Hi. If you subscribe to zookeeper mailinglist, you would see a hdfs developer working on bookeeper recently. https://issues.apache.org/jira/browse/ZOOKEEPER-465 https://issues.apache.org/jira/browse/ZOOKEEPER-959
Probably related. https://issues.apache.org/jira/browse/HDFS-234 Koji On 1/11/11 11:35 PM, "mac fang" <mac.had...@gmail.com> wrote: Hi, all, I am NOT sure if there any plan or issues of integrating Zookeeper and NameNode to resolve the Single Point issue of NameNode. With it, the DataNodes dont need to be restarted. I think this can improve the availability of the whole cluster. I heard there is a plan for JobTracker and TaskTracker, but i dont know if there is the one for HDFS. Currently, i am working on it. :) Any issues or comments are welcome. regards macf
