Uploaded image for project: 'CDAP'
  1. CDAP
  2. CDAP-13019

Spark pipelines with Error collector re-reads the input data

    XMLWordPrintableJSON

    Details

    • Type: Bug
    • Status: Resolved
    • Priority: Critical
    • Resolution: Won't Fix
    • Affects Version/s: None
    • Fix Version/s: 4.3.3
    • Component/s: Spark
    • Labels:
      None
    • Rank:
      1|i00aav:

      Description

      If a pipeline has an error collector and a spark engine then the input data is read more than once. This is reflected in the metrics for the run. Probably happens due to caching logic in the planner/error collector plugin.

      Note: The same pipeline produces right metrics count in MapReduce or if the error collector stage is removed in spark pipeline.

        Attachments

          Issue Links

            Activity

              People

              • Assignee:
                ashau Albert Shau
                Reporter:
                sree Sreevatsan Raman
              • Votes:
                0 Vote for this issue
                Watchers:
                3 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: