Yesterday I found some complicate case httpd service on heartbeat build with Redhat Enterprise (RHEL5.8 64 bit) OS.
I have 4 resources in group
Vip
Drbd_fs
mysqld
httpd
On starting resources in heartbeat everything look fine until "httpd" resource ,
It stuck on this point(httpd) for long time and return some error about "unknown error"
#On Redhat box show in crm
r_httpd_start_0 (node=node00, call=20, rc=1, status=complete): unknown error
##This error log from ha-debug
apache(r_httpd)[7708]: 2013/03/27_22:20:20 INFO: apache not running
apache(r_httpd)[7708]: 2013/03/27_22:20:20 INFO: waiting for apache /etc/conf/httpd.conf to come up Mar 27 22:20:22 node00 lrmd: [9431]: info: RA output: (r_httpd:start:stderr) /usr/lib/ocf/resource.d//heartbeat/apache: line 212: kill: (7788) - No such process
apache(r_httpd)[7708]: 2013/03/27_22:20:22 INFO: Killing apache PID 7788
apache(r_httpd)[7708]: 2013/03/27_22:20:22 INFO: apache stopped.
Mar 27 22:20:22 node00 lrmd: [9431]: info: operation start[70] on r_httpd for client 9434: pid 7708 exited with return code 1 ###
##Point seem heartbeat try to kill httpd pid on the start phase , Because it look duplicate pid so became the effect to could not running httpd via heartbeat properly.
So after I open configure on "server-status" phase in "/etc/httpd/conf/httpd.conf"
And try to restart heartbeat again this problem could solved in RHEL box.
And I also test in OS=Centos , it not happened for this kind of case even though close configure on "server-status" .
FreeOS seem better PurchaseOS in this kind of case.
Just for your information if used RHEL.
##Thank you comment from David Vossel Also
https://github.com/davidvossel/resource-agents/commit/2ce0b97d3009b4f0c94d0db83fd68b311de7bc85#commitcomment-2887917
