[Ocfs2-users] dlm_get_lock_resource/dlm_query_join_handler errors

[Mark Rife]

Hello,

 

I am running a 4-node ocfs2 cluster.

 

Our servers are running Redhat AS4, kernel 2.6.9-34.0.1.ELhugemem.

 

Our ocfs2 package versions are:

 

# rpm -qa | grep ocfs2

 

ocfs2-tools-debuginfo-1.2.1-1

ocfs2-tools-1.2.1-1

ocfs2-2.6.9-34.0.1.ELhugemem-1.2.2-1

ocfs2console-1.2.1-1

 

One of the nodes (#3) crashed. We’re rebooted node 3, but now it hangs as it
tries to rejoin the cluster.

 

On two of the nodes that are up (0 and 1), I am getting messages repeated
/var/log/messages that look like this:

 

Jul 19 09:39:40 radon6 kernel: (3994,2):dlm_query_join_handler:614 node 3
trying to join, but recovery is ongoing.

Jul 19 09:39:50 radon6 last message repeated 25 times

Jul 19 09:39:51 radon6 kernel: (27704,1):dlm_get_lock_resource:895
46A341FD43114DE4A10E7D63C5099461:M0000000000000000667f6c991b8fc9: at least
one node (3) torecover before lock mastery can begin

Jul 19 09:39:51 radon6 kernel: (3994,2):dlm_query_join_handler:614 node 3
trying to join, but recovery is ongoing.

Jul 19 09:39:51 radon6 kernel: (10183,1):dlm_get_lock_resource:895
46A341FD43114DE4A10E7D63C5099461:M00000000000000000081e17e89ae74: at least
one node (3) torecover before lock mastery can begin

Jul 19 09:39:51 radon6 kernel: (3994,2):dlm_query_join_handler:614 node 3
trying to join, but recovery is ongoing.

 

This appears to be in an infinite loop and node 3 never starts.

 

I’m not seeing the messages on node 2.

 

The cluster is up and running on 3 of the 4 servers, but I need to get all 4
nodes running again.

 

Can anyone provide any insight on what is going on or how this should be
handled?

 

Thanks!

 

Mark Rife

Oracle Applications DBA

markrife at hotmail.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://oss.oracle.com/pipermail/ocfs2-users/attachments/20060719/e96ce6ae/attachment.html