luke-sterkowicz commented on PR #1898: URL: https://github.com/apache/zookeeper/pull/1898#issuecomment-1189532721
> you can never expect that the graceful shutdown works. it adds lots of races. That's surprising because I believe during leader election zk does an equivalent of graceful shutdown internally, no? > it is better that we guarantee that ZooKeeper works well with KILL -9 There's zk and there's its clients and environment they operate in. In most cases whenever there's SIGTERM, there's also following SIGKILL after a timeout (if SIGTERM fails). So even if graceful shutdown stucks (as you claim), that shouldn't be a problem really and the parent/operator at least has a chance to try. Lack of this provided us a lots of problems in a k8s env with LB on the front of zk. And while LB may not be the most fortunate solution for zk, there will be justifiable scenarios from time to time, and people will face problems and incidents from time to time (like the original reporter of ZOOKEEPER-4400). -- 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: notifications-unsubscr...@zookeeper.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
