Re: Flink 1.15 Interval Join error after Deduplicate

The view A try to do de-duplication using event time, which will still produce update rows. if you using proc time to do de-duplication.Then the view A should only produce append only rows. Best regards, Yuxia > 2022年10月15日 上午9:50，liebin...@whu.edu.cn 写道： > > I had a problem with Interva

Re: Flink 1.15 Interval Join error after Deduplicate

Why keep the first row still produce update/delete rows? There was a similar question before. Jark said that keep the first line should produce an append-only stream and interval join can consume from it. The link to this question is: https://lists.apache.org/thread/z5w5n7o9g87dtb7p7q7p7vrvowqh

Re: Flink 1.15 Interval Join error after Deduplicate

> view A should only produce insert rows No, the view A will still produce update/delete rows. Best regards, Yuxia > 2022年10月15日 上午9:50，liebin...@whu.edu.cn 写道： > > view A should only produce insert rows

Re: SQL Changes between 1.14 and 1.15?

Thanks for raising it. It seems a bug that introduced by this pr [1]. I have created [FLINK-29651] to trace it. [1] https://github.com/apache/flink/pull/19001 [2] https://issues.apache.org/jira/browse/FLINK-26520

Re: Job Manager getting restarted while restarting task manager

Are you able to replay this scenario? Did you accidently send killing signal to the job mananger process? On Thu, 13 Oct 2022 at 4:02 PM, Puneet Duggal wrote: > Hi, > > We use session deployment mode with HA setup. Currently we have 3 job > managers and 3 task managers running on flink version 1