Task #3435
Story #3434: Investigate Timeout Problems on Staging Hazelcast
Examine the Locking exception thrown during call to updateReplicationMetadata()
100%
Description
The following is the sequence of events during which the Error occurred.
UNM heartbeat -> ORC no response
PUT /cn/v1/replicaMetadata/doi:10.6085%2FAA%2Fsbclter.451.26 -> UNM
UNM Throws -> IllegalMonitorStateException: Current thread is not owner of the lock!
UNM WARNS -> Master node has timed out its heartbeat and will be removed
UNM REMOVES -> ORC
UNM SETS -> Master address to UNM
PUT /cn/v1/replicaMetadata/doi:10.6085%2FAA%2Fsbclter.306.4 -> UNM
UNM Throws -> VersionMismatch: The requested system metadata version number 1 differs from the current version at 2
UNM Re-partitioning cluster data
UNM Merging data
UNM Sends -> Join ORC request
UNM Joins -> ORC
Although I do not know when the heartbeat monitor starts failing, from the default settings it would appear to be 300 seconds before the heartbeat timeout. Disconnect to reconnect took 15 seconds. The IllegalMonitorStateException occurred 5 seconds before the heartbeat timeout.
