Avoid reopening the stable table for each operation on follower

XMLWordPrintableJSON

    • Storage Engines - Foundations
    • None
    • None

      __clayered_can_stable_advance decides when should we reopen the stable table on the layered cursor. However, for timestamped reads, it always returns true. This means we will reopen the stable cursor for almost each every operation even there is no new checkpoint being picked up.

      We should do better and do a check if there is a new checkpoint. If not, we should avoid reopen the stable table in the layered cursor.

            Assignee:
            [DO NOT USE] Backlog - Storage Engines Team
            Reporter:
            Chenhao Qu
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated: