1996fanrui commented on PR #26619: URL: https://github.com/apache/flink/pull/26619#issuecomment-2929293805
> I'd approve the PR as is and add an IT as a separate effort. Here are some ideas: > * We could use a sink with high backpressure on writer. > * Trigger UC. > * After successful UC, fail job indefinitively (restart policy = none). > * Inspect UC that is being retained by default. Verify that channel state exists. Thanks @AHeise for the comment. My initial thought was to verify that the unaligned checkpoints worked by verifying the checkpoint duration, which really needed a benchmark. But I think verifying that the channel state existed is also a good choice. Hey @pnowojski , may I know what's your opinion? Also, I am open to whether IT is done in this PR. Anyway, I can start preparing for IT after we are aligned on the expectations for IT. -- 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: issues-unsubscr...@flink.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
