Uploaded image for project: 'Terrier Core'
  1. Terrier Core
  2. TR-289

Docid alignment is broken for MapReduce indexing when map tasks are repeated

    Details

    • Type: Bug
    • Status: Resolved
    • Priority: Critical
    • Resolution: Fixed
    • Affects Version/s: 3.6
    • Fix Version/s: 4.0
    • Component/s: None
    • Labels:
      None

      Description

      Recent indices created by MapReduce have docid alignment problems:

      Posting id 80368334 is too big for term 0 term0 Nt=34566754 TF=587985939 @{0 0 0} TFf=5215159,4448387,578322393
      Posting id 79259555 is too big for term d term21430718 Nt=22088033 TF=65422749 @{3 0 0} TFf=679062,312433,64431254

      This is /thought/ to be unrelated to compression changes.

        Attachments

          Activity

          Hide
          craigm Craig Macdonald added a comment -

          Committed r3954.

          Show
          craigm Craig Macdonald added a comment - Committed r3954.
          Hide
          craigm Craig Macdonald added a comment -

          TR v4 patch.

          Show
          craigm Craig Macdonald added a comment - TR v4 patch.
          Hide
          craigm Craig Macdonald added a comment -

          Patch which disables per-flush compression in favour of per-map compression. This version for 3.6-ish.

          Show
          craigm Craig Macdonald added a comment - Patch which disables per-flush compression in favour of per-map compression. This version for 3.6-ish.
          Hide
          craigm Craig Macdonald added a comment -

          This issue was found to be related to per-flush compression: the MapData contains information for per-flush compression, but this information may change if a map task is repeated.

          Show
          craigm Craig Macdonald added a comment - This issue was found to be related to per-flush compression: the MapData contains information for per-flush compression, but this information may change if a map task is repeated.
          Hide
          richardm Richard McCreadie added a comment -

          I don't think that I am going to make much progress on this without the indexing MR logs. Can you attach them?

          Show
          richardm Richard McCreadie added a comment - I don't think that I am going to make much progress on this without the indexing MR logs. Can you attach them?
          Hide
          craigm Craig Macdonald added a comment -

          This program can be used to test an index, used as follows:

           bin/anyclass.sh  -Dterrier.index.path=/path/to/index -Dterrier.index.prefix=data  org.terrier.applications.PrintIndexTerm
          
          Show
          craigm Craig Macdonald added a comment - This program can be used to test an index, used as follows: bin/anyclass.sh -Dterrier.index.path=/path/to/index -Dterrier.index.prefix=data org.terrier.applications.PrintIndexTerm

            People

            • Assignee:
              craigm Craig Macdonald
              Reporter:
              craigm Craig Macdonald
            • Watchers:
              0 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: