Hi Greg,

Nothing has changed. It is always the same - perhaps something else is wrong. I would suggest following:
- delete physical clients and virtual ones
- restart NW (perhaps something is in cache)
- create physical ones first (use same clientid as before)
- create virtual ones then (use same clientid as before)
- make sure that administrator or root from physical hosts is listed in remote access of the virtual client
- if on UNIX, you may wish to rerun networker.cluster script

I will look into this, I am pretty sure it is a bug though. I have 4 clusters and they have not failed over recently. All four clusters are failing with the same errors. Cheers.

Sounds like an issue with savefs then - meaning savefs is not able for given client (virtual one I assume) to list file systems owned by cluster. You should be able to see that from the cmd line by running savefs on client system:
- savefs -s server_name -vp -c physical1_node
- savefs -s server_name -vp -c physical2_node
- savefs -s server_name -vp -c virtual1_node
- savefs -s server_name -vp -c virtual2_node
- etc

I believe long time ago I did see this once and the problem was that after clustered system failed over there was a corruption at arp level left which caused confusion at OS level where is what and to what does it belong. Not sure if this is the same case or related at all - could be a bug as well - support should know it. Is this Windows 2003 perhaps (I believe I did see someone reporting sort of a issue under Windows 2003 only).

I found another user with the same issue, it seems that the "ALL" Save set specification no longer works for a virtual node on a cluster. I was able to successfully back up the virtual node by individually specifying the items to back up. I have opened a case with support. I will post the results when I have them.

Also I can confirm that nothing has changed in NetWorker behavior of resolving Cluster resources in Microsoft Cluster Server. Save set 'All' still works as before.
The root cause here may be something weird in network naming or name resolution or the binding order of NICs.
NetWorker behavior to resolve MSCS resources is also not depending on failover.