yahoNanJing commented on issue #132: URL: https://github.com/apache/arrow-ballista/issues/132#issuecomment-1216251328
> How would you support all-at-once scheduling in this setup for streaming execution? Do you mean long running tasks for streaming case? Would you submit hundreds of SQLs or jobs to the scheduler for the streaming case? If not, I think single active scheduler will satisfy your needs. Then you don't need active multiple schedulers, and for single active scheduler you don't need to save every update to the backend. For #59, I think it will work well when there are few tasks and few tasks update. However, when there are thousands of tasks for a job, even a stage, the lock and serialization and deserialization cost would be huge. That's why we added memory cache for the scheduler state previously. However, the cache layer has been removed by #59. 😠-- 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: github-unsubscr...@arrow.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
