Description of problem:
Small 2node cluster only for GFS use. Upon restarting cman on one node
other one oopsed.
Version-Release number of selected component (if applicable):
happened once so far
Steps to Reproduce:
- one node active
- other node cman restart
- first node oopsed
kernel oops, file attached
Created attachment 125938 [details]
Kernel oops output
There have been a number of bug fixes since June 2005
that may address this. I'd suggest you hold out until
FC5 and use the latest cman/dlm/gfs rpm's for that.