[Flaky test] Windows test bgw_job_stat_history flaky because custom job crashes

[Mats Kindahl]

Which test is flaky?

bgw_job_stat_history

Since when has the test been flaky?

May 13, 2024

Link to the failed test run

https://github.com/timescale/timescaledb/actions/runs/9047405772/job/24859330937

Log output

diff -u /mnt/d/a/timescaledb/timescaledb/tsl/test/expected/bgw_job_stat_history.out /mnt/d/a/timescaledb/timescaledb/build_wsl/tsl/test/results/bgw_job_stat_history.out
--- /mnt/d/a/timescaledb/timescaledb/tsl/test/expected/bgw_job_stat_history.out	2024-05-12 00:26:28.902086000 +0000
+++ /mnt/d/a/timescaledb/timescaledb/build_wsl/tsl/test/results/bgw_job_stat_history.out	2024-05-12 00:38:37.287466100 +0000
@@ -187,13 +187,13 @@
 FROM timescaledb_information.job_history
 WHERE job_id >= 1000
 ORDER BY id, job_id;
- job_id | pid | proc_schema |    proc_name     | succeeded |   config   | sqlerrcode |   err_message    
---------+-----+-------------+------------------+-----------+------------+------------+------------------
+ job_id | pid | proc_schema |    proc_name     | succeeded |   config   | sqlerrcode |             err_message             
+--------+-----+-------------+------------------+-----------+------------+------------+-------------------------------------
    1001 | t   | public      | custom_job_error | f         |            | 22012      | division by zero
    1000 | t   | public      | custom_job_ok    | t         |            |            | 
    1001 | t   | public      | custom_job_error | f         |            | 22012      | division by zero
    1000 | t   | public      | custom_job_ok    | t         | {"foo": 1} |            | 
-   1001 | t   | public      | custom_job_error | f         | {"bar": 1} | 22012      | division by zero
+   1001 | t   | public      | custom_job_error | f         |            |            | job crash detected, see server logs
 (5 rows)

Reason for flakiness

No response