ERROR: org.apache.hadoop.hbase.MasterNotRunningException: null+hbase+hadoop

6
votes


I have recently configured hadoop cluster with two machines(on ubuntu). It works fine so far. But when i try to configure hbase on the above hadoop cluster, it shows error. Here is what i did,

i have two machines. 192.168.1.110 Hadoop master 192.168.1.111 Hadoop slave

conf/hbase-env.sh

export JAVA_HOME=/usr/lib/jvm/java-6-sun-1.6.0.22 export HBASE_CLASSPATH=/home/hadoop/hadoop-0.20.2/conf export HBASE_MANAGES_ZK=true

hbase-site.xml

hbase.master->192.168.1.110:54310(Same as hadoop master:port) hbase.rootdir->hdfs://192.168.1.110:54310/hbase hbase.cluster.distributed->true hbase.zookeeper.quorum->192.168.1.110,192.168.1.111

And region srevers,
192.168.1.111

After starting hbase, i tried to run the command from hbase shell

hbase(main):001:0> list TABLE
It shows

ERROR: org.apache.hadoop.hbase.MasterNotRunningException: null

Please help me up solving the issue Thanks in advance

2
hadoophbase
With Haddop Master you mean the Name Node and with the Hadoop Slave you mean a Datanode ? Have you configured your replication correctly ? Have you tried to run a simple Map/Reduce Job on your cluster to check your setup? – khmarbaise
hi khmarbaise, The hadoop cluster setup is working fine. I have checked with a simple map reduce task(the word count example i have successfully checked), and yes, Master means the Namenode and Slave means the Datanode. Even i have tried With setting up the both the machines as slaves. Still no progress! – hbase_user
If the master:port for the hadoop cluster not different...i have in mind 9001 port? – khmarbaise
hi khmarbaise, i have set 192.168.1.110:54310 as master:port – hbase_user

2 Answers

7
votes

The error is self explaining. if you do jps with the hadoop user on your master(192.168.1.110) machine you would see that HMaster is not running. Please check the HMaster logs to see why it failed.

1
votes

There are some caveats when setting up HBase, especially a cluster. Be patient.

First of all, the cluster is organized by Zookeeper instances. Check your zookeeper logs on both nodes that they connect to each other. It is critical that you set up DNS for your machines and configure the proper domain names for the proper external IPs (192.168.1.11x in your case) in /etc/hosts. If this works (only), then your problem maybe be a different one.

Please closely follow the manual at http://hbase.eu.apache.org/book/notsoquick.html Unfortunately, there is no way around that.