Hello, sorry to bother but I want to know what’s the progress of this FLIP, we have facing the similar problem in Flink and FlinkSQL batch mode because of the resource shortage momentary and I think this may help us. Thanks!
On 2022/05/30 16:11:24 "Afonso, Nuno" wrote: > Hello everyone, > > I would like to start a discussion about FLIP-236: Asynchronous Job > Submission [1]. > > During the processing of a job submission request, Flink runs the “main” > method in a previously submitted JAR. > This method may contain operations that take arbitrarily long to complete or > genuine issues that will never generate a job. > In those cases, clients will not be able to cancel any ongoing job > submissions and will block (until a timeout is reached) before realizing that > no progress is being made. > > In [1], we propose a new set of API calls to support an asynchronous job > submission: creation, listing, completion tracking and deletion of job > submissions. > It uncouples the passive wait as well as extends control over job submissions. > > Looking forward to hearing your feedback, > Nuno > > [1] > https://cwiki.apache.org/confluence/display/FLINK/FLIP-236%3A+Asynchronous+Job+Submission > > <https://cwiki.apache.org/confluence/display/FLINK/FLIP-236%3A+Asynchronous+Job+Submission> >
