You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by "Alex Rudyy (Jira)" <ji...@apache.org> on 2019/10/01 08:02:00 UTC
[jira] [Created] (QPID-8366) [Broker-J] The loss of BDB HA majority
on invocation of house keeping operations can crash the broker
Alex Rudyy created QPID-8366:
--------------------------------
Summary: [Broker-J] The loss of BDB HA majority on invocation of house keeping operations can crash the broker
Key: QPID-8366
URL: https://issues.apache.org/jira/browse/QPID-8366
Project: Qpid
Issue Type: Task
Components: Broker-J
Affects Versions: qpid-java-broker-7.1.4, qpid-java-broker-7.1.3, qpid-java-broker-7.0.8, qpid-java-broker-7.1.2, qpid-java-broker-7.1.1, qpid-java-broker-7.0.7, qpid-java-broker-7.0.6, qpid-java-broker-7.0.5, qpid-java-broker-7.0.4, qpid-java-broker-7.1.0
Reporter: Alex Rudyy
The {{ConnectionScopedRuntimeException}} thrown from {{VirtualHost}} {{House Keeping}} thread on invocation of {{MessageStore}} operations like {{checkMessageStatus}} can crash the broker. An example of such exception stack trace (from Qpid Broker version 7.0.6) is provided below:
{noformat}
2019-09-27 07:53:38,168 ERROR [virtualhost-test-pool-1] (o.a.q.s.Main) - Uncaught exception, shutting down.
org.apache.qpid.server.util.ConnectionScopedRuntimeException: Required number of nodes not reachable
at org.apache.qpid.server.store.berkeleydb.replication.ReplicatedEnvironmentFacade.handleDatabaseException(ReplicatedEnvironmentFacade.java:495)
at org.apache.qpid.server.store.berkeleydb.replication.ReplicatedEnvironmentFacade.commit(ReplicatedEnvironmentFacade.java:332)
at org.apache.qpid.server.store.berkeleydb.AbstractBDBMessageStore.removeMessage(AbstractBDBMessageStore.java:288)
at org.apache.qpid.server.store.berkeleydb.AbstractBDBMessageStore$StoredBDBMessage.remove(AbstractBDBMessageStore.java:1090)
at org.apache.qpid.server.message.AbstractServerMessageImpl.decrementReference(AbstractServerMessageImpl.java:118)
at org.apache.qpid.server.message.AbstractServerMessageImpl.access$500(AbstractServerMessageImpl.java:37)
at org.apache.qpid.server.message.AbstractServerMessageImpl$Reference.release(AbstractServerMessageImpl.java:309)
at org.apache.qpid.server.queue.QueueEntryImpl.dispose(QueueEntryImpl.java:557)
at org.apache.qpid.server.queue.QueueEntryImpl.delete(QueueEntryImpl.java:572)
at org.apache.qpid.server.queue.AbstractQueue$11.postCommit(AbstractQueue.java:1729)
at org.apache.qpid.server.txn.AutoCommitTransaction.dequeue(AutoCommitTransaction.java:92)
at org.apache.qpid.server.queue.AbstractQueue.dequeueEntry(AbstractQueue.java:1722)
at org.apache.qpid.server.queue.AbstractQueue.dequeueEntry(AbstractQueue.java:1717)
at org.apache.qpid.server.queue.AbstractQueue.deleteEntry(AbstractQueue.java:1761)
at org.apache.qpid.server.queue.AbstractQueue.checkMessageStatus(AbstractQueue.java:2165)
at org.apache.qpid.server.virtualhost.AbstractVirtualHost$VirtualHostHouseKeepingTask.execute(AbstractVirtualHost.java:1965)
at org.apache.qpid.server.virtualhost.HouseKeepingTask$1.run(HouseKeepingTask.java:56)
at java.security.AccessController.doPrivileged(Native Method)
at org.apache.qpid.server.virtualhost.HouseKeepingTask.run(HouseKeepingTask.java:51)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308)
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180)
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at org.apache.qpid.server.bytebuffer.QpidByteBufferFactory.lambda$null$0(QpidByteBufferFactory.java:464)
at java.lang.Thread.run(Thread.java:748)
Caused by: com.sleepycat.je.rep.InsufficientAcksException: (JE 7.4.5) Transaction: -3459038252 VLSN: 10,380,435,448, initiated at: 07:53:20. Insufficient acks for policy:SIMPLE_MAJORITY. Need replica acks: 2. Missing replica acks: 2. Timeout: 15000ms. FeederState=acc3_2(3)[MASTER]
Current feeds:
acc3_1: feederVLSN=10,380,435,456 replicaTxnEndVLSN=10,380,435,396
acc3: feederVLSN=10,380,435,456 replicaTxnEndVLSN=10,380,435,396
at com.sleepycat.je.rep.impl.node.DurabilityQuorum.ensureSufficientAcks(DurabilityQuorum.java:205)
at com.sleepycat.je.rep.stream.FeederTxns.awaitReplicaAcks(FeederTxns.java:189)
at com.sleepycat.je.rep.impl.RepImpl.postLogCommitHookInternal(RepImpl.java:1426)
at com.sleepycat.je.rep.impl.RepImpl.postLogCommitHook(RepImpl.java:1385)
at com.sleepycat.je.rep.txn.MasterTxn.postLogCommitHook(MasterTxn.java:228)
at com.sleepycat.je.txn.Txn.commit(Txn.java:772)
at com.sleepycat.je.Transaction.doCommit(Transaction.java:621)
at com.sleepycat.je.Transaction.commit(Transaction.java:401)
at org.apache.qpid.server.store.berkeleydb.replication.ReplicatedEnvironmentFacade.commit(ReplicatedEnvironmentFacade.java:328)
... 25 common frames omitted
{noformat}
The issue reported with the stack trace above occurred when BDB HA {{VirtualHost}} was trying to delete an expired message, but its BDB HA group lost the majority when the {{VirtualHost}} tried to commit a BDB HA transaction for message deletion operation. The majority loss is communicated as {{ConnectionScopeRuntimeException}} to the caller. It seems we need to catch and handle {{ConnectionScopeRuntimeException}} in House Keeping operations.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
For additional commands, e-mail: dev-help@qpid.apache.org