[MDEV-27950] rpl.rpl_delayed_slave failed for test assertion 'Status should be 'Waiting until MASTER_DELAY...'' Created: 2022-02-25 Updated: 2023-04-27 |
|
| Status: | Open |
| Project: | MariaDB Server |
| Component/s: | Replication, Tests |
| Affects Version/s: | 10.6.7, 10.2, 10.3, 10.4, 10.5, 10.6, 10.7, 10.8 |
| Fix Version/s: | 10.4, 10.5, 10.6 |
| Type: | Bug | Priority: | Major |
| Reporter: | Angelique Sklavounos (Inactive) | Assignee: | Angelique Sklavounos (Inactive) |
| Resolution: | Unresolved | Votes: | 0 |
| Labels: | None | ||
| Issue Links: |
|
||||||||
| Description |
|
https://buildbot.mariadb.org/#/builders/208/builds/4366/steps/6/logs/stdio
The fail appears to usually end up as retry-pass, and is intermittent. |
| Comments |
| Comment by Angelique Sklavounos (Inactive) [ 2022-02-25 ] | ||||||||||||||||||||||
|
With the same traceback, sometimes the interpolated text is different. 10.7 507084517 https://buildbot.mariadb.org/#/builders/208/builds/4356
10.2 647e95231 https://buildbot.mariadb.org/#/builders/208/builds/4350
| ||||||||||||||||||||||
| Comment by Andrei Elkin [ 2022-03-01 ] | ||||||||||||||||||||||
|
angelique.sklavounos, let me leave this one for your detailed analysis. It looks very much as the test's own issue.
and why the test is so sure about the hardcoded 7 is the first thing to clear out. | ||||||||||||||||||||||
| Comment by Angelique Sklavounos (Inactive) [ 2022-03-04 ] | ||||||||||||||||||||||
The test adds rows to t1 that include integer values 0, 1, 2, ..., based on their order of being added. Each entry is separated by a certain sleep amount in order to test that there is a delay in the sync to the replica. The value of 7 is hardcoded here:
The macro called in line 261 will check that the replica has not synced that row yet, then sleep for a suitable amount of time, after which the row with "'stop slave and start slave: DML', 7" will now appear in the replica. In the macro, the test never seems to fail on this assertion that queries the number of rows in t1 on the replica, but rather on the replica status. It should be noted that the test does have a disclaimer of:
and the test usually seems to be retry-pass, but the initial fail does seem to be more than sporadic. |