ClaireLytt commented on PR #11675: URL: https://github.com/apache/seatunnel/pull/11675#issuecomment-5229861599
> Hi @ClaireLytt, thanks for flagging this — I dug into the actual CI runs for the current head ([a33a9ba](https://github.com/apache/seatunnel/commit/a33a9ba27c4c6926928685f58fb3cbd23297ff27)) on both attempts before replying, rather than going off the check summary alone.嗨，感谢您指出这一点。在回复之前，我仔细查看了当前代码分支在两次提交过程中的实际构建结果，而不仅仅是参考了检查总结而已。 > > Good news: I don't see any real CheckpointPlanTest / TaskTest failures for this PR. Specifically:好消息是：在这次提交中，我没有发现任何与 CheckpointPlanTest 或 TaskTest 相关的故障。具体来说： > > * Attempt 1 (finished 2026-08-07 12:48 UTC): in the `unit-test (8, ubuntu-latest)` job log, `org.apache.seatunnel.engine.server.checkpoint.CheckpointPlanTest` ran with `Tests run: 1, Failures: 0, Errors: 0`, and `org.apache.seatunnel.engine.server.dag.TaskTest` ran with `Tests run: 3, Failures: 0, Errors: 0`. Both are clean passes through the `AbstractSeaTunnelServerTest` Hazelcast setup, no errors in `before()` or anywhere else.第一次尝试（完成于 2026 年 8 月 7 日 12:48 UTC）：在 `unit-test (8, ubuntu-latest)` 任务日志中， `org.apache.seatunnel.engine.server.checkpoint.CheckpointPlanTest` 与 `Tests run: 1, Failures: 0, Errors: 0` 配合运行， `org.apache.seatunnel.engine.server.dag.TaskTest` 则与 `Tests run: 3, Failures: 0, Errors: 0` 配合运行。两者都在 `AbstractSeaTunnelServerTest` Hazelcast 环境中顺利通过了测试， `before()` 及其他环节均未出现任何错误。 > * The only non-passing job in attempt 1 was `unit-test (11, windows-latest)`, which failed — but that's unrelated to Checkpoint/Task tests too.在第一次尝试中，唯一没有通过测试的任务是 `unit-test (11, windows-latest)` ，该任务失败了——不过这与检查点/任务测试本身并无关联。 > * Attempt 2 (the rerun, finished 2026-08-08 08:07 UTC) is fully green on every unit-test job, including `unit-test (11, windows-latest)`, which passed on retry — consistent with that being a flaky/transient failure rather than anything real.第二次尝试（即重新运行后的结果，完成时间为 2026 年 8 月 8 日 08:07 UTC）中，所有单元测试都通过了，包括 `unit-test (11, windows-latest)` 。这说明这次失败只是偶发性的/暂时的问题，而非真正的故障。 > > The one job that is still not green in either attempt is `Run / Build website` (the docusaurus docs-site build), which hangs for well over an hour and then gets cancelled — that's what's making the overall `Build` check on the apache side show as failed/cancelled. That's unrelated to the Hazelcast/engine test setup and unrelated to the Iceberg connector code in this PR; it looks like an infra/docs-build flakiness issue on the CI side, not something introduced by your change.在这两次尝试中，唯一仍然无法成功完成的任务是 `Run / Build website` （Docusaurus 文档网站的构建）。该任务会持续运行一个多小时，之后就会取消。正是由于这个原因，Apache 端的 `Build` 检查才会显示为“失败/已取消”。这与 Hazelcast/引擎的测试设置无关，也与本次提交中的 Iceberg 连接器代码无关。看来，这是 CI 系统在处理文档构建时的问题，而非你所做的更改所导致的。 > > So from my side there's nothing in `CheckpointPlanTest`/`TaskTest`/`AbstractSeaTunnelServerTest` that needs fixing here, and it doesn't block this PR. If you're still seeing those specific test failures locally or in a different run, could you share the run URL or the actual failure stack trace? I'm happy to take another look, but I couldn't reproduce or find evidence of that failure in this PR's CI history. My prior approval stands either way.就我而言， `CheckpointPlanTest` / `TaskTest` / `AbstractSeaTunnelServerTest` 中没有任何需要修复的问题，这些问题也不会阻碍此次 PR 的提交。如果您仍然在本地或其他测试环境中遇到这些故障，请分享一下相关的测试 URL 或详细的故障堆栈跟踪信息。我很乐意再检查一遍，但在本次 PR 的 CI 测试记录中，我无法重现这些故障，也找不到相关的证据。无论如何，我之前的批准意见仍然有效。 Thank you so much for your reply! -- 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]
