I daily copy the Groupwise databases to the restore area on the same volume.
Quite often this results into the node receiving a poison pill from the other node because it claims the first node is dead.
My guess is that dbcopy for some reason uses so many resources that the cluster keepalive signals are suppressed.
Anyone seen this before?
I'm using SLES10SP2 and OES2SP1. DBCOPY version 3.2.
Would it help to run DBCOPY with less threads?