[ https://issues.apache.org/jira/browse/FLINK-27377?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17527324#comment-17527324 ]
Guowei Ma commented on FLINK-27377: ----------------------------------- Hi [~yittg] thanks for your reporting. Afaik the commit files action happens at the committer operator instead of the JM size after the new sink api [1]. It means this would not happen if you use the new `FlinkSink`.[2] [1] [https://cwiki.apache.org/confluence/display/FLINK/FLIP-143%3A+Unified+Sink+API] [2] [https://github.com/apache/flink/blob/master/flink-connectors/flink-connector-files/src/main/java/org/apache/flink/connector/file/sink/FileSink.java] > Job manager blocks on marking vertex finished > --------------------------------------------- > > Key: FLINK-27377 > URL: https://issues.apache.org/jira/browse/FLINK-27377 > Project: Flink > Issue Type: Bug > Reporter: Yi Tang > Priority: Major > > I have a case using FileSystem as batch sink, it will commit files by > {{FinalizeOnMaster#finalizeGlobal}} while updating state from Running to > Finished. > When {{FinalizeOnMaster#finalizeGlobal}} method is blocked, the > {{JobMaster#updateTaskExecutionState}} will also be blocked. Then the job > manager can not respond any new requests until > {{JobMaster#updateTaskExecutionState}} is complete. -- This message was sent by Atlassian Jira (v8.20.7#820007)