[ 
https://issues.apache.org/jira/browse/QPID-6102?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14144460#comment-14144460
 ] 

Keith Wall commented on QPID-6102:
----------------------------------

Andrew, could you review this commit?

> Client connection closure whilst HA transitioning state can lead to 
> unexpected Broker shutdown
> ----------------------------------------------------------------------------------------------
>
>                 Key: QPID-6102
>                 URL: https://issues.apache.org/jira/browse/QPID-6102
>             Project: Qpid
>          Issue Type: Bug
>          Components: Java Broker
>    Affects Versions: 0.31
>            Reporter: Keith Wall
>            Assignee: Andrew MacBean
>             Fix For: 0.31
>
>         Attachments: 
> 0001-QPID-6102-Java-Broker-HA-Prevent-InsufficientReplica.patch
>
>
> Whilst testing the transfer master operation whilst the group was underload I 
> ran into the following exception causing unexpected  shutdown of the previous 
> Master.
> # Bring up a group of four
> # Organise for a publisher/consumer on separate connections steadily 
> consuming messages.
> # Use transfer master to move the mastership from one node to another.
> # Following exception observed on the node from which the mastership was 
> being moved.
> {noformat}
> ########################################################################
> #
> # Unhandled Exception com.sleepycat.je.rep.MasterReplicaTransitionException: 
> (JE 5.0.97) Environment must be closed, caused by: 
> com.sleepycat.je.rep.MasterReplicaTransitionException: Environment invalid 
> because of previous exception: (JE 5.0.97) 
> mars(4):/home/keith/trunk/qpid/qpid/java/systests/target/qpid-broker/0.32-SNAPSHOT/work4/mars/config
>  com.sleepycat.je.rep.stream.MasterStatus$MasterSyncException: Master change. 
> Node master id: mars(4) Group master id: earth(3) 
> MASTER_TO_REPLICA_TRANSITION: This node was a master and must reinitialize 
> internal state to become a replica. The application must close and reopen all 
> Environment handles. Environment is invalid and must be closed. in Thread 
> IoReceiver - /192.168.0.1:55912
> #
> # Exiting
> #
> ########################################################################
> com.sleepycat.je.rep.MasterReplicaTransitionException: (JE 5.0.97) 
> Environment must be closed, caused by: 
> com.sleepycat.je.rep.MasterReplicaTransitionException: Environment invalid 
> because of previous exception: (JE 5.0.97) 
> mars(4):/home/keith/trunk/qpid/qpid/java/systests/target/qpid-broker/0.32-SNAPSHOT/work4/mars/config
>  com.sleepycat.je.rep.stream.MasterStatus$MasterSyncException: Master change. 
> Node master id: mars(4) Group master id: earth(3) 
> MASTER_TO_REPLICA_TRANSITION: This node was a master and must reinitialize 
> internal state to become a replica. The application must close and reopen all 
> Environment handles. Environment is invalid and must be closed.
>         at 
> com.sleepycat.je.rep.MasterReplicaTransitionException.wrapSelf(MasterReplicaTransitionException.java:50)
>         at 
> com.sleepycat.je.dbi.EnvironmentImpl.checkIfInvalid(EnvironmentImpl.java:1594)
>         at com.sleepycat.je.Environment.checkEnv(Environment.java:2225)
>         at 
> com.sleepycat.je.Environment.beginTransactionInternal(Environment.java:1348)
>         at 
> com.sleepycat.je.Environment.beginTransaction(Environment.java:1319)
>         at 
> org.apache.qpid.server.store.berkeleydb.AbstractBDBMessageStore.removeMessage(AbstractBDBMessageStore.java:376)
>         at 
> org.apache.qpid.server.store.berkeleydb.AbstractBDBMessageStore$StoredBDBMessage.remove(AbstractBDBMessageStore.java:1436)
>         at 
> org.apache.qpid.server.message.AbstractServerMessageImpl.decrementReference(AbstractServerMessageImpl.java:101)
>         at 
> org.apache.qpid.server.message.AbstractServerMessageImpl.access$500(AbstractServerMessageImpl.java:37)
>         at 
> org.apache.qpid.server.message.AbstractServerMessageImpl$Reference.release(AbstractServerMessageImpl.java:275)
>         at 
> org.apache.qpid.server.exchange.AbstractExchange$2.onRollback(AbstractExchange.java:571)
>         at 
> org.apache.qpid.server.txn.LocalTransaction.doRollbackActions(LocalTransaction.java:270)
>         at 
> org.apache.qpid.server.txn.LocalTransaction.rollback(LocalTransaction.java:374)
>         at 
> org.apache.qpid.server.protocol.v0_8.AMQChannel.close(AMQChannel.java:739)
>         at 
> org.apache.qpid.server.protocol.v0_8.AMQChannel.close(AMQChannel.java:715)
>         at 
> org.apache.qpid.server.protocol.v0_8.AMQProtocolEngine.closeAllChannels(AMQProtocolEngine.java:942)
>         at 
> org.apache.qpid.server.protocol.v0_8.AMQProtocolEngine.closeSession(AMQProtocolEngine.java:981)
>         at 
> org.apache.qpid.server.protocol.v0_8.AMQProtocolEngine$3.run(AMQProtocolEngine.java:1035)
>         at java.security.AccessController.doPrivileged(Native Method)
>         at javax.security.auth.Subject.doAs(Subject.java:356)
>         at 
> org.apache.qpid.server.protocol.v0_8.AMQProtocolEngine.runAsSubject(AMQProtocolEngine.java:223)
>         at 
> org.apache.qpid.server.protocol.v0_8.AMQProtocolEngine.closeSession(AMQProtocolEngine.java:1030)
>         at 
> org.apache.qpid.server.protocol.v0_8.AMQProtocolEngine.closed(AMQProtocolEngine.java:1299)
>         at 
> org.apache.qpid.server.protocol.MultiVersionProtocolEngine.closed(MultiVersionProtocolEngine.java:117)
>         at 
> org.apache.qpid.transport.network.io.IoReceiver.run(IoReceiver.java:222)
>         at java.lang.Thread.run(Thread.java:745)
> Caused by: com.sleepycat.je.rep.MasterReplicaTransitionException: Environment 
> invalid because of previous exception: (JE 5.0.97) 
> mars(4):/home/keith/trunk/qpid/qpid/java/systests/target/qpid-broker/0.32-SNAPSHOT/work4/mars/config
>  com.sleepycat.je.rep.stream.MasterStatus$MasterSyncException: Master change. 
> Node master id: mars(4) Group master id: earth(3) 
> MASTER_TO_REPLICA_TRANSITION: This node was a master and must reinitialize 
> internal state to become a replica. The application must close and reopen all 
> Environment handles. Environment is invalid and must be closed.
>         at 
> com.sleepycat.je.rep.impl.node.FeederManager.runFeeders(FeederManager.java:458)
>         at com.sleepycat.je.rep.impl.node.RepNode.run(RepNode.java:1463)
> Caused by: com.sleepycat.je.rep.stream.MasterStatus$MasterSyncException: 
> Master change. Node master id: mars(4) Group master id: earth(3)
>         at 
> com.sleepycat.je.rep.stream.MasterStatus.assertSync(MasterStatus.java:107)
>         at 
> com.sleepycat.je.rep.impl.node.FeederManager.runFeeders(FeederManager.java:390)
>         ... 1 more
> {noformat}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org
For additional commands, e-mail: dev-h...@qpid.apache.org

Reply via email to