The etcd restore procedure mentioned in below doc seems to be incomplete.

https://docs.redhat.com/en/documentation/openshift_container_platform/4.19/html-single/hosted_control_planes/index#hcp-backup-restore-on-premise

The control plane pods don't rollout automatically after following all the 4 steps. Below additional steps are required to make all the control plane pods for HCP to be running fine.

Rollout the hostedcluster manually:

oc annotate hostedcluster -n <hostedcluster-namespace> <hostedcluster-name> hypershift.openshift.io/restart-date=$(date --iso-8601=seconds)

The multus admission controller and network node identity pods still don't start.

Delete the pods for second and third members of etcd along with their PVCs:

oc delete -n $CONTROL_PLANE_NAMESPACE pvc/data-etcd-1 pod/etcd-1 --wait=false
oc delete -n $CONTROL_PLANE_NAMESPACE pvc/data-etcd-2 pod/etcd-2 --wait=false

Rollout the hostedcluster manually again:

oc annotate hostedcluster -n <hostedcluster-namespace> <hostedcluster-name> hypershift.openshift.io/restart-date=$(date --iso-8601=seconds) --overwrite

All the control plane pods start running after waiting for sometime.

    4.18.19

    100% in customer environment

Follow the doc on a baremetal HCP cluster.

https://docs.redhat.com/en/documentation/openshift_container_platform/4.19/html-single/hosted_control_planes/index#hcp-backup-restore-on-premise  

Some steps seem to be missing as the control plane pods don't start fine just by following the doc.    

Any missing steps to be added in the docs.

These steps were tested in customer environment and is is required to run the additional steps mentioned every time an etcd restore with the manual method.