Details
-
Bug
-
Status: Closed
-
Major - P3
-
Resolution: Won't Fix
-
None
-
None
-
None
Description
ISSUE SUMMARY
MongoDB 4.0+ may lose data during unclean shutdowns on macOS 10.12+. See WT-4018 for additional information.
USER IMPACT
Potential data corruption if the server is not shutdown properly.
WORKAROUNDS
We have raised an issue with the macOS team at Apple and they have properly addressed this issue as part of macOS 10.14.1.
AFFECTED VERSIONS
MongoDB production releases 4.0+ are affected by this issue.
FIX VERSION
None
RESOLUTION DETAILS
This issue has been resolved as part of macOS 10.14.1, if you are running 10.14.1 or later this issue is now fixed.
A run of the smoke tests failed for random-abort on OS X 10.12. This is another very rare instance first reported in WT-3540. The symptoms are the same, the last record from some thread is missing.
Parent: Compatibility false in-mem log false
|
Parent: Create 5 threads; sleep 10 seconds
|
Create 5 writer threads
|
Spawned 5 writer threads
|
Parent: Compatibility false in-mem log false
|
Parent: Create 5 threads; sleep 10 seconds
|
Kill child
|
Open database, run recovery and verify content
|
records-3: no record with key 3000011951
|
1 record(s) absent from 62012
|
Attachments
Issue Links
- is duplicated by
-
WT-4018 Checkpoint integrity test indicates MacOS 10.13 fsync bug
-
- Closed
-
-
WT-4398 Lost record detected during recovery in random_abort test
-
- Closed
-
- related to
-
DOCS-11753 macOS 10.13 (APFS) may not be crash safe
-
- Closed
-