There is another way. Migrate to another node in the cluster and start it. Target node must be a node which never had the same issue. If not it will not work. Not an good way, but a temporary solution. I have 5 nodes in each cluster, so i move around.