[Pacemaker] ClusterMon resource starting multiple instances of crm_mon

[Alexander Rubtsov]

The issue was observed under condition where a non-existing external agent is attempted to be executed.

Steps to reproduce:

1. Define a primitive that uses ClusterMon resource and includes the "-E" option. For example:
primitive p_ClusterMon ocf:pacemaker:ClusterMon \
    params extra_options="-E /etc/init.d/rsc-monitor" update=30 user=root \
    meta failure-timeout=120 migration-threshold=10 \
    op monitor interval=60 on-fail=restart

2. Ensure that the script, specified as the value of "-E" option, does not exist.

3. Enable this primitive

Actual result:
Each attempt of running, new crm_mon processes are generated

Expected result:
Only 1 crm_mon process is running

[Denis Meltsaykin (dmeltsaykin)]

We don't fix LOW importance bugs in 9.x.