Improve failed mysql node removal time in HA deploys.
authorChris Jones <cmsj@tenshu.net>
Fri, 4 Nov 2016 10:01:24 +0000 (10:01 +0000)
committerChris Jones <cmsj@tenshu.net>
Tue, 8 Nov 2016 10:44:14 +0000 (10:44 +0000)
commit943e49435621062a1c898de4b2edccf48c59fc16
treeb0336674ac498da80042c8858b86607c4bfd2b2c
parent2b92f3f7739f655ba88154058730d6ac07186c0e
Improve failed mysql node removal time in HA deploys.

In HA deployments, we now check mysql nodes every 1s and removed them
immediately if they are failed. Previously we would check every 2s and
allow them to fail 5 checks before being removed, producing errors from
other OpenStack services for 10s, which causes confusion and delay for
operators.
Additionally, these check options are now also a class parameter so can
be overridden by operators.

Closes-Bug: #1639189

Change-Id: I0b915f790ae5a4b018a212d3aa83cca507be05e9
manifests/haproxy.pp