[ https://issues.apache.org/jira/browse/FLINK-25634?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17475096#comment-17475096 ]
zlzhang0122 commented on FLINK-25634: ------------------------------------- I have encountered similar case and found that the config yarn.application-attempts is also affect by the config yarn.application-attempt-failures-validity-interval, the default value of the second config is 10 seconds, you can go ahead and check the config.Maybe the same case. > flink-read-onyarn-configuration > ------------------------------- > > Key: FLINK-25634 > URL: https://issues.apache.org/jira/browse/FLINK-25634 > Project: Flink > Issue Type: Bug > Affects Versions: 1.11.2 > Reporter: 宇宙先生 > Priority: Major > Attachments: image-2022-01-13-10-02-57-803.png, > image-2022-01-13-10-03-28-230.png, image-2022-01-13-10-07-02-908.png, > image-2022-01-13-10-09-36-890.png, image-2022-01-13-10-10-44-945.png, > image-2022-01-13-10-14-18-918.png > > > in flink-src.code : > !image-2022-01-13-10-14-18-918.png! > Set the number of retries for failed YARN ApplicationMasters/JobManagers in > high > availability mode. This value is usually limited by YARN. > By default, it's 1 in the standalone case and 2 in the high availability case > > in my cluster,the number of retries for failed YARN ApplicationMasters is 2 > yarn's configuration also like this > !image-2022-01-13-10-07-02-908.png! > But it keeps restarting when my taskfails, > !image-2022-01-13-10-10-44-945.png! > I would like to know the reason why the configuration is not taking effect. > sincere thanks! -- This message was sent by Atlassian Jira (v8.20.1#820001)