question Re: unable to start nodemanager in CDH 5.16.2 in Support Questions

question Re: unable to start nodemanager in CDH 5.16.2 in Support Questions https://community.cloudera.com/t5/Support-Questions/unable-to-start-nodemanager-in-CDH-5-16-2/m-p/300154#M220061 Hello <a href="https://community.cloudera.com/t5/user/viewprofilepage/user-id/22673">@Nagamalleswara</a> , thank you for sharing with us the detailed description of how your NM seems to have issues on a certain node. Next time you hit this problem, can you please check if you see multiple NM instances on a single node?In some cases the CM agent might gets into a bad state. When you hit the issue, please run the following command and see if it showes 2 NM processes running on the hostssh into the host ;   become root # export SUPER_CONF=/var/run/cloudera-scm-agent/supervisor/supervisord.conf# /usr/lib64/cmf/agent/build/env/bin/supervisorctl -c $SUPER_CONF status52372-yarn-NODEMANAGER       RUNNING    pid 71771, uptime 2:21:1952723-yarn-NODEMANAGER       FATAL      Exited too quickly (processlog may have details) If so, it means that the CM-agent process was bad. In this case please:1) Proceed to kill the PID. Verify that the PID does not exist and neither is anything else listening on the port.2) <A href="https://docs.cloudera.com/documentation/enterprise/6/6.3/topics/cm_ag_agents.html" target="_self">Restart cloudera-scm-agent</A> service cloudera-scm-agent restart3) Start the RM process from CM Please share with us your findings. If you are not hitting the above described issue it still helps us to narrow down further this issue. Please also check if the NM heap size on this node is set to at least 1Gb just to rule out stop the world GC activity rendering the NM unresponsive. Thank you: Ferenc Tue, 21 Jul 2020 15:33:09 GMT Bender 2020-07-21T15:33:09Z