Currently, startup_recovery_for_restore_restarts.js only waits, after inserting into a sentinel collection, for a log message that indicates an unstable checkpoint has completed. However, we must also wait for the log message that indicates the start of checkpoint, to make sure the insert is covered.
Wait for unstable checkpoint to begin and complete after inserting sentinel collection in startup_recovery_for_restore_restarts.js
- Votes:
-
0 Vote for this issue
- Watchers:
-
2 Start watching this issue
- Created:
- Updated:
- Resolved: