[
https://issues.apache.org/jira/browse/QPID-8749?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18100720#comment-18100720
]
ASF GitHub Bot commented on QPID-8749:
--------------------------------------
dakirily opened a new pull request, #423:
URL: https://github.com/apache/qpid-broker-j/pull/423
This PR addresses JIRA
[QPID-8749](https://issues.apache.org/jira/browse/QPID-8749), fixing the issue
with AMQP connections being stuck while processing pending writes
> [Broker-J] AMQP connections may not close while processing pending writes
> -------------------------------------------------------------------------
>
> Key: QPID-8749
> URL: https://issues.apache.org/jira/browse/QPID-8749
> Project: Qpid
> Issue Type: Bug
> Components: Broker-J
> Affects Versions: qpid-java-broker-10.0.1
> Reporter: Daniil Kirilyuk
> Priority: Minor
> Fix For: qpid-java-broker-10.1.0
>
>
> Idle AMQP connections could not be closed sometimes, producing following
> stacktraces in thread dumps:
> {code:java}
> "IO-/XX.XXX.XX.XXX:45532" #159java.lang.Thread.State: RUNNABLE
> cpu=169433659.40mselapsed=416877.64s
> at java.util.Collections.addAll(Collections.java:5867)at
> org.apache.qpid.server.bytebuffer.QpidByteBufferFactory.write(...)at
> org.apache.qpid.server.transport.NonBlockingConnection.writeToTransport(...)at
> org.apache.qpid.server.transport.NonBlockingConnection.processPending(...)
> {code}
> The issue seems to be caused by race condition when NonBlockingConnection
> processes pending writes. In NonBlockingConnection#doWork() value of the
> _closed variable may change between the check and processPending() call. The
> _scheduledTime is not reset and may become stale. Additionally the writing
> process should be optimized by batching the byte buffers.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
