Require PIT restore checkpoints to cover prior writes in disagg_pit_restore.py

XMLWordPrintableJSON

    • Type: Task
    • Resolution: Gone away
    • Priority: Major - P3
    • None
    • Affects Version/s: None
    • Component/s: None
    • None
    • Query Optimization
    • 200
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      The DisaggPeriodicPITRestore hook accepted any checkpoint that was new to its own bookkeeping map. GetLastCheckpoint can still return a checkpoint predating the hook's fsync (for example when the forced checkpoint has not reached CRS yet), so the hook paired the current data capture with an older PIT and the post-restore validation can fail with recently created collections missing.

            Assignee:
            Henri Nikku
            Reporter:
            Henri Nikku
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: