vagrant-hadoop-spark-cluster

https://github.com/dnafrance/vagrant-hadoop-spark-cluster

1. Introduction

Vagrant project to spin up a cluster of 4, 32-bit CentOS6.5 Linux virtual machines with Hadoop v2.6.0 and Spark v1.1.1.

Ideal for development cluster on a laptop with at least 4GB of memory.

1. node1 : HDFS NameNode + Spark Master
2. node2 : YARN ResourceManager + JobHistoryServer + ProxyServer
3. node3 : HDFS DataNode + YARN NodeManager + Spark Slave
4. node4 : HDFS DataNode + YARN NodeManager + Spark Slave

2. Prerequisites and Gotchas to be aware of

1. At least 1GB memory for each VM node. Default script is for 4 nodes, so you need 4GB for the nodes, in addition to the memory for your host machine.
2. Vagrant 1.7 or higher, Virtualbox 4.3.2 or higher
3. Preserve the Unix/OSX end-of-line (EOL) characters while cloning this project; scripts will fail with Windows EOL characters.
4. Project is tested on Ubuntu 32-bit 14.04 LTS host OS; not tested with VMware provider for Vagrant.
5. The Vagrant box is downloaded to the ~/.vagrant.d/boxes directory. On Windows, this is C:/Users/{your-username}/.vagrant.d/boxes.

3. Getting Started

1. Download and install VirtualBox
2. Download and install Vagrant.
3. Run vagrant box add centos65 http://files.brianbirkinbine.com/vagrant-centos-65-i386-minimal.box
4. Git clone this project, and change directory (cd) into this project (directory).
5. Download Hadoop 2.6 into the /resources directory
6. Download Spark 1.1.1 into the /resources directory
7. Download Java 1.8 into the /resources directory
8. Run vagrant up to create the VM.
9. Run vagrant ssh to get into your VM.
10. Run vagrant destroy when you want to destroy and get rid of the VM.

4. Modifying scripts for adapting to your environment

You need to modify the scripts to adapt the VM setup to your environment.

1. List of available Vagrant boxes
2. ./Vagrantfile
   To add/remove slaves, change the number of nodes:
   line 5: numNodes = 4
   To modify VM memory change the following line:
   line 13: v.customize ["modifyvm", :id, "--memory", "1024"]
3. /scripts/common.sh
   To use a different version of Java, change the following line depending on the version you downloaded to /resources directory.
   line 4: JAVA_ARCHIVE=jdk-8u25-linux-i586.tar.gz
   To use a different version of Hadoop you've already downloaded to /resources directory, change the following line:
   line 8: HADOOP_VERSION=hadoop-2.6.0
   To use a different version of Hadoop to be downloaded, change the remote URL in the following line:
   line 10: HADOOP_MIRROR_DOWNLOAD=http://apache.crihan.fr/dist/hadoop/common/stable/hadoop-2.6.0.tar.gz
   To use a different version of Spark, change the following lines:
   line 13: SPARK_VERSION=spark-1.1.1
   line 14: SPARK_ARCHIVE=$SPARK_VERSION-bin-hadoop2.4.tgz
   line 15: SPARK_MIRROR_DOWNLOAD=../resources/spark-1.1.1-bin-hadoop2.4.tgz
4. /scripts/setup-java.sh
   To install from Java downloaded locally in /resources directory, if different from default version (1.8.0_25), change the version in the following line:
   line 18: ln -s /usr/local/jdk1.8.0_25 /usr/local/java
   To modify version of Java to be installed from remote location on the web, change the version in the following line:
   line 12: yum install -y jdk-8u25-linux-i586
5. /scripts/setup-centos-ssh.sh
   To modify the version of sshpass to use, change the following lines within the function installSSHPass():
   line 23: wget http://pkgs.repoforge.org/sshpass/sshpass-1.05-1.el6.rf.i686.rpm
   line 24: rpm -ivh sshpass-1.05-1.el6.rf.i686.rpm
6. /scripts/setup-spark.sh
   To modify the version of Spark to be used, if different from default version (built for Hadoop2.4), change the version suffix in the following line:
   line 32: ln -s /usr/local/$SPARK_VERSION-bin-hadoop2.4 /usr/local/spark

