DanielLeens commented on PR #12258: URL: https://github.com/apache/seatunnel/pull/12258#issuecomment-5711758423
CI follow-up for head `bd08491a32` (fork run `35097572107`). I read the failing job logs directly rather than the `Build` pointer check. **The PR's own blocker is closed.** `updated-modules-integration-test-part-6` - the job that failed deterministically on `MariaDbCDCCheckpointRestoreIT.testMysqlCdcRestoresAfterCheckpointedFullPipelineFailure` on every previous head - is green on both JDK 8 and JDK 11, as is `part-2` (the MariaDB JDBC E2E classes). **`Build` is still red from 4 jobs, none of which runs code this PR touches:** | Job | Failure | Classification | |---|---|---| | `updated-modules-integration-test-part-5` (8 and 11) | `OpengaussCDCIT.testAddFieldWithRestore:476` `ConditionTimeout` | Tracked dev-wide failure, #12344 (reported as 24/24 failing legs on unrelated PRs); fix in flight in #12346. Scheduled here only because `connector-cdc-opengauss` depends on `connector-jdbc`. | | `updated-modules-integration-test-part-3` (8) | `Could not resolve dependencies for connector-jdbc-e2e-part-2`: `org.testcontainers:database-commons:1.20.1`, `com.sap.cloud.db.jdbc:ngdbc:2.21.11` | Maven artifact resolution failure before any test ran; the JDK 11 leg of the same job passed. | | `unit-test (11, windows-latest)` | `FileCollectReaderBehaviorTest.rediscoversFileAfterInactiveCursorClosed:101` `ConditionTimeout` in `seatunnel-edge-agent-connector` | Module not touched by this PR; the JDK 8 Windows leg and both Ubuntu legs passed. | **Path to a green `Build`:** the `part-3` and Windows unit-test legs should clear on a rerun of the failed jobs (`gh run rerun 35097572107 --failed` - @merlau, this has to come from your side, the run lives on your fork and maintainers cannot rerun it). The two `part-5` legs will keep failing on any rerun until #12346 (or another fix for #12344) lands on `dev` and this branch picks it up, so I would not spend CI on them before that. On the one non-blocking review note (a narrow unit test for the offset-counter handoff): `loadStartingOffsetState` is private and needs a live connection/binlog client, so covering it in isolation means reshaping production code for testability. I'd rather not do that on an approved head and restart a 4h CI run that is currently exposed to #12344; happy to do it as a follow-up. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
