Description of problem: If a clustered service is running with "apache" as a resource and then "pkill httpd" is ran the service will got into a failed state. The reason that service goes into the failed state is because the pid file was already cleaned up and no longer exist. Thus checking for existence of the pid file will always fail. The solution is to only call the "stop_generic $PIDFILE" if the pid file actual exist. The service will then go into failed state. Version-Release number of selected component (if applicable): rgmanager-2.0.46-1.el5_3.3 How reproducible: Every time. Steps to Reproduce: 1. create apache resource in a service 2. start service 3. pkill httpd 4. clustat Actual results: The service will be in failed state and does not recover correctly. Expected results: The service should be restarted correctly. Additional info: Patch to resolve the issue is attached to the bz.
Created attachment 355302 [details] Patch that checks for existence of .pid file
This really looks like a duplicate of https://bugzilla.redhat.com/show_bug.cgi?id=449394
*** This bug has been marked as a duplicate of bug 449394 ***