Atlassian support's suggested method of recovering from a cluster panic is to restart tomcat (and perhaps blow away the work/Catalina/localhost dir).

This is not a good option for those that are trying to maintain 24/7 operation of Confluence, which is one of the primary reasons we bought confluence massive (in addition to being able to scale up as needed).

It would be much better if Atlassian provided a suggested method of having confluence (or some external container app, process, script) restart itself automatically if it goes into cluster panic.