Comment 18 for bug 1952363

We're adding a third vip for all ha applications, and found some of them to be affected by this (magnum, gnocchi, barbican) while others are not (aodh, heat, ceilometer, openstack-dashboard).

On the affected applications we have to remove the hacluster relation, which causes a service downtime, and then readd for it to work.

Meanwhile on the non affecteds the third vip is added, works fine but is not included on the proper resource group. e.g: grp_gnocchi_vips. We add it manually afterwards with crm configure edit.