I'm running an experiment with storm and wanted to see how it behaved when
a supervisor gets killed. Would the running bolts move to other
supervisors in the cluster?
I'm unable to restart the killed supervisor for reasons I don't quite
follow. Here's the stack trace I get when I try to start it back up:
java.lang.RuntimeException: Version already exists or data already exists
at backtype.storm.utils.VersionedStore.createVersion(VersionedStore.java:68)
at backtype.storm.utils.VersionedStore.createVersion(VersionedStore.java:62)
at backtype.storm.utils.LocalState.persist(LocalState.java:45)
at backtype.storm.utils.LocalState.put(LocalState.java:34)
at
backtype.storm.daemon.supervisor$mk_synchronize_supervisor$this__4692.invoke(supervisor.clj:312)
at backtype.storm.event$event_manager$fn__2484.invoke(event.clj:24)
at clojure.lang.AFn.run(AFn.java:24)
at java.lang.Thread.run(Thread.java:662)
I tried clearing out the contents of storm.local.dir and the problem occurs
each time.
