Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-4020

test_timestamp_abort detected records after stable

    XMLWordPrintable

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Major - P3
    • Resolution: Duplicate
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      While running make check on my branch for WT-3959, I encountered a failure:

      CONFIG: test_timestamp_abort -C -h WT_TEST.timestamp-abort -T 5 -t 10
      Create checkpoint thread
      Create timestamp thread
      Create 5 writer threads
      Thread 0 starts at 0
      Thread 1 starts at 1000000000
      Thread 4 starts at 4000000000
      Thread 2 starts at 2000000000
      Thread 3 starts at 3000000000
      Checkpoint 0 complete at stable 26108.
      Checkpoint 1 complete at stable 50232.
      Checkpoint 2 complete at stable 88874.
      Parent: compatibility: true, in-mem log sync: false, timestamp in use: true
      Parent: Create 5 threads; sleep 10 seconds
      CONFIG: test_timestamp_abort -C -h WT_TEST.timestamp-abort -T 5 -t 10
      Kill child
      Open database, run recovery and verify content
      Got stable_val 88874
      records-0: COLLECTION record with key 17716 record ts 88878 > stable ts 88874
      records-0: COLLECTION record with key 17717 record ts 88883 > stable ts 88874
      records-0: COLLECTION record with key 17718 record ts 88942 > stable ts 88874
      records-0: COLLECTION record with key 17719 record ts 88947 > stable ts 88874
      records-0: COLLECTION record with key 17720 record ts 88953 > stable ts 88874
      records-0: COLLECTION record with key 17721 record ts 88957 > stable ts 88874
      ...
      [LOTS of records listed here]
      

      It is not reproducible (so far) on either my Mac or my AWS box. I have the original, pre-recovery copy of the database directory.

      While this was on my branch for WT-3959, it is not entirely clear to me that it is directly related to the changes on that branch. It also feels like it could be a possible variation of WT-3998 or some of the other failures we've been seeing lately.

        Attachments

          Activity

            People

            • Assignee:
              backlog-server-execution Backlog - Execution Team
              Reporter:
              sue.loverso Susan LoVerso
            • Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: