[GitHub] [flink] XComp commented on pull request #23101: [FLINK-32667][dispatcher] Do not save job without restart strategy to job writer and store

["XComp (via GitHub)" <gi...@apache.org>]

XComp commented on PR #23101:
URL: https://github.com/apache/flink/pull/23101#issuecomment-1658366241

   I'm wondering whether you're doing too much here: Aren't you just interested in not adding the `JobGraph` to the `JobGraphStore`? What about coming up with a custom implementation of `JobGraphWriter` that filters out jobs in `JobGraphWriter#putJobGraph` that serves as a wrapper around another `JobGraphWriter` instance and deals with the filtering of `JobGraph`. 
   
   Additionally, I'm not sure whether you want to filter out `JobResourceRequirements` updates. I feel like we want the request to fail if the user accidentally requested the rescaling for a job that is not allowed to rescale. :thinking:
   
   For the `JobResultStore`, I'd argue that we're not gaining much because the `JobResultStore` is used at the end of the job's lifecycle sending only some requests to S3.
   
   In this sense, I feel like we don't need to introduce yet another component in the `Dispatcher` codebase but rather rely on the `JobGraphWriter` interface for this. WDYT?


-- 
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-unsubscribe@flink.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org