-
Type:
Improvement
-
Resolution: Won't Do
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
None
-
None
I observed this while waiting for PR testing to finish. unit_test_bucket06 on windows64 took >30 minutes. The following logs suggest one particular test is the main cause:
[2023/01/09 23:13:53.890] [pid:5964]: None ... ok [2023/01/09 23:13:53.890] [pid:5964]: None ... ok [2023/01/09 23:13:53.890] [pid:5964]: None ... ok [2023/01/09 23:14:16.264] [pid:5964]: None ... ok [2023/01/09 23:14:16.264] [pid:5964]: None ... ok [2023/01/09 23:14:16.264] [pid:5964]: None ... ok [2023/01/09 23:14:17.896] [pid:5964]: None ... ok [2023/01/09 23:14:17.896] [pid:5964]: None ... ok [2023/01/09 23:14:17.896] [pid:5964]: None ... ok [2023/01/09 23:14:20.423] [pid:5964]: None ... ok [2023/01/09 23:14:20.423] [pid:5964]: None ... ok [2023/01/09 23:14:20.635] [pid:5964]: Test salvage in a 'wt' process on a table that is purposely damaged. (column.fail-point) ... [pid:5964]: after salvaging, file has 274/1000 entries [2023/01/09 23:14:20.659] ok [2023/01/09 23:14:20.659] [pid:5964]: None ... skipped "Creating many tables shouldn't fill the cache (enable with --long)" [2023/01/09 23:14:28.853] [pid:5964]: None ... ok [2023/01/09 23:14:28.853] [pid:5964]: None ... ok [2023/01/09 23:14:28.853] [pid:5964]: None ... ok [2023/01/09 23:14:30.507] [pid:5964]: None ... ok [2023/01/09 23:14:30.507] [pid:5964]: None ... ok [2023/01/09 23:14:30.507] [pid:5964]: None ... skipped 'requires a non-diagnostic build' [2023/01/09 23:24:47.727] [pid:5964]: None ... ok [2023/01/09 23:24:47.727] [pid:5964]: None ... ok [2023/01/09 23:24:47.727] [pid:5964]: None ... ok [2023/01/09 23:24:47.880] [pid:5964]: Test create in a 'wt' process (rS) ... ok [2023/01/09 23:24:48.809] [pid:5964]: Spot check various combinations of configuration options. (table-lsm) ... ok [2023/01/09 23:25:50.694] [pid:5964]: None ... ok [2023/01/09 23:25:50.694] [pid:5964]: None ... ok [2023/01/09 23:25:50.694] [pid:5964]: None ... ok [2023/01/09 23:27:22.040] [pid:5964]: None ... ok [2023/01/09 23:27:22.040] [pid:5964]: None ... ok [2023/01/09 23:27:22.040] [pid:5964]: None ... ok [2023/01/09 23:27:40.765] [pid:5964]: None ... ok [2023/01/09 23:27:40.876] [pid:5964]: Test iterating a table that contains one preexisting entry (fix) ... ok [2023/01/09 23:27:41.101] [pid:5964]: None ... ok [2023/01/09 23:27:41.101] [pid:5964]: None ... ok [2023/01/09 23:27:41.101] [pid:5964]: None ... ok [2023/01/09 23:27:41.805] [pid:5964]: None ... ok [2023/01/09 23:27:41.805] [pid:5964]: None ... ok [2023/01/09 23:27:41.805] [pid:5964]: None ... ok [2023/01/09 23:27:41.806] [pid:5964]: Test salvage in a 'wt' process, using an empty table (column.no-fail-point) ... skipped 'extension "encryptors/rotn" not built' [2023/01/09 23:27:45.220] [pid:5964]: None ... ok [2023/01/09 23:27:45.220] [pid:5964]: None ... ok [2023/01/09 23:27:45.220] [pid:5964]: None ... skipped 'extension "encryptors/sodium" not built' [2023/01/09 23:27:46.425] [pid:5964]: None ... ok [2023/01/09 23:27:46.425] [pid:5964]: None ... ok [2023/01/09 23:27:46.425] [pid:5964]: None ... ok
We can see a 10 minute gap in the timestamps and several minute gaps. We should investigate why it is a lot slower than the ubuntu2004 equivalent.