Comment 5 for bug 1423116

To test this patch:
* reproduce the isse (Note, the reproducing steps are unknown, please update the bug description then discovered),
* for affected node, manually update the OCF file for Rabbitmq RA
* for affected node, kill -9 the hanged monitor action and its forks , example http://paste.openstack.org/show/176859/
* it is expected that rabbit node will rejoin the cluster and the issue couldn't be reproduced anymore.