What were you trying to do that didn't work?

DC election of Pacemaker was stuck because actions in the current batch were stuck due to a hang

What is the impact of this issue to you?

Pacemaker did not run any action due to the current batch being stuck

Please provide the package NVR for which the bug is seen:

2.1.9-1

How reproducible is this bug?:

Reproducible when actions in the current batch hang

Steps to reproduce

1. Force resource agent to hang
2. Make a pacemaker to execute a resource agent
3. Kill a DC node, then Pacemaker will execute the resource agent and get stuck, because actions in the current batch is not finishing. Pacemaker does not elect DC node until resource agent is finished executing

Expected results

Pacemaker should not be stuck because the current batch isn't finished in time or hung. There should be a configuration parameter that can be tuned to force parallel batch execution

Actual results

Pacemaker gets stuck if current batch action hangs

Is there a way to mitigate this kind of behaviour? Either configuration to enforce "minimum" batch action or have timeout for current batch action so that if current action is hung then it will be retried later