Problem/Motivation
Random test failure - Table 'sequences' already exists.
https://www.drupal.org/pift-ci-job/1704722
Testing Drupal\FunctionalTests\Installer\InstallerExistingConfigDirectoryTest
E 1 / 1 (100%)
Time: 4.66 seconds, Memory: 4.00 MB
There was 1 error:
1) Drupal\FunctionalTests\Installer\InstallerExistingConfigDirectoryTest::testInstaller
Exception: Drupal\Core\Database\SchemaObjectExistsException: Table 'sequences' already exists.
Drupal\Core\Database\Schema->createTable()() (Line: 614)
https://www.drupal.org/pift-ci-job/1705699
Testing Drupal\Tests\jsonapi\Functional\EntityViewModeTest
...SE... 8 / 8 (100%)
Time: 1.29 minutes, Memory: 6.00 MB
There was 1 error:
1) Drupal\Tests\jsonapi\Functional\EntityViewModeTest::testPostIndividual
Drupal\Core\Database\SchemaObjectExistsException: Table 'sequences' already exists.
Proposed resolution
| Comment | File | Size | Author |
|---|---|---|---|
| #12 | 3143812-12-1000x-fail.patch | 1.75 KB | quietone |
| #11 | 3143812-11-1000x-fail.patch | 1.72 KB | quietone |
| #10 | 3143812-10-1000x-fail.patch | 525 bytes | quietone |
| #9 | 3143812-9-1000x-fail.patch | 1.74 KB | quietone |
Comments
Comment #2
krzysztof domańskiComment #3
krzysztof domańskiComment #4
krzysztof domańskihttps://www.drupal.org/pift-ci-job/1841152
Comment #9
quietone commentedrandom test failures are bugs, changing status.
The reports for this failure are all on 9.1
9.1.x-dev test with PHP 7.3.x-dev & MySQL 5.7 - HEAD
3143618-10.patch test with PHP 7.4 & MySQL 5.7 - #3143618: Replace space (U+0020) with No-Break Space (U+00A0) to minimize false positives
2716019-95.patch test with PHP 7.3 & MySQL 8 - #2716019: View titles in breadcrumb and metatag title don't get properly translated
There have been no other mentions of this failing on #2829040: [meta] Known intermittent, random, and environment-specific test failures.
Making a patch to run this test 1,000 times and see it it will fail.
Comment #10
quietone commentedTry again
Comment #11
quietone commentedMaybe third time lucky.
Comment #12
quietone commentedDarn. I lost the
repeat: 1000Comment #13
quietone commentedI forgot to change the status
Comment #14
quietone commentedDiscussed in #bugsmash with borisson_ and longwave. longwave thought that this test fails when run concurrently, so 1000 repeats (as done in the patch) does not prove anything. They also thought that there is supposed to be a lock that prevents the problem. borisson_ pointed out that if this issue is before the lock was added, then it can probably be closed.
When was the lock added?
Comment #15
quietone commentedYes, this still occurs,
https://www.drupal.org/pift-ci-job/2423331
Comment #18
quietone commentedUpdating the title per Special issue titles.
Comment #20
dcam commentedThis issue has to be outdated for the main branch (Drupal 12) because the sequences table has been removed. See #3335756: Drop sequences table in Drupal 12. But I'm not going to close it right now since Drupal 11 will be supported until Drupal 13 is released. I don't know if it still occurs on that branch or not.
Comment #21
mstrelan commentedIt might still be an issue because sequences used to be the first table created when installing the site. We might just see another table instead, I think keyvalue.
Comment #22
smustgrave commentedAny objections to closing this one? Haven't see any errors and we aren't skipping any tests in this file. Can always re-open.
Comment #23
mstrelan commentedFine with me. I think outdated makes sense as the table no longer exists.