We are setting up a three node cluster running NW6.5sp4a and Groupwise
6.5.5. The nodes are HPDL380 server with 2 Qlogic 2300 HBAs. We have been
testing the multipath which seems to work ok using the following commands
in the Startup.NCF:

If I pull the fiber cable from one of the HBAs, that HBA is deactivated
and the storage is then seen through the other HBA. Then we tried loading
the cluster resource that has the GW Domain and pulling one of the cables
from the HBA, the server then sees the storage using the other HBA but
deactivates the pool and causes the resource to go Comatose. Should I also
be using the LSIMPE.CDM? Any ideas on what we are missing so that the
resource continues to run using the other HBA instead of putting it
Comatose. The errors we see on the logger screen are listed below:

May 17, 2006 11:15:31 am NSS-3.23d-xxxx: comnPool.c[2672]
Pool GWDOM2VS: System data I/O error 22(zlssMSAP.[1796]).
Block 0(file block 0)(ZID 0)
************************************************** *************************
************************************************** ******
************************************************** *************************
************************************************** ******
5-17-2006 11:15:31 am: COMN-3.23-1092 [nmID=A0025]
NSS-3.00-5001: Pool BAZ-NC12/GWDOM2VS is being deactivated.
An I/O error (22(zlssMSAP.c[1796])) at block 0(file block 0)(ZID 0)
has compromised pool integrity.