Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-13751

Investigate why multiple checkpoints are required to fully reclaim space

    • Type: Icon: Task Task
    • Resolution: Unresolved
    • Priority: Icon: Major - P3 Major - P3
    • None
    • Affects Version/s: None
    • Component/s: Checkpoints
    • None
    • Storage Engines

      WT-13340, introduced a check in the checkpoint skip logic to proceed on files that have available space at the end of the file that can be truncated at the end of a checkpoint. This should be able to recover all the space in a file that has been completely deleted. Once all the pages are marked as obsolete and checkpoint cleanup removes all the obsolete pages, it should only take two checkpoints to move the extents from the allocated list to the available list and be truncated. However, the number of checkpoints at the end of test_checkpoint33.py required to recover all the space is at least 3 and increases as the file size increases.
      We should investigate why this is the case.

            Assignee:
            backlog-server-storage-engines [DO NOT USE] Backlog - Storage Engines Team
            Reporter:
            sean.watt@mongodb.com Sean Watt
            Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

              Created:
              Updated: