On Sat, Nov 24, 2012 at 10:31 AM, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> I am also using Ubuntu 12.04, Zookeeper 3.4.4 HBase 0.94.2 and Hadoop 1.0.4. (64-bit nodes), I finally managed to have the HBase cluster up and running, below is the line in my /etc/hosts for your reference:
> #127.0.0.1 localhost
> 127.0.0.1 localhost.localdomain localhost
> According to my set up experience, below are my advices:
> 1) /etc/hosts: should not comment out 127.0.01 in /etc/hosts
> 2) Zookeeper: do not sync its "data" and "datalog" folders to other Zookeeper servers in your deployment
> 3) check your start procedures:
> - check your firewall policies, make sure each server can use the required TCP/IP ports, especially port 2181 in your case
> - start Zookeeper first, need to make sure all other servers can access Zookeeper servers, use "/bin/zkCli.sh -server XXXX" or "echo ruok | nc XXXX 2181" to test all Zookeepers from each HBASE server.
> - start Hadoop, use JPS to make sure Namenode, SecondaryNameNode, Datanodes up and running, check LOG files of each servers
> - start MapReduce if you need it
> - start HBase, use JPS to check HBase's HMaster and HRegionServers, then wait a while use JPS to check HMaster and HRegionServers again, if them all HBASE servers gone but HADOOP still up and running, most likely it would be HBASE configure issue in hbase-site.xml related to ZooKeeper settings or ZooKeeper configure/data issues.
> Hope these help and good luck.
Thanks ac for the clean instructions. We have an ubuntu callout here
on localhost in /etc/hosts:
http://hbase.apache.org/book.html#basic.prerequisites What else would
you suggest we add to the reference guide?
P.S. I used to have a friend named AC but in his case it stood for