Uploaded image for project: 'Spark'
  1. Spark
  2. SPARK-14963

YarnShuffleService should use YARN getRecoveryPath() for leveldb location

Log workAgile BoardRank to TopRank to BottomAttach filesAttach ScreenshotVotersStop watchingWatchersCreate sub-taskConvert to sub-taskLinkCloneLabelsUpdate Comment AuthorReplace String in CommentUpdate Comment VisibilityDelete Comments
    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Resolved
    • Major
    • Resolution: Fixed
    • 1.6.1
    • 2.1.0
    • Shuffle, Spark Core, YARN
    • None

    Description

      The YarnShuffleService, currently just picks a directly in the yarn local dirs to store the leveldb file. YARN added an interface in hadoop 2.5 getRecoverPath() to get the location where it should be storing this.

      We should change to use getRecoveryPath(). This does mean we will have to use reflection or similar to check for its existence though since it doesn't exist before hadoop 2.5

      Attachments

        Issue Links

        Activity

          This comment will be Viewable by All Users Viewable by All Users
          Cancel

          People

            jerryshao Saisai Shao Assign to me
            tgraves Thomas Graves
            Votes:
            0 Vote for this issue
            Watchers:
            6 Stop watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Issue deployment