Improve diagnosability for invariant failure in WiredTigerRecordStore::waitForAllEarlierOplogWritesToBeVisible()

XMLWordPrintableJSON

    • Type: Improvement
    • Resolution: Fixed
    • Priority: Major - P3
    • 4.9.0
    • Affects Version/s: None
    • Component/s: Storage
    • None
    • Fully Compatible
    • Execution Team 2021-02-22
    • 115
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      Having trouble diagnosing a test failure.

      waitForAllEarlierOplogWritesToBeVisible() has an invariant check that there is no open txn on the storage engine when it begins to wait (so it can see the latest oplog entries).

      Changes:

      • Add the RecoveryUnit's _state via toString(getState())
      • Add the readConcern setting – UPDATE: doesn't seem worthwhile to link read concern files into the WT integration layer b/c no AutoGetColl* is taken so no read source selection occurs w/ possible snapshot opening
      • Is there any way to add whether it's part of a multi-document transaction??

            Assignee:
            Dianna Hohensee (Inactive)
            Reporter:
            Dianna Hohensee (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Created:
              Updated:
              Resolved: