When restoring with `--force-new-cluster` there might still be a need to revision bump and mark the revision as compacted to reset watchers.

We have implemented this upstream last time only on etcdctl snapshot restores, we would also need to have this as a restore option in etcd itself.

This spike should bring this topic upstream, see if we can reuse the code paths also during the restart using --force-new-cluster.

-> Pitch the carry patch from https://github.com/openshift/etcd/pull/300 upstream