5. Post Provisioning

After you have provisioned the cluster, you need to run some commands to initialize your Hadoop cluster. SSH into node1 using
vagrant ssh node-1 Commands below require root permissions. Change to root access using sudo su or create a new user and grant permissions if you want to use a non-root access. In such a case, you'll need to do this on VMs.
Issue the following command.

1. $HADOOP_PREFIX/bin/hdfs namenode -format myhadoop

Start Hadoop Daemons (HDFS + YARN)

SSH into node1 and issue the following commands to start HDFS.

1. $HADOOP_PREFIX/sbin/hadoop-daemon.sh --config $HADOOP_CONF_DIR --script hdfs start namenode
2. $HADOOP_PREFIX/sbin/hadoop-daemons.sh --config $HADOOP_CONF_DIR --script hdfs start datanode

SSH into node2 and issue the following commands to start YARN.

1. $HADOOP_YARN_HOME/sbin/yarn-daemon.sh --config $HADOOP_CONF_DIR start resourcemanager
2. $HADOOP_YARN_HOME/sbin/yarn-daemons.sh --config $HADOOP_CONF_DIR start nodemanager
3. $HADOOP_YARN_HOME/sbin/yarn-daemon.sh start proxyserver --config $HADOOP_CONF_DIR
4. $HADOOP_PREFIX/sbin/mr-jobhistory-daemon.sh start historyserver --config $HADOOP_CONF_DIR

Test YARN

Run the following command to make sure you can run a MapReduce job.

yarn jar /usr/local/hadoop/share/hadoop/mapreduce/hadoop-mapreduce-examples-2.6.0.jar pi 2 100

Start Spark in Standalone Mode

SSH into node1 and issue the following command.

1. $SPARK_HOME/sbin/start-all.sh

Test Spark on YARN

You can test if Spark can run on YARN by issuing the following command. Try NOT to run this command on the slave nodes.

$SPARK_HOME/bin/spark-submit --class org.apache.spark.examples.SparkPi \
    --master yarn-cluster \
    --num-executors 10 \
    --executor-cores 2 \
    lib/spark-examples*.jar \
    100

Test Spark using Shell

Start the Spark shell using the following command. Try NOT to run this command on the slave nodes.

$SPARK_HOME/bin/spark-shell --master spark://node1:7077

Then go here https://spark.apache.org/docs/latest/quick-start.html to start the tutorial. Most likely, you will have to load data into HDFS to make the tutorial work (Spark cannot read data on the local file system).

6. Web UI

You can check the following URLs to monitor the Hadoop daemons.

1. NameNode
2. ResourceManager
3. JobHistory
4. Spark

7. References

This project was put together with great pointers from all around the internet. All references made inside the files themselves. Primaily this project is forked from Jee Vang's vagrant project

https://github.com/sequenceiq/docker-spark
http://blog.sequenceiq.com/blog/2015/01/09/spark-1-2-0-docker/

Apache Spark on Docker

This repository contains a Docker file to build a Docker image with Apache Spark. This Docker image depends on our previous Hadoop Docker image, available at the SequenceIQ GitHub page. The base Hadoop Docker image is also available as an official Docker image.

Pull the image from Docker Repository

$sudo docker pull sequenceiq/spark:1.5.1

Building the image

$sudo docker build --rm -t sequenceiq/spark:1.5.1 .

