[ https://issues.apache.org/jira/browse/FLINK-23180?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17409394#comment-17409394 ]
Piotr Nowojski commented on FLINK-23180: ---------------------------------------- [~wind_ljy], I think you are right, it doesn't make sense to create unused directories. If checkpointing is disabled, we should skip it. I would still suggest to keep it initialised eagerly, at the very least in order to fail fast. I'm not sure how easy/difficult would it be to implement it, as I don't remember the code. Would you like to work on this issue [~wind_ljy]? > Initialize checkpoint location lazily in DataStream Batch Jobs > -------------------------------------------------------------- > > Key: FLINK-23180 > URL: https://issues.apache.org/jira/browse/FLINK-23180 > Project: Flink > Issue Type: Improvement > Components: Runtime / Checkpointing > Affects Versions: 1.11.0 > Reporter: Jiayi Liao > Priority: Major > > Currently batch jobs will initialize checkpoint location eagerly when > {{CheckpointCoordinator}} is created, which will create lots of useless > directories on distributed filesystem. -- This message was sent by Atlassian Jira (v8.3.4#803005)