-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: None
-
5
-
Storage Engines 2020-03-09
-
v4.4
This might be related to WT-5665 but it is not clear they're the same. This ticket may provide a different or easier path to reproduce.
The Evergreen recovery-stress-test has been failing consistently and I can reproduce manually running the command. The failure seen is:
[2020/02/26 23:22:01.041] Kill child [2020/02/26 23:22:01.041] Open database, run recovery and verify content [2020/02/26 23:22:01.041] Got stable_val 4980 [2020/02/26 23:22:01.041] test_timestamp_abort: FAILED: shadow search failure: WT_NOTFOUND: item not found [2020/02/26 23:22:01.041] process aborting
The command to reproduce is:
test_timestamp_abort -m -h WT_TEST.timestamp-abort -T 169 -t 12
In my manual runs I also see the opposite failure:
Kill child Open database, run recovery and verify content Got stable_val 5188 test_timestamp_abort: FAILED: shadow search success process aborting Aborted (core dumped)