Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Fixed
-
Discovery Impl 1.0.10
-
None
Description
There is a possibility for a race condition in VotingHandler.cleanupTimedoutVotings when multiple nodes in a cluster try to delete a timed-out vote. Currently, this results in an error, eg:
<...> *ERROR* [pool-6-thread-4] org.apache.sling.discovery.impl.common.View remove: Could not remove node: org.apache.sling.api.resource.PersistenceException: Unable to delete resource at /var/discovery/impl/ongoingVotings/ <...> . Resource does not exist.
This error is misleading, as this is just two instances trying to delete the same resource - which should not cause an error at all. An info though would be appropriate.