I'm working on deploying a four node file cluster on OES2 SP1. The cluster has twelve resources which are SAN attached 1TB volumes. Each node has two QLogic HBAs.

Everything seems to work well, EXCEPT nss /PoolActivate and /PoolDeactivate takes 25-30 seconds each with no data on the volumes! This makes cluster resources moves take way too long.

In my experience, pool activations and deactivations take a couple seconds at most even with lots of data on a large volume. I've been trying to pin the problem on some part of the stack, but it seems like the slow-up is happening somewhere in the nss kernel modules.

Can anybody point me in the right direction on this? Should I suspect multipathing? Anybody running a similar config?