DanielLeens commented on PR #12099: URL: https://github.com/apache/seatunnel/pull/12099#issuecomment-5552504705
## CI triage on `e559edbc8` (self-review note — this PR is mine, so posting as a plain comment) `Build` went red on two jobs. Since `rocketmq-connector-it` is exactly the module this PR changes, I did not assume flake — traced both from the job logs: **1. `rocketmq-connector-it (8, ubuntu-latest)` — `RocketMqIT` 43/87 red.** Error mix in the log: 311× `MQClientException CODE: 17 No topic route info in name server` + 485× `ROCKETMQ-11 Failed to get topic min and max` (the known broker route-metadata race — that is what took down the 6 source test methods across all 7 container legs), and 28× `Consume offset should be visible ... expected: not <null>` on `testSinkRocketMq` (a sink can't surface committed offsets when the broker's route metadata is broken). No shutdown/close/interrupt-related exception anywhere in the log. Decisive counter-evidence against this PR's `close()` ordering change being the cause: **`rocketmq-connector-it (11, ubuntu-latest)` passed 87/87 on this identical commit.** Same code, same broker image, same offset checks — one runner's broker fixture flaked, the other's didn't. **2. `unit-test (11, windows-latest)`.** No Maven failure at all: every surefire summary is green (including the new `RocketMqConsumerThreadCloseTest`, 2/2 passed in that very job); the log just stops mid-build at ~80 min with the test step still recorded as `in_progress` and no step conclusion — a runner-level termination, under the job's 90-minute `timeout-minutes`. The other three unit-test legs (8/ubuntu, 11/ubuntu, 8/windows) all passed on this commit. **Action:** job-level rerun of `rocketmq-connector-it (8)` submitted (run `33943742413`, now attempt 2, in progress). GitHub refuses a second job-level rerun while one is in flight, so the Windows unit-test rerun is queued behind it — I'll re-run it as soon as attempt 2 completes rather than stacking reruns. No code change needed on this PR from either failure. -- 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]
