Most of the RFPs we get from VNF vendors and communication providers is to provide high-availability for VM instances in the following faulty conditions:
1. VM failure,
2. vSwitch failure
3. Connectivity failure
4. Compute node failure
RH has documented procedure but is not OSPd aware hence not supported in production. This feature request track implementation of VM high availability through OSPd.
https://access.redhat.com/solutions/2661641https://access.redhat.com/documentation/en/red-hat-openstack-platform/9/single/high-availability-for-compute-instances/
“While implementing High Availability for Instances, no upgrade or scale-up operations are possible. Any attempts to do so will fail.
In addition, enabling High Availability for Instances will prevent you from using the director to upgrade the overcloud in the future. This applies to both major and minor upgrades.
As upgrade is must for any production types deployment, we need this implementation through OSPd.