If from the /var/log/messages you are receiving network / o2net related errors and some OCFS2 / system functionality is not available or not working properly, generating a ‘tcpdump’ over the network interconnect may help to diagnose the problem for root cause analysis (RCA).
Step 1: Start tcpdump on all nodes:
# tcpdump -i -C 20 -s 10000 -S \
> -w /tmp/`hostname -s`_tcpdump.log -ttt 'port 7777' &
note that ethX stands for the network interface that carries the OCFS2 net (o2net) traffic.
Step 2: Reproduce the problem or wait until the problem reproduces. Note that if the problem is happening when the system boots up, it may be adequate to restart the o2cb and ocfs2 services. If the problem does not reproduce, then the above tcpdump needs to be put into the service scripts.
