[ 
https://issues.apache.org/jira/browse/KAFKA-13878?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Greg Harris updated KAFKA-13878:
--------------------------------
    Component/s: KafkaConnect

> Connect deadlock in WorkerConnector on desired state change
> -----------------------------------------------------------
>
>                 Key: KAFKA-13878
>                 URL: https://issues.apache.org/jira/browse/KAFKA-13878
>             Project: Kafka
>          Issue Type: Bug
>          Components: KafkaConnect
>            Reporter: Greg Harris
>            Priority: Major
>
> We've experienced multiple instances of deadlocks where the connector thread 
> is blocked indefinitely with the following stacktrace:
> {noformat}
> "connector-thread-myconnector" #2059323 prio=5 os_prio=0 cpu=123.43ms 
> elapsed=147352.41s tid=0x00007f5588160de0 nid=0x18be in Object.wait()  
> [0x00007f550e3fe000]
>    java.lang.Thread.State: WAITING (on object monitor)
>     at java.lang.Object.wait(java.base@16.0.2/Native Method)
>     - waiting on <no object reference available>
>     at java.lang.Object.wait(java.base@16.0.2/Object.java:320)
>     at 
> org.apache.kafka.connect.runtime.WorkerConnector.doRun(WorkerConnector.java:149)
>     - locked <0x00000005d2253818> (a 
> org.apache.kafka.connect.runtime.WorkerConnector)
>     at 
> org.apache.kafka.connect.runtime.WorkerConnector.run(WorkerConnector.java:118)
>     at 
> java.util.concurrent.Executors$RunnableAdapter.call(java.base@16.0.2/Executors.java:515)
>     at 
> java.util.concurrent.FutureTask.run(java.base@16.0.2/FutureTask.java:264)
>     at 
> java.util.concurrent.ThreadPoolExecutor.runWorker(java.base@16.0.2/ThreadPoolExecutor.java:1130)
>     at 
> java.util.concurrent.ThreadPoolExecutor$Worker.run(java.base@16.0.2/ThreadPoolExecutor.java:630)
>     at java.lang.Thread.run(java.base@16.0.2/Thread.java:831){noformat}
> This appears to be caused by a race condition where a notify() is never 
> delivered to a task, causing the connector thread to wait indefinitely for a 
> notify() call that never comes.
> This causes the connector to be unable to process further state changes, or 
> generate task configurations.



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

Reply via email to