Solr
  1. Solr
  2. SOLR-3065

Let overseer process cluster state changes asynchronously

    Details

    • Type: Improvement Improvement
    • Status: Closed
    • Priority: Minor Minor
    • Resolution: Fixed
    • Affects Version/s: 4.0-ALPHA
    • Fix Version/s: 4.0-ALPHA
    • Component/s: SolrCloud
    • Labels:
      None

      Description

      Currently the overseer updates clusterstate.json on almost every change - one change at a time. This is not efficient when there are a lot of changes happening in short period of time (for example when a number of hosts are started at once).

      It would be better if changes were published on timely manner instead.

        Activity

        Hide
        Sami Siren added a comment -

        we might want to at least log warnings on the '// XXX stop processing, exit' spots, at least for debugging purposes - I have not for the moment

        sure. that's a good idea.

        Show
        Sami Siren added a comment - we might want to at least log warnings on the '// XXX stop processing, exit' spots, at least for debugging purposes - I have not for the moment sure. that's a good idea.
        Hide
        Sami Siren added a comment -

        we might want to at least log warnings on the '// XXX stop processing, exit' spots, at least for debugging purposes - I have not for the moment

        sure. that's a good idea.

        Show
        Sami Siren added a comment - we might want to at least log warnings on the '// XXX stop processing, exit' spots, at least for debugging purposes - I have not for the moment sure. that's a good idea.
        Hide
        Mark Miller added a comment -

        Thanks Sami - committing this now. Only comment is that we might want to at least log warnings on the '// XXX stop processing, exit' spots, at least for debugging purposes - I have not for the moment, but if you agree I will add them.

        Show
        Mark Miller added a comment - Thanks Sami - committing this now. Only comment is that we might want to at least log warnings on the '// XXX stop processing, exit' spots, at least for debugging purposes - I have not for the moment, but if you agree I will add them.
        Hide
        Sami Siren added a comment -

        -process cluster state updates asynchronously (now once per 500ms)
        -make sure overseer is still leader before updating the state

        Show
        Sami Siren added a comment - -process cluster state updates asynchronously (now once per 500ms) -make sure overseer is still leader before updating the state

          People

          • Assignee:
            Mark Miller
            Reporter:
            Sami Siren
          • Votes:
            0 Vote for this issue
            Watchers:
            0 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Development