jerryshao commented on issue #9165: URL: https://github.com/apache/gravitino/issues/9165#issuecomment-3691083247
> Okay, I understand what you mean. You're saying that an HTTP job template can be used to implement a new executor on a distributed scheduler, and the local executor is only used for process verification. > > So, I'd like to ask, if high-concurrency production tasks are delegated to the distributed scheduler, does that mean the entire job lifecycle management is also handled by the distributed scheduler? > Yes, the entire lifecycle of the job should be managed by the external scheduler. You can treat Gravitino as a job submitter; after submission, Gravitino will not handle further things. > Or is it necessary to periodically poll the job status and persist it to the job's MySQL table? Pulling the status periodically provides a convenient way to check the job status. Theoretically, we can directly get the status from the scheduler; there is no need to store it in Gravitino. But this invocation "from client to Gravitino server to the scheduler" may be time-consuming; storing it in Gravitino can save the time of invocation. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
