Multipathing status changed

[Peter J West]

We're been running an 8 node multi-site Lefthand solution for some time now and when it was originally set up a lot of time was taken configuring the multi-pathing to ensure we had full redundancy.

I don't know where but somewhere along the line this has changed and when I check the datastore status now it's showing as 'Partial/No Redundancy' for all volumes.

Is this expected?  Or more imporantly is there a way to get back to Full Redundancy?

I suspect the change occurred when we upgraded vSphere from v4.1 to v5.0.  We're now running v5.1 and have 3 VMKernel adapters bound to the iSCSI Software Adapter.

I'm guessing (but would appreciate confirmation) that the system considers us to only have partial redundancy because we're still ultimately dependent on a SPOF which would be the iSCSI Software Adapter.

Grateful for any comments on this one.

Pete

[oikjn]

are you targeting the VIP of the cluster for your LUNs?  If so, you should have all the redundancy required since any node that fails while it is acting as an individual LUN target will automatically failover to the next available node.  It really shouldn't take a lot of time to setup the SAN specifically "for redundancy" unless you are talking about many LUNs on many servers which is time consuming just because of the quantity.  I"m a hyper-V shop so I won't speak about the specifics of how to setup the ESX hosts, but if you read the setup guide for ESX+StoreVirtual it really doesn't take that much to get it correct/redundant.  Now, it might be that ESX is just reporting non-redundant because it only things you are connecte to a single IP and it assumes a single IP is not a clustered SAN and that is why its giving you a warning and then I bet it probably can be ignored.