Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-23010

Committed Snapshot Meaning Incorrect

    XMLWordPrintable

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Critical - P2
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 3.3.4
    • Component/s: Replication
    • Labels:
      None
    • Backwards Compatibility:
      Fully Compatible
    • Operating System:
      ALL
    • Backport Requested:
      v3.2
    • Sprint:
      Repl 11 (03/11/16), Repl 12 (04/01/16)
    • Linked BF Score:
      0

      Description

      Currently with the changes for the SERVER-22276 (and SERVER-22495) the committed snapshot is not correctly being updated based on the replica set config and journal write concern flag.

      Rules to apply to correct the problem:

      The meaning of the snapshot depends on the replica set config (“writeConcernMajorityJournalDefault”) which indicates if it is tied to the durable or applied optime.

      Notable Exceptions (when both aren't true or false)

      1. Applied Snapshot OpTime, with J:True
        • Wait for snapshot for writeOpTime (Applied)
        • Wait for replication of durableOpTime >= writeOpTime (new code)
      2. Durable Snapshot OpTime, with J:False
        • Do not wait for snapshot, and you cannot read your own write
        • Wait for replication of majority optime (applied = normal)

        Attachments

          Issue Links

            Activity

              People

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

                Dates

                • Created:
                  Updated:
                  Resolved: