HBASE SETUP

Set Up Pseudo-Distributed Mode

1. Verify Installation Requirements

– Java, password-less SSH

2. Configure Java

3. Configure the use of HDFS

– Specify the location of Namenode

– Configure replication

4. Make sure HDFS is running

5. Start HBase

6. Verify HBase is running

6

1: Verify Installation Requirements

• SSH installed, sshd must be running

– Just like Hadoop

– Need password-less SSH to all the nodes including yourself

– Required for both pseudo-distributed and fully-distributed modes

2: Configure Java

• vi <HBASE_HOME>/conf/hbase-env.sh

export JAVA_HOME=/usr/java/jdk1.6.0

9

3: Configure the use of HDFS

• Point to HDFS for its filesystem

– Edit <hbase_home>/conf/hbase-site.xml

– hbase.rootdir property:

– Example: hdfs://localhost:9000/hbase

– dfs.replication property:1

<property>

<name>hbase.rootdir</name>

<value>hdfs://localhost:9000/hbase</value>

</property>

...

4: Make sure HDFS is running

– Easy way is to check web-based management console

• http://localhost:50070/dfshealth.jsp

– Or use command line

• $ hdfs dfs -ls /

13

5: Start HBase

14

$ cd <hbase_home>/bin

$ ./start-hbase.sh

starting master, logging to /home/hadoop/Training/logs/hbase/hbase-hadoop-masterhadoop-laptop.out

6: Verify HBase is Running

15

$ hbase shell

HBase Shell; enter 'help<RETURN>' for list of supported commands.

$ hadoop fs -ls /hbase

Found 5 items

drwxr-xr-x - hadoop supergroup 0 2011-12-31 13:18 /hbase/-ROOTdrwxr-

xr-x - hadoop supergroup 0 2011-12-31 13:18 /hbase/.META.

drwxr-xr-x - hadoop supergroup 0 2011-12-31 13:18 /hbase/.logs

drwxr-xr-x - hadoop supergroup 0 2011-12-31 13:18 /hbase/.oldlogs

-rw-r--r-- 1 hadoop supergroup 3 2011-12-31 13:18 /hbase/hbase.version

HBase data and metadata is stored in HDFS16

By default HBase manages Zookeeper daemon for you

• Logs by default go to <hbase_home>/logs

– Change the default by editing <hbase_home>/conf/hbaseenv.sh

• export HBASE_LOG_DIR=/new/location/logs

HBase comes with web based management

– http://localhost:60010

Display cluster's status via status command

– hbase> status

– hbase> status 'detailed'