[
https://issues.apache.org/jira/browse/DIRMINA-1111?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16844815#comment-16844815
]
Jonathan Valliere commented on DIRMINA-1111:
--------------------------------------------
Emmanuel is trying to figure out why {{reset()}} was in there at all. Might
end up just removing that call. For the time being, this fixes the problem.
{code:java}
diff --git
a/mina-core/src/main/java/org/apache/mina/core/polling/AbstractPollingIoProcessor.java
b/mina-core/src/main/java/org/apache/mina/core/polling/AbstractPollingIoProcessor.java
index 1553a99..9501074 100644
---
a/mina-core/src/main/java/org/apache/mina/core/polling/AbstractPollingIoProcessor.java
+++
b/mina-core/src/main/java/org/apache/mina/core/polling/AbstractPollingIoProcessor.java
@@ -1170,7 +1170,7 @@
// The first unwritten empty buffer must be
// forwarded to the filter chain.
if (buf.hasRemaining()) {
- buf.reset();
+ buf.clear();
failedRequests.add(req);
} else {
IoFilterChain filterChain = session.getFilterChain();
diff --git
a/mina-core/src/main/java/org/apache/mina/core/write/DefaultWriteRequest.java
b/mina-core/src/main/java/org/apache/mina/core/write/DefaultWriteRequest.java
index 8324c6a..23604fa 100644
---
a/mina-core/src/main/java/org/apache/mina/core/write/DefaultWriteRequest.java
+++
b/mina-core/src/main/java/org/apache/mina/core/write/DefaultWriteRequest.java
@@ -226,7 +226,7 @@
if (message instanceof IoBuffer) {
// duplicate it, so that any modification made on it
// won't change the original message
- this.message = ((IoBuffer)message).duplicate();
+ this.message = ((IoBuffer)message).slice();
}
{code}
> 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)
