What Can I Do If launcher-job Is Terminated by Yarn When a Flink Task Is Submitted?
Symptom
Launcher-job is terminated by Yarn due to insufficient heap size when a Flink task is submitted on the management plane.
Solution
Increase the heap size of the launcher-job queue.
- Log in to the active OMS node as user omm.
- Change the value of job.launcher.resource.memory.mb in /opt/executor/webapps/executor/WEB-INF/classes/servicebroker.xml to 2048.
- Run the sh /opt/executor/bin/restart-executor.sh command to restart the executor process.
Parent topic: Job Development
- Symptom
- Solution