noodle-learns-programming / node-hbase Goto Github PK

 ___   _           _          _     _ ______                   
|   \ | |         | |        | |   | |____  \                  
| |\ \| | ___   __| |_____   | |___| |____)  )_____  ___ _____ 
| | \   |/ _ \ / _  | ___ |  |  ___  |  __  ((____ |/___) ___ |
| |  \  | |_| ( |_| | ____|  | |   | | |__)  ) ___ |___ | ____|
|_|   |_|\___/ \____|_____)  |_|   |_|______/\_____(___/|_____) New BSD License

Node HBase is a Node.JS client for the Apache HBase database. It use the Rest API (Stargate) to communicate with HBase. Currently, all the API is implemented and the data exchange format is JSON (but protocol buffer could follow).

Apache HBase is part of the Hadoop ecosystem. It describes itself as the Hadoop database optimized for random, realtime read/write access to big data. It is an open-source, distributed, versioned, column-oriented store modeled after Google Bigtable.

Client features include:

• Intuitive API following HBase naming conventions
• Documentation and tests
• Full Implementation of the REST API
• Transparent encoding/decoding of values
• Scanner and filter support implementing the stream.Readable API
• Kerberos Support

Via npm:

npm install hbase

This code create a new HBase instance, create a table and a column family, insert a few records and traverse them.

var assert = require('assert');
var hbase = require('hbase');

hbase({ host: '127.0.0.1', port: 8080 })
.table('my_table' )
.create('my_column_family', function(err, success){
  this
  .row('my_row')
  .put('my_column_family:my_column', 'my value', function(err, success){
    this.get('my_column_family', function(err, cells){
      this.exists(function(err, exists){
        assert.ok(exists);
      });
    });
  });
});

Options accepts a krb5 object. Password and keytab authentication are supported. Refer to the krb5 package for additionnal information on how to configure the krb5 option.

Using a keytab:

var hbase = require('hbase');
hbase({
  host: '127.0.0.1',
  port: 8080,
  "krb5": {
    "principal": "{username}@{REALM}",
    "keytab": "{path/to/keytab}",
    "service_principal": "HTTP@{fqdn}"
  }
})
.version();

Using a password:

var hbase = require('hbase');
hbase({
  host: '127.0.0.1',
  port: 8080,
  "krb5": {
    "principal": "{username}@{REALM}",
    "password": "{password}",
    "service_principal": "HTTP@{fqdn}"
  }
})
.version();

The scanner implement the stream.Readable API. For ease of usage, an optional callback argument may be provided. For example:

client
.table('node_table')
.scan({
  startRow: 'my_row',
  maxVersions: 1
}, function(err, rows){
  console.log(err, rows);
});

is equivalent to:

var rows = [];
scanner = client
.table('node_table')
.scan({
  startRow: 'my_row',
  maxVersions: 1
});
scanner.on('readable', function(){
  while(chunk = scanner.read()){
    rows.push(chunk);
  }
});
scanner.on('error', function(err){
  console.log(err);
});
scanner.on('end', function(){
  console.log(rows);
});

It can be quite a pain to figure out what options can be sent with a scanner request. You will find a lot of examples inside the Scanner test and also look at the examples published by Marc Trudel.

Tests are executed with mocha. Before running the tests the first time, copy the file "./test/properties.json.sample" to "./test/properties.json" and make the appropriate changes.

If using the HDP sandbox, start the virtual machine, log-in as "root", start Ambari start_ambari.sh, start HBase start_hbase.sh and start the HBase REST server /usr/lib/hbase/bin/hbase rest -p 60080.

To run the tests:

make test

When testing against HBase secured with Kerberos, you must create a table with the right ownership.

kinit hbase
hbase shell
create 'node_table1', {NAME => 'node_column_family', VERSIONS => 5}
grant 'ryba', 'RWC', 'node_table'

You can use the example located in "test/properties.json.krb5" to configure the test. It comes pre-configured for Ryba configured in development cluster mode.

• Official Apache HBase project
• REST server bundled with HBase (Stargate)

• David Worms: https://github.com/wdavidw
• Michael Kurze: https://github.com/michaelku
• Michal Taborsky: https://github.com/whizz