openstack-cinder-volume preupgrade validation fails in HA environment because the service is running in active/passive mode

[Marius Cornea]

openstack-cinder-volume preupgrade validation fails in HA environment because the service is running in active/passive mode so the service is active on a single controller node at a time:

[stack@undercloud-0 ~]$ ## controller-0
[stack@undercloud-0 ~]$ ssh heat-admin@192.168.0.11 /usr/bin/systemctl show 'openstack-cinder-volume' --property ActiveState | grep '\bactive\b'
ActiveState=active
[stack@undercloud-0 ~]$ ## controller-1
[stack@undercloud-0 ~]$ ssh heat-admin@192.168.0.17 /usr/bin/systemctl show 'openstack-cinder-volume' --property ActiveState | grep '\bactive\b'
[stack@undercloud-0 ~]$ ## controller-2
[stack@undercloud-0 ~]$ ssh heat-admin@192.168.0.27 /usr/bin/systemctl show 'openstack-cinder-volume' --property ActiveState | grep '\bactive\b'

[Marius Cornea (mcornea)]

We'll probably see this issue with all the pacemaker services which run in active/passive manner: cinder-backup, manila-share

[Sofer Athlan-Guyot (sofer-athlan-guyot)]

Hi,

I think this is resolved by https://review.openstack.org/#/c/440488/