Uploaded image for project: 'DSpace'
  1. DSpace
  2. DS-1958

Discovery OutOfMemoryError when indexing Large Bitstreams

    Details

    • Attachments:
      0
    • Comments:
      9
    • Documentation Status:
      Not Required

      Description

      Discovery SolrLogger changes in DSpace 3.0 cause entire Item Bitstreams to be copied into Strings (which then get copied into byte arrays and finally copied into response output to Solr) This is inefficient and causes Indexing to bound to JVM Memory footprint.

      SolrLogger originally passed the Bitstream.retrieve() InputStream to Solr for indexing, Solr treats InputStreams differently than Strings and holds them until needed in the Http Post request.

      The following change assures a consistent memory footprint for Solr Indexing.

      https://github.com/DSpace/DSpace/pull/515

        Attachments

          Activity

            People

            • Assignee:
              mwood Mark H. Wood
              Reporter:
              mdiggory Mark Diggory
            • Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: