SEPURI-SAI-KRISHNA commented on PR #12650: URL: https://github.com/apache/seatunnel/pull/12650#issuecomment-6310607896
CI update on `b82484a95`, and a correction to how I read the last two runs. The remaining red is **not** an infrastructure cancellation. It is this job hitting its own wall clock: - attempt 5: `updated-modules-integration-test-part-6` on JDK 11 and 17, both stopped at 210.5 minutes - attempt 6: the same job on JDK 17 only, stopped at 210.8 minutes `backend.yml` **on this branch** sets `timeout-minutes: 210` for `updated-modules-integration-test-part-6`, and GitHub records a job that exceeds its limit as `cancelled` rather than `failed`, which is what made it look like an external cancel. I checked the 300 minute value first and it misled me: that is the limit on current `dev`, not here. #11929 (`4a6f0fe82`, merged 2026-10-17) raised every integration shard to 300, and this branch predates it by three days. So re-running cannot clear it. The shard needs more than 210 minutes for the module set this change selects, and only a sync picks up the larger budget. Everything else is now green. Attempt 6 is **61 success, 23 skipped, 1 cancelled, 0 failures**, and in particular the `updated-modules-integration-test-part-7` Databend CDC failure from the earlier run passed this time, so there are no failing tests left on this PR. @DanielLeens your suggestion to re-run only the failed jobs was the right call on the information we both had, and it did clear part-7. The part-6 shard was never going to pass that way. Next step on my side: sync the latest `dev` and push, which brings in #11929 and gives part-6 the 300 minute budget. I checked the merge first. The only file this PR shares with `dev` since the merge base is `ArrowToSeatunnelRowReader.java`, and #11929 changed the constructor and `close()` while this PR changes `convertSeatunnelRow`, so the two do not overlap and the fix is unaffected. One thing to expect after the sync: the four `unit-test` legs will go red on `LLMBooleanOutputTest.readsLegacyFactoryWrapperWithValidationDisabled`, an `InvalidClassException` on the `AbstractMultiCatalogTransform` serialVersionUID that arrived with #12194 and is tracked in #12929. It is unrelated to this change and currently reddens everything based on current `dev`. -- 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]
