ginolard (5/7/2013)
Sorry to hijack but we are experiencing a similar issue on our 3-node cluster. Two of the disks attached to an HP MSA SAN reguarly "lose ownership".The iSCSI initiator properties for ALL targets are set to
Local Adapater: Default
Initiator IP: Default
Target Portal IP: Default
Are you saying these values should be changed to this?
Local Adapater: Microsoft iSCSI Initiator
Initiator IP: x.x.x.x
Target Portal IP: x.x.x.x
I also notice that the binding order is DIFFERENT on each node, God knows what idiot set that up but I guess they should be exactly the same?
Two things ...
Each case is different, so I suggest open a new thread and post your issue, so any other user or Perry can step in and help. Not that I mind you for hijack my thread though 🙂 ... but you may get better attention.
2nd thing is ... the MS engineer advice was to change the binding order and put the fixed IP on iSCSI initiatior, yes, that is correct. But I do differ on the 1st advice though. This is how I usually configure my Clusters (binding order(
-Public
-Heartbeat
-SAN
This is why MS is recommending
-Public
-SAN
-Heartbeat
And binding order should BE THE SAME ON ALL NODES!
I do agree that the iSCSI initiator should use actual IPs though, and that's what I will change today with the help of my IT resource.