Yolean/kubernetes-kafka

ZooKeeper produce a zombie processes

ilya-shaydullin opened this issue · 4 comments

Hello!

Every 5 minutes created ~30 zombie processes

5697 ?        Ssl    0:10  \_ /opt/java/openjdk/bin/java -Xmx512M -Xms512M -server -XX:+UseG1GC -XX:MaxGCPauseMillis=20 -XX:InitiatingHeapOccupancyPercent=35 -XX:+ExplicitGCInvokesConcurrent -Djava.awt.headless=true -Xlog:gc*:file=/opt/kafka/bin/../logs/zookeeper-gc.log:
 8891 ?        Z      0:00  |   \_ [sh] <defunct>
 8928 ?        Z      0:00  |   \_ [sh] <defunct>
 8948 ?        Z      0:00  |   \_ [sh] <defunct>
 9300 ?        Z      0:00  |   \_ [sh] <defunct>
 9340 ?        Z      0:00  |   \_ [sh] <defunct>
 9361 ?        Z      0:00  |   \_ [sh] <defunct>
 9712 ?        Z      0:00  |   \_ [sh] <defunct>
 9751 ?        Z      0:00  |   \_ [sh] <defunct>
 9771 ?        Z      0:00  |   \_ [sh] <defunct>
10467 ?        Z      0:00  |   \_ [sh] <defunct>
10512 ?        Z      0:00  |   \_ [sh] <defunct>
10532 ?        Z      0:00  |   \_ [sh] <defunct>
10912 ?        Z      0:00  |   \_ [sh] <defunct>
10960 ?        Z      0:00  |   \_ [sh] <defunct>
10980 ?        Z      0:00  |   \_ [sh] <defunct>
11323 ?        Z      0:00  |   \_ [sh] <defunct>
11360 ?        Z      0:00  |   \_ [sh] <defunct>
11382 ?        Z      0:00  |   \_ [sh] <defunct>
11762 ?        Z      0:00  |   \_ [sh] <defunct>
11798 ?        Z      0:00  |   \_ [sh] <defunct>
11818 ?        Z      0:00  |   \_ [sh] <defunct>
12162 ?        Z      0:00  |   \_ [sh] <defunct>
12202 ?        Z      0:00  |   \_ [sh] <defunct>
12252 ?        Z      0:00  |   \_ [sh] <defunct>
13302 ?        Z      0:00  |   \_ [sh] <defunct>
13370 ?        Z      0:00  |   \_ [sh] <defunct>
13409 ?        Z      0:00  |   \_ [sh] <defunct>
14126 ?        Z      0:00  |   \_ [sh] <defunct>
14192 ?        Z      0:00  |   \_ [sh] <defunct>
14212 ?        Z      0:00  |   \_ [sh] <defunct>
14529 ?        Z      0:00  |   \_ [sh] <defunct>
14600 ?        Z      0:00  |   \_ [sh] <defunct>
14620 ?        Z      0:00  |   \_ [sh] <defunct>
14938 ?        Z      0:00  |   \_ [sh] <defunct>
......

hi,
any news here? I see the same on my cluster k8s 19.7 with containerD .
did u found a workaround or a fix for it?

hi,
any news here? I see the same on my cluster k8s 19.7 with containerD .
did u found a workaround or a fix for it?

This solution worked for me

https://www.back2code.me/2020/02/zombie-processes-back-in-k8s/

More info that i found
https://srcco.de/posts/kubernetes-liveness-probes-are-dangerous.html
https://www.youtube.com/watch?v=QKI-JRs2RIE&ab_channel=CNCF%5BCloudNativeComputingFoundation%5D

10x! it solved it for me also

See also #337 and #340.