On Tue, Mar 7, 2017 at 10:10 AM Charan Reddy G <reddychara...@gmail.com> wrote:
> But due to the limitation of ZK (which is well documented thing in ZK > https://zookeeper.apache.org/doc/r3.4.5/zookeeperProgrammers.html#ch_ > zkWatches), "Because watches are one time triggers and there is latency > between getting the event and sending a new request to get a watch you > cannot reliably see every change that happens to a node in ZooKeeper. Be > prepared to handle the case where the znode changes multiple times between > getting the event and setting the watch again." > The way I understand this is that you might get less notifications than the number of changes on the z-node, but when you set the watch, it should be tied to the z-node version that you have already read, so you should be get notified for any change after that. Resuming, you might not see all intermediate states, but you should guaranteed to converge to final state. Probably Flavio can give more insight on this. Matteo
