OBSOLETE: Go to HBase wiki main page

Here are the steps involved checking out hbase and making it run. Takes about ten minutes.

  1. Download the latest stable release from the HBase download page

  2. Move the download to wherever you want to run hbase from, say to ~/hbase

  3. If running a remote HDFS that you want hbase to connnect to, edit conf/hbase-site.xml and add a property for hbase.rootdir. Set it to hdfs://NAMENODE_HOST:NAMENODE_PORT/HBASE_ROOTDIR. For example, hdfs://localhost:9000/hbase. Otherwise, use the default: file:///tmp/hbase-${user.home}/hbase UNLESS you are running on windows (TODO: Restore what you do here when on windows -- St.Ack).

  4. Edit conf/hbase-env.sh and define JAVA_HOME

  5. Make sure that you can log into localhost without a password (see Hadoop Quick Start, "Setup passphraseless ssh"). Test this out first! New versions of ssh require accepting a new foreign host key before logging in, even if the host is configured with authorized keys.

  6. Start hbase with ~/hbase/bin/start-hbase.sh (hbase logs are viewable in ~/hbase/logs by default)
  7. Enter hbase shell by running ~/hbase/bin/hbase shell; type 'help<ENTER>' to see shell help

  8. Have fun with hbase
  9. Stop the hbase servers with ~/hbase/bin/stop-hbase.sh. Wait until the servers are finished stopping. Avoid killing servers (though a 'kill pid' should be fine as opposed to a 'kill -9 pid').

Based on an hadoop-dev list posting by Dennis Kubes, Sun, 21 Oct 2007 23:09:46 -0500.

Hbase/10Minutes (last edited 2009-10-05 21:30:49 by stack)