gianm opened a new pull request #7558: BaseAppenderatorDriver: Fix potentially overeager segment cleanup. URL: https://github.com/apache/incubator-druid/pull/7558 Here is a thing that I think can go wrong: 1. We push some segments, then try to publish them transactionally. 2. The segments are actually published, but the 200 OK response gets lost (connection dropped, whatever). 3. We try again, and on the second try, the publish fails (because the transaction baseline start metadata no longer matches). 4. Because the publish failed, we delete the pushed segments. 5. But this is bad, because the publish didn't really fail, it actually succeeded in step 2. I haven't seen this in the wild, but thought about it while reviewing #7537. This patch also cleans up logging a bit, making it more accurate and somewhat less chatty.
---------------------------------------------------------------- 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] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
