hi there everybody !

we have a nw6sp5 stretched cluster (2 nodes, one cluster volume on a
msa1000, both servers see the volume via SSP, multipathing off).
hba: qlogic, ql2300.ham version 6.50.26 (the only version which can see
the sbd partition !!; i already tried all the versions).
cpus: intel xeon 2,4 GHz
ram: 2048 MB
ProLiant Support Pack 7.20
both nodes are in a server lan.
parameters (like min + max connections etc...) have been tuned for cluster

the problem:
we had an abend on a cluster node because one compaq module was causing
trouble (cpqasm.nlm).
well the node went up again and forced a remirror process (at this time
full working condition, about 400 connections on the active cluster node).

strange things happened then:
the remirror process was at about 50-60 % when suddenly one cluster node
was forced in an abend (it was the passive one i think).
what we could see when executing "cluster stats display" was that the
ticks were going up (our heartbeat is 8 seconds);
well for me it looked like the remirror process was killing the heartbeat
process......but....is that possible, since i think the heartbeat process
should be of highest priority.

well....we never experienced such problems since we went online with our
cluster. we' ve been testing cluster before going live, but i don't know
what the problem is.
nothing was patched since we went live, except nw6sp5.

any help is appreciated.....