zhangshenghang commented on PR #11500: URL: https://github.com/apache/seatunnel/pull/11500#issuecomment-6251285801
Thanks @DanielLeens — here are the failing job links for head `b0101f3fc3` (fork run [39013187961](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961)), plus why I did not rerun them. Two distinct causes, 5 lanes × JDK 11/17, and as far as I can tell neither comes from this PR's diff: **1) Empty-stream `PreconditionViolationException` — 4 lanes** - `all-connectors-it-2` ([JDK17](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961/job/117593855105) / [JDK11](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961/job/117593855063)): `IcebergSourceIT` - `all-connectors-it-6` ([JDK17](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961/job/117593855269) / [JDK11](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961/job/117593855263)): `FlinkMetricsIT` - `mysql-cdc-connector-it` ([JDK17](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961/job/117593855244) / [JDK11](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961/job/117593855251)): `MysqlCDCWithFlinkSchemaChangeIT` - `transform-v2-it-part-2` ([JDK17](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961/job/117593855130) / [JDK11](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961/job/117593855122)): `TestSparkDateTimeTransformIT` All four jobs ran with `RUN_ENGINES: ["zeta"]` (the `run-engines` step from #12812, merged to dev on Oct 16; this branch picked it up in the dev sync that produced `b0101f3f`). These tests only build Flink/Spark containers, so under a zeta-only engine list the `@TestContainers` factory returns zero containers and JUnit rejects the `@TestTemplate` before anything actually runs. The inputs are identical on a rerun, so it would just reproduce — that's why I skipped it. Cross-checks that point away from this change: the same head passes every lane that exercises the engine change itself (`engine-k8s-it` both JDKs, `engine-v2-it` both JDKs, code style); and an unrelated PR (#12377, cancellation-state handling, also ran with `RUN_ENGINES: ["zeta"]`) fails exactly the same 5 lanes with identical signatures in its Oct 18 run [39104328939](https://github.com/zhangshenghang/seatunnel/actions/runs/39104328939) (e.g. [transform-v2-it-part-2](https://github.com/zhangshenghang/seatunnel/actions/runs/39104328939/job/117643680293), [hbase](https://github.com/zhangshenghang/seatunnel/actions/runs/39104328939/job/117643680248)). My Oct 14 run on `f5fd0fdd` (the head right before that dev sync) had all five of these lanes green — its failures were `engine-k8s-it` / `engine-v2-it` (JDK11), addressed since. **2) `hbase-connector-it` ([JDK17](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961/job/117593855061) / [JDK11](https://github.com/zhangshenghang/seatunnel/actions/runs/39013187961/job/117593855017)): `HbaseIT`, 22/22 errors** — the `seatunnelhub/hbase-standalone:2.4.9` container never finished master init (`KeeperException$NoNodeException: NoNode for /hbase/hbaseid`, then `PleaseHoldException: Master is initializing` until the retries ran out). The same failure shows on #12377's run above, so it looks like the container/runner side rather than this change. I'm not sure yet what the cleanest upstream fix is (skip engine-mismatched `@TestTemplate` tests when their engine is not in `RUN_ENGINES`, or keep the broad lanes on the full engine list) — happy to help once there's agreement. Until then I'll watch dev and re-run/re-push this branch as soon as the filtering issue is addressed upstream. Also re-checked F6: the PR description no longer mentions the JDBC converter change. -- 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]
