Monitoring Tools Pacemaker check ineffective with containerized services

[Matt Wisch]

Starting with the Pike release of Red Hat OSP/RDO/TripleO deployments, many services, including those managed by Pacemaker are deployed inside of Docker containers.

The output structure of pcs status or crm_mon -1 is different for these services making the check non-functional.

pacemaker_host_check.py should support passing in Docker bundle resource names and successfully returning a status.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/575187

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/575187
Committed: https://git.openstack.org/cgit/openstack/osops-tools-monitoring/commit/?id=f7fb864ea1b02888b6cbc0d4b01107ee210afdfc
Submitter: Zuul
Branch: master

commit f7fb864ea1b02888b6cbc0d4b01107ee210afdfc
Author: Matt Wisch <email address hidden>
Date: Wed Jun 13 12:00:53 2018 -0600

    Add support for Pacemaker managed Docker container resources

    Added ability to pass in the resource name of a docker bundle
    and successfully filter the host list for systems where the
    container is in a Master/Slave/Started state.

    Also corrected the error handling for non 0 return codes on
    the pcs status or crm_mon commands. Errors for pcs are returned
    on stderr and were not being captured in the command output.

    Change-Id: Id3a858a785d5062feda59362c1a7f69ab04b861e
    Closes-Bug: 1776731