Home > Apache Mesos, Hadoop, Kafka, Open Source Projects, Zookeeper > Open Source Cloud Formation with Minotaur for Mesos, Kafka and Hadoop

Open Source Cloud Formation with Minotaur for Mesos, Kafka and Hadoop

Today I am happy to announce “Minotaur” which is our Open Source AWS based infrastructure for managing big data open source projects including (but not limited too): Apache Kafka, Apache Mesos and Cloudera’s Distribution of Hadoop. Minotaur is based on AWS Cloud Formation.

The following labs are currently supported:

Supervisor

Supervisor is a Docker-based image that contains all the necessary software to manage nodes/resources in AWS.

Supervisor set-up

  • clone this repo to repo_dir
  • cd to repo_dir/supervisor folder

Before trying to build docker image, you must put some configuration files under config directory:

a) aws_config

This file is just a regular aws-cli config, you must paste your secret and access keys, provided by Amazon in it:

[default]
output = json
region = us-east-1
aws_access_key_id = SECRET_KEY
aws_secret_access_key = ACCESS_KEY

Do not add or remove any extra whitespaces (especially before and after “=” sign in keys)

b) private.key

This is your private SSH key, public part of which is registered on Bastion host

c) environment.key

This is a shared key for all nodes in environment Supervisor is supposed to manage.

d) ssh_config

This is a regular SSH config file, you have to change your_username only (this is the one registered on Bastion).

BASTION_IP is handled dynamically when container is built.

# BDOSS environment
Host 10.0.2.*
    IdentityFile ~/.ssh/environment.key 
    User ubuntu
    ProxyCommand  ssh -i ~/.ssh/private.key your_username@BASTION_IP nc %h %p
Host 10.0.*.*
    IdentityFile ~/.ssh/environment.key 
    User ubuntu
    ProxyCommand  ssh -i ~/.ssh/private.key your_username@BASTION_IP nc %h %p
  • exec up.sh:

If this is the first time you’re launching supervisor – it will take some time to build.

Subsequent up’s will take seconds.

Using supervisor

Now you can cd to /deploy/labs/ and deploy whatever you want

Example:


minotaur lab deploy mesosmaster -e bdoss-dev -d test -r us-east-1 -z us-east-1a
Creating new stack 'mesos-master-test-bdoss-dev-us-east-1-us-east-1a'...
Stack deployed.

this will spin up a mesos master node in “testing” deployment.

awsinfo

Supervisor has a built-in “awsinfo” command, which relies on AWS API and provides brief info about running machines. It is also capable of searching through that info.

Usage example

awsinfo – will display brief info about all nodes running in AWS

root@supervisor:/deploy# awsinfo
Cloud:  bdoss/us-east-1
Name                                Instance ID  Instance Type  Instance State  Private IP      Public IP      
----                                -----------  -------------  --------------  ----------      ---------      
nat.bdoss-dev                       i-c46a0b2a   m1.small       running         10.0.2.94       54.86.153.142  
bastion.bdoss-dev                   i-3faa69de   m1.small       running         10.0.0.207      None           
mesos-master.test.bdoss-dev         i-e80ddc09   m1.small       terminated      None            None           
mesos-slave.test.bdoss-dev          i-e00ddc01   m1.small       terminated      None            None           

awsinfo mesos-master – will display info about all mesos-master nodes running in AWS.

root@supervisor:/deploy/labs# awsinfo mesos-master
Cloud:  bdoss/us-east-1
Name                                Instance ID  Instance Type  Instance State  Private IP      Public IP      
----                                -----------  -------------  --------------  ----------      ---------      
mesos-master.test.bdoss-dev         i-e80ddc09   m1.small       terminated      None            None           

awsinfo 10.0.2 – match a private/public subnet

root@supervisor:/deploy/labs# awsinfo 10.0.2
Cloud:  bdoss/us-east-1
Name                                Instance ID  Instance Type  Instance State  Private IP      Public IP      
----                                -----------  -------------  --------------  ----------      ---------      
nat.bdoss-dev                       i-c46a0b2a   m1.small       running         10.0.2.94       54.86.153.142  
mesos-master.test.bdoss-dev         i-e96ebd08   m1.small       running         10.0.2.170      54.172.160.254 

Vagrant

If you can’t use Docker directly for some reason, there’s a Vagrant wrapper VM for it.

