Recently I was working with a customer where we had very sporadic remote management related problems to random Hyper-V hosts that were deployed on Windows Server 2012.
The affected servers would occasionally drop off from Server Manager (host not contactable, or data retrieval errors) and additional when trying to add the nodes to a failover cluster we were seeing RPC errors. We could also not remote Powershell to these nodes (similar WinRM related error messages ensued)
The most confusing part of this problem was that it was not consistent. Ordinarily I would blame the firewall, or corruption in the WinRM services… however this problem occurred on different nodes at different times and connectivity was fine from certain hosts but blocked from others.
I happened to chance on a very similar problem whilst researching the issue which related to large send offload (LRO) being enabled on the NIC used to contact the hosts. The customer mentioned that they had experienced problems with this in the past so we disabled this and rebooted one of the affected nodes. As if by magic the node was once again contactable and we could add it in to a failover clustering. Hope this helps. ThanksSteveH