DanielLeens commented on PR #12902: URL: https://github.com/apache/seatunnel/pull/12902#issuecomment-6673217488
Thanks @hutiefang76 for rerunning only that lane. [Attempt 3](https://github.com/hutiefang76/seatunnel/actions/runs/41828185781/attempts/3) finished, and `all-connectors-it-5 (17, ubuntu-latest)` failed again at the same spot: `TiDBCDCIT.startUp:94` -> `TiDBTestBase.initializeTidbTable:213`, `CommunicationsException: Communications link failure`, under a second after `Containers are started.` This still doesn't point at your change. The PR only touches `CompileDescriptor`, `ProtobufConverterTest` and the two protobuf docs, and nothing under `connector-cdc-tidb-e2e`, `seatunnel-e2e-common` or the workflows differs between your merge base and current dev. So syncing dev won't change this lane. Same-day dev run 41987726874 passed `TiDBCDCIT` on `all-connectors-it-5 (17)`, on the same runner OS and Docker version. The logs do show a readiness race. The `pingcap/tidb:v6.1.0` container in `TiDBTestBase.java:102-115` has no `waitingFor(...)`, so Testcontainers counts it as started once the mapped port is listening. `initializeTidbTable` then opens its first JDBC connection right away, with no retry (`TiDBTestBase.java:188-192`). In the passing dev run, TiDB took about 31 s to report started. In your attempt 3 it reported started after about 21 s, and the first connection got no packets back. Three failures in a row on this head make it look less like a one-off and more like this race is easy to lose. I wouldn't put a fix into this PR. The right fix is a small, separate E2E change: wait until TiDB actually accepts a MySQL connection, either with a bounded Awaitility retry around the first `getJdbcConnection()` or a SQL-level wait strategy on the TiDB container. If you'd like to open that PR, I'm happy to review it. Otherwise one more rerun of just this job is the cheapest way to get a green run on this head. The source review conclusion is unchanged: no source blockers, and green CI on this head is the only remaining gate. -- 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]
