CODEWITHSUNDEEP
hadoop
Abbas Gadhia
Abbas Gadhia

Reputation: 15090

Node Manager Stop - Hadoop 2.2.0

On invoking ./stop-yarn.sh, I get

"nodemanager did not stop gracefully after 5 seconds: killing with kill -9".

Is this something I should be concerned about? What should I do to avoid this error?

Upvotes: 4

Views: 6867

Answers (3)

Prashan Singh
Prashan Singh

Reputation: 21

The reason is because the Resource Manager (RM) is stopped before Node Manager (NM). The NM needs to unregister itself from the RM. So when NM tries to connect with RM (to unregister itself) it fails as RM is already stopped. And so a timeout occurs (after waiting 5 sec - mention in start-yarn.sh) causing NM to be killed and not allowing to stop gracefully.

Upvotes: 2

Md Danish Alam
Md Danish Alam

Reputation: 46

There is not way to do a graceful stop till yet. you can follow this link https://issues.apache.org/jira/browse/YARN-914 for tracking and feature changes.

Upvotes: 1

Alberto Ayora
Alberto Ayora

Reputation: 9

You could use "hadoop job" to see if you have jobs still running and use the option "-kill " http://hadoop.apache.org/docs/r1.0.4/commands_manual.html#job

Upvotes: 0

Related Questions