Uploaded image for project: 'Couchbase Server'
  1. Couchbase Server
  2. MB-45958

FTS: concurrent index partitions movement during rebalance seems to be slow.

    XMLWordPrintable

Details

    Description

      Build: 7.0.0-5017

      Scenario is the following:

      Having cluster {kv} {fts}

      I'm doing the following steps:

      • set¬†maxConcurrentPartitionMovesPerNode = 1
      • create bucket `default`
      • load 100,00 docs into it, docs are attached
      • creating default fts index on `default`¬†
      • set number of partitions to be 6 for fts index
      • waiting for indexing finish
      • start rebalance-in {kv,fts} node
      • waiting for rebalance-in end, calculating rebalance time.
      • rebalance {kv, fts} node out
      • waiting for fts indexing finish
      • set¬†maxConcurrentPartitionMovesPerNode = 2
      • rebalance in {kv, fts} node once again
      • comparing 1'st rebalance time and 2'nd.¬†

      It is not 100% reproducible, but sometimes, 2'nd time is longer than 1'st one.

      Attachments

        For Gerrit Dashboard: MB-45958
        # Subject Branch Project Status CR V

        Activity

          People

            Sreekanth Sivasankaran Sreekanth Sivasankaran (Inactive)
            evgeny.makarenko Evgeny Makarenko (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Gerrit Reviews

                There are no open Gerrit changes

                PagerDuty