General Information.

Apache Ambari is a api based tool for provisioning, managing, and monitoring Apache Hadoop clusters. The set of Hadoop components that are currently supported by Ambari includes:

• Apache Hadoop - HDFS
• Apache Hadoop - MapReduce
• Apache Hive
• Apache HCatalog
• Apache HBase
• Apache Zookeeper
• Apache Oozie
• Apache Pig
• Apache Sqoop
• Ambari enables System Administrators to:

Provision a Hadoop Cluster
Ambari provides an easy-to-use, step-by-step wizard for installing Hadoop services across any number of hosts.
Ambari handles configuration of Hadoop services for the cluster.
Manage a Hadoop Cluster
Ambari provides central management for starting, stopping, and reconfiguring Hadoop services across the entire cluster.
Monitor a Hadoop Cluster
Ambari provides a dashboard for monitoring health and status of the Hadoop cluster.
Ambari leverages Ganglia for metrics collection.
Ambari leverages Nagios for system alerting and will send emails when your attention is needed (e.g., a node goes down, remaining disk space is low, etc)

Checking out AMBARI-666

The following uses Ambari git for the development process. Please keep in mind git is read only and patches will have to be uploaded and committed via svn.

git clone git://git.apache.org/ambari.git
cd ambari
git checkout -b AMBARI-666 origin/AMBARI-666

We'll refer to the top-level "ambari" directory as AMBARI_DIR in this document.

Building Ambari

After checking out Ambari, you can build ambari with the steps below:

$ mvn clean install 

Note: The above is for building the jars. See below for starting/deploying the Ambari Server and the Ambari Agent.

Building Artifacts and Running the Server/Agent

To build tarball artifacts run:

$ mvn clean package assembly:assembly

This will build the Server and Agent tarballs.

NOTE: Run everything as root below.

Running the Ambari Server:

Execute the following, where INSTALL_DIR is a directory where you want to run the Server:

$ cp ambari-server/target/ambari-server-*-dist.tar.gz INSTALL_DIR
$ cd INSTALL_DIR
$ tar -zxvf ambari-server-*-dist.tar.gz

Edit the Ambari properties configuration file:

$ vim /etc/ambari/ambari.properties

Add the following entry:

security.server.keys_dir=INSTALL_DIR/ambari-server-1.0.3-SNAPSHOT/keystore

NOTE: replace INSTALL_DIR with the actual directory path.

$ INSTALL_DIR
$ cd ambari-server-*
$ export AMBARI_PASSPHRASE="tmp"
$ cp /etc/ambari/ambari.properties etc/ambari-server/conf/
$ java -cp etc/ambari-server/conf/:lib/ambari-server/* org.apache.ambari.server.controller.AmbariServer

NOTE: the conf directory should come first in the java classpath because guice-persist.jar includes a log4j.properties which will cause the Ambari log4j.properties to not load.

The Server will launch and run as a foreground process. Scripting will come later as we move along to run the server with proper confs/envs.

To check if the server is running and healthy, you can go to:

http://localhost:8080/api/check

The status will show as running. Or from a command line, you can run:

curl localhost:8080/api/check

NOTE: The health check is static response for now.

Running the Ambari Agent.

The Ambari Agent requires python 2.6 to run on a host. RHEL5/CentOS5 has python 2.4 installed by default. To install python 2.6 (NOTE: you can install both 2.4/2.6 on the host), install the epel repo and then run:

$ yum install python26.

Python needs setuptools to be able to setup the project for development. Download http://pypi.python.org/packages/2.6/s/setuptools/setuptools-0.6c11-py2.6.egg#md5=bfa92100bd772d5a213eedd356d64086^{Image Removed} and run:

$ sh setuptools-0.6c11-py2.6.egg

To setup the Ambari Agent:

$ mkdir -p /var/ambari /var/run/ambari /var/log/ambari /tmp/ambari-agent/

The following dev process is easier than using setup.py and is a quicker turnaround. As soon as we have a good deployment story (rpm) we will add those in.

$ cd ambari-agent/src/main/python/ambari_agent

• Ambari agent requires facter to get the node host information. Download facter tarball from http://downloads.puppetlabs.com/facter/facter-1.6.10.tar.gz^{Image Removed} and uninstall in a directory.
• Edit facter_home in AmbariConfig.py to the directory where facter is installed. The agent executes the script facter_home/bin/facter to get all the node information and also uses df to get the mount point information.
• Ambari agent requires puppet. Download puppet tarball from http://www.puppetlabs.com/downloads/puppet/puppet-2.7.9.tar.gz^{Image Removed} and uninstall in a directory.
• Edit puppet_home in AmbariConfig.py to the directory where puppet is installed. The agent executes puppet_home/bin/puppet to run puppet scripts.

To run the agent:

$ export AMBARI_PASSPHRASE="tmp"

Copy puppet modules to /var/lib/ambari-agent/puppet/

$ cp -R ../../puppet/modules/ /var/lib/ambari-agent/puppet
$ python2.6 main.py

The Ambari Agent will just try connecting to the server. To see the logs, take a look at:

/var/log/ambari-agent/ambari-agent.log

Building and Running Ambari Web

Ambari Web uses Brunch (http://brunch.io/) as the application assembler for development purposes. Brunch is used to build and deploy automatically in the background as the developer modifies source files (JS, LESS/CSS, Handlebars template files). Brunch can also start up a Node.js-based web server with a single command so that you can easily run the application from the application root directory without any special configuration.

To build, deploy and run Ambari Web:

1. Install NodeJS and NPM from http://nodejs.org/download/
2. Install Brunch

   Code Block
   $ npm install -g brunch

3. Download dependent packages into ambari-web/node_modules directory

   Code Block
   $ cd AMBARI_DIR/ambari-web $ npm install

To run Ambari Web so that it works with mock data independently from Ambari Server:

1. Start test server:

   Code Block
   $ brunch watch --server (or simply "brunch w --s")

2. Edit ambari-web/app/initialize.js and set to test mode (testMode = true). You can optionally set to go to the install wizard or directly into the app dashboard.

   Code Block
   App.testMode = true; App.skipBootstrap = true; App.alwaysGoToInstaller = true;

3. Access http://localhost:3333^{Image Removed} from your web browser. When you change Ambari Web source files, changes are immediately reflected.

To run Ambari Web with Ambari Server integration:

$ cd AMBARI_DIR
$ mvn clean package assembly:assembly

Note: currently the mvn command does not put ambari-web in the Maven-generated package. This will be fixed in the future.

$ cd AMBARI_DIR/ambari-server/ambari/target/ambari-server-*-dist/ambari-server-*
$ java -cp etc/ambari-server/conf/:lib/ambari-server/* org.apache.ambari.server.controller.AmbariServer
$ cd AMBARI_DIR/ambari-web
$ brunch watch

Access http://localhost:8080^{Image Removed} from your web browser. When you change Ambari Web source files, changes are immediately reflected.

Setting up Ambari in Eclipse

$ mvn clean eclipse:eclipse

After doing the above you should be able to import the project via Eclipse "Import > Maven > Existing Maven Project". Choose the root directory where you cloned the git repository. You should be able to see the following projects on eclipse:

ambari
|
|- ambari-project
|- ambari-server
|- ambari-agent
|- ambari-web

Select the top-level "ambari pom.xml" and click Finish.