Uploaded image for project: 'Spark'
  1. Spark
  2. SPARK-1993

Let users skip checking output directory

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Resolved
    • Major
    • Resolution: Duplicate
    • None
    • None
    • Spark Core
    • None

    Description

      SPARK-1100 made our behavior consistent with Hadoop (we won't let users replace an existing output directory). We should add a flag to let advanced users skip this. Something like "spark.hadoop.skipOutputChecks". We'd just need to check for this when we check the specs. See this patch:

      https://github.com/apache/spark/commit/3a8b698e961ac05d9d53e2bbf0c2844fcb1010d1

      Attachments

        Issue Links

          is related to

          Improvement - An improvement or enhancement to an existing feature or task. SPARK-1677 Allow users to avoid Hadoop output checks if desired

          • Major - Major loss of function.
          • Resolved

          Activity

            People

              Unassigned Unassigned
              pwendell Patrick Wendell
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: