Removing a shard with in-progress transaction coordinators can leave transactions on participants permanently in prepare

XMLWordPrintableJSON

    • Type: Bug
    • Resolution: Unresolved
    • Priority: Major - P3
    • None
    • Affects Version/s: None
    • Component/s: Sharding
    • Cluster Scalability
    • ALL
    • v8.1, v8.0
    • Sharding 2020-11-02, Sharding 2020-11-16, Sharding 2020-11-30, Sharding 2020-12-14, Sharding 2020-12-28, Sharding 2021-01-11
    • None
    • 3
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      The following scenario can occur:

      1. Transaction coordinator on shard X makes a commit/abort decision after hearing from all participants
      2. Shard X is removed and shut down before transaction coordinator sends decision to participant shards
      3. Participant shards are stuck permanently with prepared transactions

        1. 0001-No-repro-when-transitioning-to-dedicated-config-serv.patch
          7 kB
        2. 0001-Repro-SERVER-50143.patch
          6 kB

            Assignee:
            [DO NOT USE] Backlog - Cluster Scalability
            Reporter:
            Matthew Saltz (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            10 Start watching this issue

              Created:
              Updated: