCSI NICs with MPIO enabled.