Running the image

• if using boot2docker make sure your VM has more than 2GB memory
• in your /etc/hosts file add $(boot2docker ip) as host 'sandbox' to make it easier to access your sandbox UI
• open yarn UI ports when running container

$sudo docker run -it -p 8088:8088 -p 8042:8042 -h sandbox sequenceiq/spark:1.5.1 bash

or

$sudo docker run -it -h sandbox sequenceiq/spark:1.5.1 bash

$cd /usr/local/spark 

or

$sudo docker run -d -h sandbox sequenceiq/spark:1.5.1 -d

Versions

Hadoop 2.6.0 and Apache Spark v1.5.1 on Centos 

Testing

There are two deploy modes that can be used to launch Spark applications on YARN.

Disable logs

Just execute this command in the spark directory:

cp conf/log4j.properties.template conf/log4j.properties

Edit log4j.properties:

Replace at the first line:

log4j.rootCategory=INFO, console

by

log4j.rootCategory=WARN, console   or  log4j.rootCategory=ERROR, console

Save and restart your shell. It works for me for Spark 1.1.0 and Spark 1.5.1 on OS X.

YARN-client mode

In yarn-client mode, the driver runs in the client process, and the application master is only used for requesting resources from YARN.

# run the spark shell
spark-shell \
--master yarn-client \
--driver-memory 1g \
--executor-memory 1g \
--executor-cores 1

# execute the the following command which should return 1000
scala> sc.parallelize(1 to 1000).count()

YARN-cluster mode

In yarn-cluster mode, the Spark driver runs inside an application master process which is managed by YARN on the cluster, and the client can go away after initiating the application.

Estimating Pi (yarn-cluster mode):

# execute the the following command which should write the "Pi is roughly 3.1418" into the logs
# note you must specify --files argument in cluster mode to enable metrics
spark-submit \
--class org.apache.spark.examples.SparkPi \
--files $SPARK_HOME/conf/metrics.properties \
--master yarn-cluster \
--driver-memory 1g \
--executor-memory 1g \
--executor-cores 1 \
$SPARK_HOME/lib/spark-examples-1.5.1-hadoop2.6.0.jar

Estimating Pi (yarn-client mode):

# execute the the following command which should print the "Pi is roughly 3.1418" to the screen
spark-submit \
--class org.apache.spark.examples.SparkPi \
--master yarn-client \
--driver-memory 1g \
--executor-memory 1g \
--executor-cores 1 \
$SPARK_HOME/lib/spark-examples-1.5.1-hadoop2.6.0.jar

Install Docker on Ubuntu 14.04 LTS

Introduction

Docker is a container-based software framework for automating deployment of applications. “Containers” are encapsulated, lightweight, and portable application modules.

Pre-Flight Check

• These instructions are intended for installing Docker.
• I’ll be working from a Liquid Web Core Managed Ubuntu 14.04 LTS server, and I’ll be logged in as root.

Step 1: Installation of Docker

First, you’ll follow a simple best practice: ensuring the list of available packages is up to date before installing anything new.
apt-get update

Let’s install Docker by installing the docker-io package:
apt-get -y install docker.io

Link and fix paths with the following two commands:
ln -sf /usr/bin/docker.io /usr/local/bin/docker
sed -i '$acomplete -F _docker docker' /etc/bash_completion.d/docker

Finally, and optionally, let’s configure Docker to start when the server boots:
update-rc.d docker defaults

Step 2: Download a Docker Container

Let’s begin using Docker! Download the fedora Docker image:
docker pull ubuntu

Step 3: Run a Docker Container

Now, to setup a basic ubuntu container with a bash shell, we just run one command. docker run will run a command in a new container, -i attaches stdin and stdout, -t allocates a tty, and we’re using the standard ubuntu container.
docker run -i -t ubuntu /bin/bash

That’s it! You’re now using a bash shell inside of a ubuntu docker container.
To disconnect, or detach, from the shell without exiting use the escape sequence Ctrl-p + Ctrl-q.

There are many community containers already available, which can be found through a search. In the command below I am searching for the keyword debian:
docker search debian