> On Oct. 7, 2013, 5 p.m., Neha Narkhede wrote: > > core/src/main/scala/kafka/consumer/ZookeeperConsumerConnector.scala, line > > 319 > > <https://reviews.apache.org/r/14497/diff/1/?file=361459#file361459line319> > > > > This comment is a bit misleading. The data change watch for the > > /brokers/topics path is only set in reinitializeConsumer() which is invoked > > on startup and when new topics are discovered through wildcards. Doesn't > > zkclient automatically re-register watches once they are fired ? > > Guozhang Wang wrote: > I am not sure if the zkClient auto re-register upon watcher fired, but I > read from the source code that the watcher will be re-registered upon reading > the subscribed path. I made the above comment since in rebalance the > getPartitionAssignmentForTopics function call will read the /brokers/topics > path and hence set the watchers. > > Guozhang Wang wrote: > This is what I found for IZkDataListener (the English seems not very easy > to understand though :P ): > > "An IZkDataListener can be registered at a ZkClient for listening on zk > data changes for a given path. Node: Also this listener re-subscribes it > watch for the path on each zk event (zk watches are one-timers) is is not > guaranteed that events on the path are missing (see > http://zookeeper.wiki.sourceforge.net/ZooKeeperWatches). An implementation of > this class should take that into account. "
http://pastebin.com/pZQAXnxd. Zkclient does re-register the watch *before* reading the path, which is what I wanted to check. I think we are good to go here. "I made the above comment since in rebalance the getPartitionAssignmentForTopics function call will read the /brokers/topics path and hence set the watchers." This is not true. It re-registers the watch when it fires the data change event, not when you read the path. It is worth making that clear in your comment. - Neha ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/14497/#review26705 ----------------------------------------------------------- On Oct. 7, 2013, 5:25 p.m., Guozhang Wang wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/14497/ > ----------------------------------------------------------- > > (Updated Oct. 7, 2013, 5:25 p.m.) > > > Review request for kafka. > > > Bugs: KAFKA-1075 > https://issues.apache.org/jira/browse/KAFKA-1075 > > > Repository: kafka > > > Description > ------- > > KAFKA-1075.v2 > > > KAFKA-1075.v1 > > > Diffs > ----- > > core/src/main/scala/kafka/consumer/ZookeeperConsumerConnector.scala > 08b4b7218f62f876eb95628bd6db78c0956c7f04 > > Diff: https://reviews.apache.org/r/14497/diff/ > > > Testing > ------- > > > Thanks, > > Guozhang Wang > >
