Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-26198

Balancer can return Status::OK for a manual moveChunk while the migration is still waiting on waitForDelete

    XMLWordPrintable

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Major - P3
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 3.3.14
    • Component/s: Sharding
    • Labels:
      None
    • Backwards Compatibility:
      Fully Compatible
    • Operating System:
      ALL
    • Sprint:
      Sharding 2016-10-10
    • Linked BF Score:
      0

      Description

      Mongos sends moveChunk command, balancer on config server schedules it, then config server steps down, interrupting the balancer at the point where the migration has been committed but waitForDelete has not finished on the shard. MigrationManager::executeManualMigration does a check on failure to see if the moveChunk was actually successful, and if so, sets the Status to OK.

      If this is the case, we should not set Status to OK if the error is identifiably due to config interrupt/stepdown.

        Attachments

          Activity

            People

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

              Dates

              • Created:
                Updated:
                Resolved: