Hi, Using ActiveMQ-5.11.1. Topology : [shared-storage-master-slave ] with kahadb Kahadb store location is on NFS.
Yesterday we faced a strange issue where suddenly failover broker opened it's transport connector and started accepting connections. Both primary and failover were serving the clients. Please note that there were not restart. Details: >From the logs: Failvoer logs says that it checks each 10 seconds to acquire the lock but for that time, it didn't check for 3 minutes and then suddenly acquires the lock over kahadb. And on the other side primary broker was continuously serving the clients. *This is very strange but how both primary and failvoer can serve the clients simultaneously ? Is there some bug in kahadb locking procedure? * Failover logs: [20150425 *08:02:57*.356 EDT (main) org.apache.activemq.store.SharedFileLocker#doStart 58 INFO] - Database /local/pod/activemq/kahadb/lock is locked... wait ing 10 seconds for the database to be unlocked. Reason: java.io.IOException: File '/local/pod/activemq/kahadb/lock' could not be locked. [20150425* 08:03:07*.359 EDT (main) org.apache.activemq.store.SharedFileLocker#doStart 58 INFO] - Database /local/pod/activemq/kahadb/lock is locked... wait ing 10 seconds for the database to be unlocked. Reason: java.io.IOException: File '/local/pod/activemq/kahadb/lock' could not be locked. [20150425 *08:03:17*.365 EDT (main) org.apache.activemq.store.SharedFileLocker#doStart 58 INFO] - Database /local/pod/activemq/kahadb/lock is locked... wait ing 10 seconds for the database to be unlocked. Reason: java.io.IOException: File '/local/pod/activemq/kahadb/lock' could not be locked. [20150425 *08:06:21*.963 EDT (main) org.apache.activemq.store.kahadb.MessageDatabase$Metadata#read 168 INFO] - KahaDB is version 5 [20150425 *08:06:21*.983 EDT (main) org.apache.activemq.store.kahadb.MessageDatabase#recover 603 INFO] - Recovering from the journal ... [20150425 08:06:22.052 EDT (main) org.apache.activemq.store.kahadb.MessageDatabase#recover 616 INFO] - Recovery replayed 5670 operations from the journal in 0.085 seconds. [20150425 08:06:25.426 EDT (main) org.apache.activemq.store.kahadb.MessageDatabase$Metadata#read 168 INFO] - KahaDB is version 5 [20150425 08:06:25.457 EDT (main) org.apache.activemq.store.kahadb.MessageDatabase#recover 603 INFO] - Recovering from the journal ... [20150425 08:06:25.458 EDT (main) org.apache.activemq.store.kahadb.MessageDatabase#recover 616 INFO] - Recovery replayed 33 operations from the journal in 0.009 seconds. Thanks, Anuj -- View this message in context: http://activemq.2283324.n4.nabble.com/ActiveMQ-slave-broker-opening-it-s-connector-tp4695677.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
