Why does my OES2 (Linux) cluster node hang when stopping NCS?

Every once in a while I read something out on the ‘Net that that is so important, and not well-known, that I think I really need to pass it along to anyone who might be reading my blog as well… This is one of those things.

The other day, I was troubleshooting my own Linux-based OES2 cluster services, since it appeared the servers would hang while shutting down or stopping the NCS services.

In my research, I came across the answer in — not surprisingly — the Novell Support Forums. It is a valuable resource! Here is a link to that thread and answer, but I’ll post it below for your benefit.

From forum user Gerrit Beul, A.K.A. “gbeul” (and thanks to others):

“go to /etc/sysconfig/boot
open it in vi
search for variable RUN_PARALLEL
this is set by default to ‘yes’
change it to ‘no’
now init 6 and shutdown works fine”

This simple fix makes a huge difference in the timing and performance of cluster service stop and start. Without the setting change, it took my servers a very, very long time to stop NCS and exit the cluster. Another work-around is to never stop or start the NCS services, and only ever “join” or “leave” the cluster…. but how realistic is that? You have to start or stop it eventually…. don’t you?


Leave a Comment

Your email address will not be published. Required fields are marked *