CryoThrust commented on PR #12156: URL: https://github.com/apache/seatunnel/pull/12156#issuecomment-5994148402
Synced onto `dev` at `eb2eb1ddc` first — the branch was 40 commits behind, rebase was clean, and the Zeta tests still pass (`StateTransitionCleanupTest`: 7 run, 0 failures). Head is now `95504b5fa`, and a fresh run is already going (`37306887373`), so this reply is about the old head `449b51782` while the new one builds. **On the two jobs you wanted ruled out.** Both are green on that head — I pulled them from the job list of run `35681243175` (attempt 2, the Sep 29 rerun): ``` unit-test (8, ubuntu-latest) | success engine-v2-it (8, ubuntu-latest) | success unit-test (11, ubuntu-latest) | success unit-test (8/11, windows-latest) | success ``` So neither `unit-test` nor `engine-v2-it` is a failure on this head. The rerun the thread mentioned did surface them once, but they are not red in the run that is actually attached to the head. **What is red, and why it can't be this diff.** Ten jobs in that run, all connector e2e, and every failure is in container/image setup or an unrelated e2e assertion: | Job | Failing test | Error | |---|---|---| | `all-connectors-it-2 (8/11)` | `IcebergSourceIT.startUp:161` | `ContainerFetch Can't get Docker image` | | `all-connectors-it-6 (8/11)` | `DatabendIT.startUp:249`, `DatabendCDCSinkIT.startUp:221` | `ContainerFetch Can't get Docker image` | | `all-connectors-it-7 (8/11)` | `S3FileWithFilterIT.startUp:76` | `ContainerLaunch Container startup failed` | | `paimon-connector-it (8/11)` | `PaimonWithS3IT.startUp:151` | `ContainerFetch Can't get Docker image` | | `doris-connector-it (8)` | `DorisErrorIT.testDoris:85` | `expected: <true> but was: <false>` (1 of 49) | | `rocketmq-connector-it (11)` | `RocketMqIT` | 14 failures + 29 errors of 87 — the offset-visibility case already filed as #12383 | These are container-registry failures during test `startUp`, not assertion failures in product code, and the diff on this branch is limited to `PhysicalPlan` / `JobStatus` plus their tests, none of which is on the classpath of a connector e2e module. **Baseline.** I checked whether the same check is red elsewhere right now rather than only arguing from the diff: unrelated open PRs #12623, #12618 and #12616 all currently show a failing `Build` check on their own heads, touching completely different modules. That is the same pattern DanielLeens described earlier in this thread. If the fresh run on `95504b5fa` comes back with the same shape, the list above should let us classify it without another round trip; if any job fails in a way that isn't in this list, I'll dig into that one specifically. -- 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]
