adriancole commented on pull request #6039: URL: https://github.com/apache/incubator-pinot/pull/6039#issuecomment-707063779
functionally, this works if the "add table" commands are done last (because effectively, you end up needing to wait for everything anyway). Verified here https://github.com/hypertrace/pinot/pull/48 It may seem not that much better if done in java vs shell scripts, especially as everything has to be ready first. However, java is easier to test, there's less network problems, and a chance to be considered a bootstrap operation (crash the process if it fails, participate in overall health) I understand that there's some debate about where to stuff the add table command. I also understand there is some angst about this temporarily using the "service" role. This was to perform the least change as literally ServiceManager is designed to work on services. I don't really care if this is called a service or not. However, there's a lack of an alternative also. This brings to the next action. I'm hoping someone maybe @kishoreg can find an agreeable place to house this logic. Once those debates are over, I can take a look at backfilling tests etc in that context. Of course, if someone wants this code to do on own, they can do that also. NEXT STEP: knowing a role ADD_TABLE won't fly, await an alternative. ---------------------------------------------------------------- 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: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org
