MSA Storage

Adding 2nd FC HBA connection to production MSA1000 Cluster

New Member

Adding 2nd FC HBA connection to production MSA1000 Cluster

I mistakenly asked this at the end of another similar thread but thought I should post it separately to insure I get additional info if it is available:

I have a Microsoft cluster consisting of 2 DL380s with Windows 2003 Enterprise Edition.

Right now each DL380 has 2 fca2214 adapters installed in them but only 1 of each of the adapters is connected via fiber cable to my MSA1000. My MSA1000 also has dual switches.

I need to install Securepath for Windows and connect up the 2nd FC HBA on each server with a little impact to this production environment as possible. My original questions was: Can I safely install Securepath and hook up the 2nd FC HBA connections on each server? Or do I need to back out the cluster services first before installing Securepath and hooking up the 2nd connection on each server?

Steven Clementi recommended against it but Dilbert responded:

No, I have successfully moved from a non-redundant to redundant config. on an existing cluster.

Whilst I agree that this is not the "normal" approach I can't see why it isn't both acceptable and supportable.

Dilbert, any tips on doing this safely? Do you remember the sequence of what you actually did? Do I need to stop the cluster services on the inactive server while I install Securepath and hook up the FC HBA?

Any additional detail anyone can provide would be greatly appreciated.

Thanks, Rochelle
New Member

Re: Adding 2nd FC HBA connection to production MSA1000 Cluster

Ciao Rochelle,
I followed the following steps serveral times with no problems:

First of all let's configure Cluster resources inside Cluster Administrator to not failover and move all cluster resources on one node.

Than write down all the disk signtures (with Microsoft dumpcfg.exe)
Install secure path on the node without resources
reboot the server, plug the second fc cable and check again the signatures.
try to move one resource at time.
if all goes ok you can proceed to the next sever.

Only once I had to evict one node and rejoin the cluster...