[
https://issues.apache.org/jira/browse/QPID-6386?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14317228#comment-14317228
]
Rob Godfrey commented on QPID-6386:
-----------------------------------
Sadly (but unsurprisingly) the stack traces don't really say much. The
IoSender is waiting until it is given more data... It's locked on an object
that should be notified if the sender is closed... and it should never get into
the wait if the IoSender close() method has been called. And IoSender.close()
should be being called immediately after the close command is sent on the wire
to the client.
Obviously I'm missing something, but I don't immediately see how it gets into
this state.
I'll have another look in the morning when maybe my brain is fresher
> IoSender thread leak
> --------------------
>
> Key: QPID-6386
> URL: https://issues.apache.org/jira/browse/QPID-6386
> Project: Qpid
> Issue Type: Bug
> Components: Java Broker
> Affects Versions: 0.30
> Environment: java version "1.7.0_71"
> OpenJDK Runtime Environment (rhel-2.5.3.2.el6_6-x86_64 u71-b14)
> OpenJDK 64-Bit Server VM (build 24.65-b04, mixed mode)
> Reporter: Mark Soderquist
> Attachments: Qpid6386Bug.java, screenshot-1.png
>
>
> We are running a client that opens and closes a connection for every message
> that it sends. It does this as fast as possible. When testing right after the
> server has started there is normally zero or one IoSender thread. After a
> several minutes there is always one IoSender and sometimes two. The count
> continues to increment one at a time roughly every five minutes. I've been
> running the server now for about 30 minutes with the client continually
> sending messages and we now have six "orphaned" threads. Stopping the client
> does not clear the threads.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
