내용으로 건너뛰기

Kill yarn jobs that not running automatically

1

I want to know if the option exist in EMR or automated way to kill the spark job that says initial job not accepted any resources. Some of my cluster is transient that run off batch hour without progressing stuck in accepted state.

Thanks in advance

질문됨 3년 전816회 조회
1개 답변
3
수락된 답변

Hello,

There is no option to kill the application automatically when initial job has not accepted any resources. Because this is not an error on yarn side but the job is waiting for resource allocation and it will run after getting resources to launch executors.

So you may consider to create a script that periodically polls application logs and kill the application if there are logs "Initial job has not accepted any resources. Check your cluster UI to ensure that workers are registered and have sufficient resources" and kill it if the progress keeps 0% for the duration you specified.

Found this external reference that might help to achieve this. You can set to run the script on the primary node configuring bootstrap action when creating the cluster. Hope this answered your query.

AWS
지원 엔지니어
답변함 3년 전
전문가
검토됨 3년 전
전문가
검토됨 3년 전
  • Thank you for the detailed information. I shared a feedback for this as feature request. 👏

로그인하지 않았습니다. 로그인해야 답변을 게시할 수 있습니다.

좋은 답변은 질문에 명확하게 답하고 건설적인 피드백을 제공하며 질문자의 전문적인 성장을 장려합니다.

관련 콘텐츠