Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-25686

Config server manual move chunk request won't rejoin migrations if they waitForDelete

    XMLWordPrintable

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Major - P3
    • Resolution: Fixed
    • Affects Version/s: 3.3.12
    • Fix Version/s: 3.3.12
    • Component/s: Sharding
    • Labels:
      None

      Description

      The config server's manual move chunk internal call _configsvrMoveChunk is intended to be idempotent in the absence of other metadata changing activity on the system and the continuous stepdown suite relies on this behaviour.

      However, in the case where waitForDelete is specified and the move chunk operation on the shard is already at the stage where cleanup happens, _configsvrMoveChunk may report a retried operation as successful before the cleanup has completed.

      This causes tests, which rely on the range cleanup to complete to fail when run under the continuous config stepdown suite.

        Attachments

          Activity

            People

            • Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: