I have two mirrored iSCSI based SANs which are based on NWSB6.5 sp1. One of the SANs is connected over a fiber cable from a remote building.


I have a 2-node cluster which is mirroring all partitions including the SBD partition.

The server that's running Groupwise had some corruptions in the NSS Pool which NSS Rebuild could not fix. I REMed out iscsinit connect statement out of autoexec.ncf and rebooted the server so that I could wipe out the Pool on one partition and restore from backup. (I wanted to keep one partition in case we had to resort to data recovery)

When the server cameup, I started to get the following error in Logger:

CLUSTER-<FATAL.-<6059>: Failed to Read all SLOTs data from SBD partition(202).

Now, when the server was coming up normal (even after the corruption in the Pool) both SANs were being read and all partitions (including SBD) showed to be mirrored and in-sync.

Any ideas on why the server would not come up reading only one of the SANs?

Thank you,