[
https://issues.apache.org/jira/browse/LOG4J2-3619?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17633986#comment-17633986
]
Gary ,Verwers commented on LOG4J2-3619:
---------------------------------------
[~Ragi] I found the consumer thread waiting for access to the next log event.
What appears to have happened is the event printing process started on the
first stackoverflow event and it failed. But it failed in a way that it did
not release the lock. When the consumer comes around again it stalls waiting
for a lock that will never be released. You can see the Batch event processor
in a wait state even though the ringbuffer is full of log events it has not
processed.
> Log4j stops logging when the log message contains a
> java.lang.StackOverflowError
> --------------------------------------------------------------------------------
>
> Key: LOG4J2-3619
> URL: https://issues.apache.org/jira/browse/LOG4J2-3619
> Project: Log4j 2
> Issue Type: Bug
> Components: Core
> Affects Versions: 2.17.2
> Environment: We run Weblogic 12 in a Linux environment.
> Reporter: Gary ,Verwers
> Priority: Major
> Attachments: BatchEventProcessor.png, CauseEvent.png,
> IncomingReferences.png, cusor.png, gatinngSequence.png, log4j2.xml,
> thread-dump-analyses.png, threadDump.txt
>
>
> The RingBuffer Consumer stop printing logs when it comes to a log event with
> an exception connected with a StackOverFlowError attached. We have had
> this occur several dozen times.
> Both the cursor and the gate point at the same place in the event buffer and
> very quickly all the threads on the server are frozen waiting for the ring
> buffer to clear up space. But the consumer is just stuck waiting a not
> printing event. It has failed with a stack overflow from spring MVC mapping
> and it has failed with a stack overflow from Jackson JSON ObjectMapper with
> object level recursion triggering a stack overflow.
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
