HeartSaVioR edited a comment on pull request #30770: URL: https://github.com/apache/spark/pull/30770#issuecomment-747154451
Let's be clear, it is relying on luck "as it is", it requires non-trivial change to not rely on luck. E.g. You can make state store coordinator to track executors being inactive but not evicted as well, and take all (but with preference for batches these executor are loaded) into account for calculating preferred location. That's the prerequisite of replication I've mentioned before, but it can be used for this PR as well (till TTL takes place). Now my previous comments make sense to you? ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
