So you have decided to run Windows Server 2008 R2, using the new Clustered Shared Volumes features and you are going to have a SAN based on iSCSI. Good for you, then you will properly also want to set it up according to best practices, hmm lets see…
Two physical servers, with 4 NIC’s that is perfect. I use one NIC for management, one NIC as the Hyper-V network switch and the other two NIC’s as iSCSI NIC’s and as a best practices you bind off everything except for IP on the iSCSI NIC’s
THAT WILL MAKE Cluster Shared Volumes to FAIL !!!
Microsoft require you to enable the “Microsoft Network Client” and “File and Printer Sharing for Microsoft Networks” on all network that takes part of the cluster, in other words you have two choices here.
1. Do not use the iSCSI NIC’s as “cluster NIC’s”, in other word make sure that they are listed as disabled for cluster use (You cannot use them for Internal or Client use). Bad thing about this is that in that case they cannot be used as heartbeat either.
2. Enable “Microsoft Network Client” and “File and Printer Sharing for Microsoft Networks” on the iSCSI NIC’s, (All NIC’s) bad thing about that is that you will have some traffic on the iSCSI network that is “unwanted”.
For some reason there is no KB article that I have found on this little issue, the problem is that it works perfect as long as you do not use the clustered shared volumes and Live Migration, if you just use Quick Migration it works. It is easy to see if there is anything wrong with the configuration, just open c:\clusterstorage\volume1 on one of the nodes, then try to open the same path on the other node. If it freezes up, you know what the problem is. Option one or option two. I usually take option number two
There are two other articles about this issue, sorry to say, not in English.
If you open the event log, you will find this:
Log Name: System
Event ID: 5120
Task Category: Cluster Shared Volume
Description: Cluster Shared Volume ‘Volume1′ (’Cluster Disk 1′) is no longer available on this node because of ‘STATUS_BAD_NETWORK_PATH(c00000be)’. All I/O will temporarily be queued until a path to the volume is reestablished.