In case of an unclean shutdown, a dangling lock file is left in the global persistent location. Currently this is being ignored, causing a number of issues with rejoining the cluster.

Add an unclean-shutdown-action attribute to the global state configuration with the following possible values:

• NONE : Prevents startup of the cache manager if a dangling lock file is found in the persistent global state.
• FAIL: Clears the persistent global state if a dangling lock file is found in the persistent global state.
• PURGE: Clears the persistent global state if a dangling lock file is found in the persistent global state.