ztison commented on PR #25218: URL: https://github.com/apache/flink/pull/25218#issuecomment-2401913141
Hi, I was wondering if it's advisable to change the default slot assigner. Wouldn't it be more appropriate to keep the default setting, which spreads the workload across as many workers as possible? This approach enhances the system's availability and resilience. The system will need to migrate the entire state to a new TM if it crashes, is it what we want in default implementation? Isn't it a better approach to have a config option to activate the proposed behavior as defined in FLINK-36426? Then we could use a new slot assigner instead of the `DefaultSlotAssigner` and `StateLocalitySlotAssigner`. I'm also wondering if the same strategy can be applied in the Resource Manager to allocate slots from a minimized number of Task Managers (TMs). -- 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-unsubscr...@flink.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
