Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-14395

Crash during a checkpoint should not advance the oldest timestamp

    • Type: Icon: Bug Bug
    • Resolution: Unresolved
    • Priority: Icon: Major - P3 Major - P3
    • None
    • Affects Version/s: None
    • Component/s: Checkpoints
    • Storage Engines, Storage Engines - Persistence
    • StorEng - Defined Pipeline

      If the checkpoint crashes before it finishes—that is, updates the turtle file—the timestamps should not advance based on the checkpoint.

      In the following sequence of actions, we should always end up with oldest timestamp being 50:

      1. Set the oldest timestamp to 50.
      2. Do more work and create some checkpoints.
      3. Set the oldest timestamp to 100.
      4. Crash towards the end of the checkpoint, right before the turtle file is updated.
      5. Restart.

      At this point, I'm assuming that the oldest timestamp should be 50. However, I see that:

      • If the connection does not enable logging, the timestamp is 50.
      • If the connection enables logging, the timestamp is 100.

            Assignee:
            backlog-server-storage-engines [DO NOT USE] Backlog - Storage Engines Team
            Reporter:
            peter.macko@mongodb.com Peter Macko
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Created:
              Updated: