kbendick commented on issue #3791: URL: https://github.com/apache/iceberg/issues/3791#issuecomment-999840375
Ok. And can you describe the tables (are they partitioned, how big are they) and also, how youv'e configured the procedure calls (do they have a filter, do they group commits or do they use the default behavior of committing all at the end)? I think this can be solved. I've just started researching more, and I am still working on a reproduction case. If you have one you can send or create a gist for, that would be awesome. But I'm seeing a few options. We can pass a `scheduler` in to [make the expiration more synchronous](https://github.com/ben-manes/caffeine/wiki/Compute#synchronous-listeners), and then alternatively we can also consider trying to set both `refreshAfter` and `expireAfter`, as `refreshAfter` will mark the entry for removal, and then on the next round, it will return the [old value and then evict the entry (so the next get will return right away)](https://github.com/ben-manes/caffeine/issues/204#issuecomment-345312109). -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
