cluster-recheck-interval is too long

[Andrea Ieri]

The hacluster charm does not currently override the upstream default of the cluster recheck interval, and that interval is therefore set to 15 minutes. Due to pacemaker being inherently event-driven[0], having a reasonably low default is a precondition for consistently supporting quicker timers (e.g. for allowing failcounts to timely expire - lp#1802310).
I therefore suggest we lower the default to 60 seconds.

[0] https://clusterlabs.org/pacemaker/doc/en-US/Pacemaker/1.1/html-single/Pacemaker_Explained/#s-rules-recheck

[Andrea Ieri (aieri)]

I have proposed a patch in https://review.openstack.org/#/c/619590/

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/619590
Committed: https://git.openstack.org/cgit/openstack/charm-hacluster/commit/?id=81b386d37955f3e65b10657b8939f7f110bbfa1f
Submitter: Zuul
Branch: master

commit 81b386d37955f3e65b10657b8939f7f110bbfa1f
Author: Andrea Ieri <email address hidden>
Date: Thu Nov 22 15:29:54 2018 +0100

    Lower cluster-recheck-interval to 1 minute

    The upstream default for cluster-recheck-interval is 15 minutes. This
    renders any quick timer (i.e. failure-timeout) effectively meaningless,
    as in many cases their expiration will only be checked 4 times per hour.
    This commit lowers cluster-recheck-interval to 1 minute, which ensures a
    fine enough granularity for specifying timers on a scale of minutes,
    without overly stressing the policy engine.

    Change-Id: I756d53e7267e6fa1b6fb348e219e30bafe757360
    Closes-Bug: 1804667