redhat cluster

[V S Chakravarthi]

hi,

how to find cause for a cluster service to go to failed state.
i checked messages and cluster i do not get clue.

service failed with return 11 and then went to failed state.

i couldnt start it manually and then i had to restart the cluster service then it started.

-chakri

[Vitaly Karasik_1]

>i checked messages and cluster i do not get clue.

- what do you see in the cluster log?
- you can increase log level for cluster

[Steven E. Protter]

Shalom,

all cluster messages go into /var/log/messages

in one session:
tail -f /var/log/messages

in another window:
recreate the failure condition. You should see more.

if not here is a basic checklist:
1) Check the startup script. If anything, even something seemingly ubiquitous returns an return code other than zero, the service will go into a failed state.
2) Networking must be perfect. No inconsistencies in /etc/hosts or DNS recored that may be used.
3) shut down all daemons, fenced, rgmanager, cman ccsd on all nodes. Make sure cluster.conf is the same on all nodes.
4) Restart all nodes after step 3.

SEP

[V S Chakravarthi]

hi,

Problem is i cannot reproduce.Also, logs doesnt give much information.

thanks
chakri