Before doing anything with Vagrant, complete the above steps for Docker, but don’t execute up.sh script

Just cd into vagrant directory and exec vagrant up, then vagrant ssh (nothing special here yet).

When you will exec vagrant ssh, docker container build process will spawn up immediately, so wait a bit and let it complete.

Now you’re inside a Docker container nested in Vagrant VM and can proceed with deployment in the same manner as it’s described for docker.

All the following vagrant ssh‘s will spawn Docker container almost immediately.

Once you are inside of the supervisor image, the minotaur.py script may be used to provision an environment and labs. The rest of this readme assumes that the script is executed from within the supervisor container.

Minotaur Commands

List Infrastructure Components

root@supervisor:/deploy# ./minotaur.py infrastructure list
Available deployments are: ['bastion', 'iampolicies', 'iamusertogroupadditions', 'nat', 'sns', 'subnet', 'vpc']

Print Infrastructure Component Usage

root@supervisor:/deploy# ./minotaur.py infrastructure deploy bastion -h
usage: minotaur.py infrastructure deploy bastion [-h] -e ENVIRONMENT -r REGION
                                                 -z AVAILABILITY_ZONE
                                                 [-i INSTANCE_TYPE]

optional arguments:
  -h, --help            show this help message and exit
  -e ENVIRONMENT, --environment ENVIRONMENT
                        CloudFormation environment to deploy to
  -r REGION, --region REGION
                        Geographic area to deploy to
  -z AVAILABILITY_ZONE, --availability-zone AVAILABILITY_ZONE
                        Isolated location to deploy to
  -i INSTANCE_TYPE, --instance-type INSTANCE_TYPE
                        AWS EC2 instance type to deploy

Deploy Infrastructure Component

In this example, the bdoss-dev bastion already existed, so the CloudFormation stack was updated with the current template.

root@supervisor:/deploy# ./minotaur.py infrastructure deploy bastion -e bdoss-dev -r us-east-1 -z -us-east-1a
Template successfully validated.
Updating existing 'bastion-bdoss-dev-us-east-1-us-east-1a' stack...
Stack updated.

List Labs

List all supported labs.

root@supervisor:/deploy# ./minotaur.py lab list
Available deployments are: ['clouderahadoop', 'gokafkaconsumer', 'gokafkaproducer', 'kafka', 'mesosmaster', 'mesosslave', 'zookeeper']

Print Lab Usage

Print the kafka lab usage.

root@supervisor:/deploy# ./minotaur.py lab deploy kafka -h
usage: minotaur.py lab deploy kafka [-h] -e ENVIRONMENT -d DEPLOYMENT -r
                                    REGION -z AVAILABILITY_ZONE [-n NUM_NODES]
                                    [-i INSTANCE_TYPE] [-v ZK_VERSION]
                                    [-k KAFKA_URL]

optional arguments:
  -h, --help            show this help message and exit
  -e ENVIRONMENT, --environment ENVIRONMENT
                        CloudFormation environment to deploy to
  -d DEPLOYMENT, --deployment DEPLOYMENT
                        Unique name for the deployment
  -r REGION, --region REGION
                        Geographic area to deploy to
  -z AVAILABILITY_ZONE, --availability-zone AVAILABILITY_ZONE
                        Isolated location to deploy to
  -n NUM_NODES, --num-nodes NUM_NODES
                        Number of instances to deploy
  -i INSTANCE_TYPE, --instance-type INSTANCE_TYPE
                        AWS EC2 instance type to deploy
  -v ZK_VERSION, --zk-version ZK_VERSION
                        The Zookeeper version to deploy
  -k KAFKA_URL, --kafka-url KAFKA_URL
                        The Kafka URL

Deploy Lab

Deploy a 3-broker Kafka cluster.

root@supervisor:/deploy# ./minotaur.py lab deploy kafka -e bdoss-dev -d kafka-example -r us-east-1 -z us-east-1a -n 3 -i m1.small    
Template successfully validated.
Creating new 'kafka-bdoss-dev-kafka-example-us-east-1-us-east-1a' stack...
Stack deployed.

/*******************************************
 Joe Stein
 Founder, Principal Consultant
 Big Data Open Source Security LLC
 Twitter: @allthingshadoop
********************************************/

		
  1. No comments yet.
  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 62 other followers

%d bloggers like this: