Daniel Schierbeck
Daniel Schierbeck
Can you try updating your version of librdkafka? It looks to me like the configuration setting isn't present in the version you're using. And please report back what version _did_...
@magicguitarist would you have time to write a PR that catches the exception and provides a more helpful error message?
Can you open a PR mentioning this in the README?
I'm hoping that we can leave the consumer API as-is, and just have this be a configuration. Since both `process` and `process_batch` work on single-partition data, we should be able...
It seems like the specs are broken because – can you investigate?
@daviduebelacker can you try rebasing on latest master? Maybe that fixes the build.
If your session timeout is 5 minutes, that means that a consumer can theoretically process messages for 5 minutes before figuring out that the group has been re-synchronized. At that...
Also: by opting to use rolling updates you'll get these sorts of issues. I'm also not quite sure why you'd get more availability – the whole group needs to re-sync...
Try to simplify your config and see if that doesn't resolve the issue. It's tricky to debug these things, unfortunately. The problem could be in Racecar, rdkafka-ruby, or in librdkafka.
@thijsc this looks more like a problem in rdkafka-ruby or perhaps librdkafka? Should this be handled in Racecar at all in your opinion? As far as I understand, the "group...