centos9: ha resourced not restarted as expected during minor update
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tripleo |
Fix Released
|
High
|
Damien Ciabrini |
Bug Description
With pacemaker 2.1, various attributes and their value have changed in the CIB, and we don't track the necessary information anymore in our resource scripts:
. The master/slave terminology has been superseded by promoted/unpromoted role values.
. The class name of the resources configured in the cluster start with 'ocf:' rather than 'ocf::'.
This naming changes made the tracking of the resource restart fail:
# TRIPLEO_
Mon Sep 6 11:06:09 AM UTC 2021: Restarting ovn-dbs-bundle locally on 'undercloud-0'
ovn-dbs-bundle successfully restarted
Mon Sep 6 11:06:16 AM UTC 2021: Waiting until ovndb_servers has restarted on undercloud-0 and is in state Unpromoted
Mon Sep 6 11:06:16 AM UTC 2021: Will probe resource state with the following XPath pattern: //bundle/
Mon Sep 6 11:06:16 AM UTC 2021: Waiting for ovndb_servers to transition to role Unpromoted on undercloud-0
Mon Sep 6 11:06:20 AM UTC 2021: Waiting for ovndb_servers to transition to role Unpromoted on undercloud-0
Mon Sep 6 11:06:24 AM UTC 2021: Waiting for ovndb_servers to transition to role Unpromoted on undercloud-0
[...]
Fix proposed to branch: master /review. opendev. org/c/openstack /tripleo- heat-templates/ +/807572
Review: https:/