gaborgsomogyi commented on pull request #29729: URL: https://github.com/apache/spark/pull/29729#issuecomment-696621270
@zsxwing it's not explicitly written but the short answer is that the user needs to migrate the application. A little bit more detailed if one wants to authorize with `AdminClient` then `client.id` can be used as alternative (if similar solution is the expectation). @HeartSaVioR the weird behavior related `assign` is true but I would like to add a correction. > One can commit back offests with a consumer and such case group id is used. Of course this doesn't apply here at all since Spark is not committing any offsets. I've no idea about future but I have a strong feeling this stays like this (I mean executors won't commit back offsets). As a general thought I'm not super big fan of `group.id` or now `client.id`. It can be considered as a shared secret which is not the top solution of security(though the possibility is there). My personal opinion is to go on with ACLs. ---------------------------------------------------------------- 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]
