RussellSpitzer commented on pull request #30167: URL: https://github.com/apache/spark/pull/30167#issuecomment-717965512
I think the use case you are specifying is rather unlikely compared to users who set this property and expect it to work or not work. It's hard for me to imagine someone relying on the fact that any exception in catalog initialization is ok, the domain of possible failures is just too large. It's more like their application works in spite of their intent, rather than in coordination with it. From my perspective, the vast majority of users who set this property are expecting to use it and if it doesn't work they will stop and change the configuration but we make harder to determine by ignoring the exception and even harder than that by swallowing the stack trace. While the flip side is much harder for me to imagine, a user ok with the behavior which can occur for any number of reasons they cannot actually predict. So again at minimum we should be logging the exception and trace but I don't think we actually disrupt any real use cases by failing fast either. ---------------------------------------------------------------- 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]
