Uploaded image for project: 'Flink'
  1. Flink
  2. FLINK-26652

Failing to cleanup a job should not fail the Flink Cluster in Session Mode

    XMLWordPrintableJSON

Details

    Description

      We introduced the option to disable the retryable cleanup in FLINK-26331. This should make Flink fall back to the 1.14- functionality with just printing a warning in session mode.

      Instead, a RetryException is thrown which causes Flink to fail fatally. For Job and Application Mode failing fatally is ok because it doesn't affect other builds. But for session mode, we want to print a warning, instead.

      Attachments

        Issue Links

          is caused by

          Sub-task - The sub-task of the issue FLINK-26331 Make max retries configurable

          • Major - Major loss of function.
          • Resolved
          Testing discovered

          Bug - A problem which impairs or prevents the functions of the product. FLINK-26690 DefaultJobGraphStore.globalCleanup does not trigger the cleanup when retrying the cleanup after a failover

          • Blocker - Blocks development and/or testing work, production could not run
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. FLINK-26680 SessionDispatcherLeaderProcess#onAddJobGraph crashes if the job is marked as deleted

          • Blocker - Blocks development and/or testing work, production could not run
          • Closed
          links to

          Web Link GitHub Pull Request #19102

          Web Link GitHub Pull Request #19117

          Activity

            People

              mapohl Matthias Pohl
              mapohl Matthias Pohl
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: