[
https://issues.apache.org/jira/browse/DIRMINA-1111?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16844793#comment-16844793
]
Guus der Kinderen commented on DIRMINA-1111:
--------------------------------------------
{{rewind()}} looks like a simpler solution than resetting-if-marked, but
there's a small difference in that the position would always revert to 0,
instead of the last marked position. I'm not sure if that makes any difference
to code that depends on this.
In any case, preventing the {{InvalidMarkException}} would be achieved by both,
and that is what seems to prevent the CPU-churning. I'd be happy with either
solution.
> 100% CPU (epoll bug) on 2.1.x, Linux only
> -----------------------------------------
>
> Key: DIRMINA-1111
> URL: https://issues.apache.org/jira/browse/DIRMINA-1111
> Project: MINA
> Issue Type: Bug
> Affects Versions: 2.1.2
> Reporter: Guus der Kinderen
> Priority: Major
> Attachments: image-2019-05-21-11-37-41-931.png
>
>
> We're getting reports
> [reports|https://discourse.igniterealtime.org/t/openfire-4-3-2-cpu-goes-at-100-after-a-few-hours-on-linux/85119/13]
> of a bug that causes 100% CPU usage on Linux (the problem does not occur on
> Windows).
> This problem occurs in 2.1.2. but does _not_ occur in 2.0.21.
> Is this a regression of the epoll selector bug in DIRMINA-678 ?
> A stack trace of one of the spinning threads:
> {code}"NioProcessor-3" #55 prio=5 os_prio=0 tid=0x00007f0408002000 nid=0x2a6a
> runnable [0x00007f0476dee000]
> java.lang.Thread.State: RUNNABLE
> at sun.nio.ch.EPollArrayWrapper.epollWait(Native Method)
> at sun.nio.ch.EPollArrayWrapper.poll(EPollArrayWrapper.java:269)
> at sun.nio.ch.EPollSelectorImpl.doSelect(EPollSelectorImpl.java:93)
> at sun.nio.ch.SelectorImpl.lockAndDoSelect(SelectorImpl.java:86)
> - locked <0x00000004c486b748> (a sun.nio.ch.Util$3)
> - locked <0x00000004c486b738> (a java.util.Collections$UnmodifiableSet)
> - locked <0x00000004c420ccb0> (a sun.nio.ch.EPollSelectorImpl)
> at sun.nio.ch.SelectorImpl.select(SelectorImpl.java:97)
> at
> org.apache.mina.transport.socket.nio.NioProcessor.select(NioProcessor.java:112)
> at
> org.apache.mina.core.polling.AbstractPollingIoProcessor$Processor.run(AbstractPollingIoProcessor.java:616)
> at
> org.apache.mina.util.NamePreservingRunnable.run(NamePreservingRunnable.java:64)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> at java.lang.Thread.run(Thread.java:748)
> Locked ownable synchronizers:
> - <0x00000004c4d03530> (a
> java.util.concurrent.ThreadPoolExecutor$Worker){code}
> More info is available at
> https://discourse.igniterealtime.org/t/openfire-4-3-2-cpu-goes-at-100-after-a-few-hours-on-linux/85119/13
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
