Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-51610

Support retrying write statements within a transaction

    • Type: Icon: Task Task
    • Resolution: Gone away
    • Priority: Icon: Major - P3 Major - P3
    • None
    • Affects Version/s: None
    • Component/s: Replication, Sharding
    • Labels:
      None
    • Sharding NYC
    • Sharding 2020-11-02, Sharding 2020-11-16

      As part of this ticket (or spun off tickets):

      • Extend TransactionParticipant to track optional statement ids for transaction commands and return the response a write would have returned originally if a previously executed statement id is received
        • Optional because drivers will not include them, only a router
        • Should support all retryable writes commands, i.e. insert, update, delete, and findAndModify
      • Extend TransactionRouter to send statement ids to shards
        • This is only needed for internal transactions used to execute a retryable write, so this does not need to be default behavior
      • Extend chunk migration session history migration to migrate history for committed transactions to enable reconstructing statement responses for retries across migrations

            Assignee:
            backlog-server-sharding-nyc [DO NOT USE] Backlog - Sharding NYC
            Reporter:
            jack.mulrow@mongodb.com Jack Mulrow
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Created:
              Updated:
              Resolved: