How to Deploy Spark Instance Using Ansible 2.0 in FIWARE Lab

How to deploy Spark HA instance using
Ansible 2.0 in FIWARE Lab
Fernando López Aguilar
Technological Expert. IT and Cloud Architect
fernando.lopezaguilar@telefonica.com
@flopezaguilar

2
Cookies of the Dark Side
• Use OpenStack locally.
• Use Ansible.
• Use both together.

3
Come to the Dark Side
We use OpenStackClient

4
OpenStack client
 Unified shell command structure and a common language to describe
operations in OpenStack.
 Remote interaction with any OpenStack environment (if it is open).
 Easy install, just execute: pip install python-openstackclient.
 Easy to use, just execute: $ openstack --help

ubuntu@ubuntu:~/malaga$ sudo apt install python-pip
…
ubuntu@ubuntu:~/malaga$ sudo apt install virtualenv
…
ubuntu@ubuntu:~/malaga$ virtualenv env
New python executable in
/home/ubuntu/malaga/env/bin/python
Installing setuptools, pip, wheel...done.
ubuntu@ubuntu:~/malaga$ source env/bin/activate
(env) ubuntu@ubuntu:~/malaga$ pip install
> python-openstackclient
…
5
OpenStackClient hands-on
 Create a new directory
(e.g. malaga)
 If you do not have, install
pip and virtualenv.
 Create your virtualenv
(virtualenv is a tool to
create isolated Python
environments).
 Activate the isolated
python environment.
 Install the
OpenStackClient

6
OpenStack client hands-on
 After the installation you can execute the command shell openstack.
 To obtain information about the openstack command and its
subcommands, run it:
$ openstack help [subcommand]
$ openstack --help
 Typing openstack switches to the interactive mode.
Typing help [subcommand] for information.
To exit the interactive mode, type quit.

7
OpenStack client hands-on
• Next steps: we need the
credentials of your OpenStack
environment.
• Just login into FIWARE Lab Cloud
Portal (https://cloud.lab.fiware.org/)
and get your data:

ubuntu@ubuntu:~/malaga$ more
fernando.lopezaguilar@telefonica.com-openrc
export
OS_USERNAME=fernando.lopezaguilar@telefonica.com
export OS_PASSWORD=
export OS_TENANT_NAME="Fernando Lopez cloud”
export OS_REGION_NAME=Spain2
export OS_AUTH_URL=http://130.206.84.8:4730/v3/
8
 File downloaded a file with
format:
<your user name>-openrc
 Content:
 OS_USERNAME
 OS_PASSWORD
 OS_TENANT_NAME
 OS_REGION_NAME
 OS_AUTH_URL

export OS_PROJECT_DOMAIN_NAME=default
export OS_USER_DOMAIN_NAME=default
export OS_IDENTITY_API_VERSION=3
export PS1='(`basename "$VIRTUAL_ENV`)[u@FIWARE
Lab W(keystone_user)]$ ’
ubuntu@ubuntu:~/malaga$ source
fernando.lopezaguilar@telefonica.com-openrc
9
 You need to edit the file
and add the following.
 I suggest to add also the
following line.
 Move the file to your work
directory and load it.

10
 Your turn…
 Get help about openstack client
 Create a keypair.
 List of keypairs.

11
THEY HAVE ANIBLE… !!!
I’ve been to the dark side…

12
Ansible
 Ansible is an IT automation tool.
 Manages machines in an agent-less manner.
 Uses of OpenSSH for transport and a YAML language.
 Main focus in software deployment and system configuration.

13
What do you need to know?
 Connect to a remote machine using SSH.
 Interact with the bash command-line shell.
 Install packages.
 Use the sudo command.
 Check and set file permissions.
 Start and stop services.
 Set environment variables.
 Write scripts (any language).

14
Previous action
 Install ansible (v 2.2.0.0)
 Using openstackclient, deploy a server.
 Create a keypair
 Create security group an assign rule
 Create a server

15
First steps with Ansible
 Define inventory file (hosts).
 INI-like format and looks like this::
 <server name>: name of the server
 ansible_ssh_host: IP of the server
 ansible_ssh_user: user to access via ssh
 ansible_ssh_private_key_file: key pair to access the server

ubuntu@ubuntu:~/malaga$ ansible ansible-test -i hosts -m
ping
ansible-test | SUCCESS => {
"changed": false,
"ping": "pong"
}
16
First steps with Ansible
 Connect to the server
named <server name>
and invoke the ping
module.

---
invoice: 12345
product:
- id: 987
amount: 2
description: basketball ball
- id: 760
amount: 1
description: football ball
18
Ansible
 Process description in YAML
description file.
 YAML is a human friendly data
serialization standard for all
programming languages.
 A deployment is described in a
“playbook” (e.g.: deploy a web
application).
 A “playbook” can define several
processes o deployments

19
Ansible inventory
 Ansible works against multiple systems in your infrastructure at the
same time.
 By defaults, saved in the location /etc/ansible/hosts but can be
specified a different inventory with –i command.
 Not only is this inventory configurable, but you can also use multiple
inventory files at the same time.
 You can use a script to generate dynamically the inventory.

20
Using Variables: About Jinja2
 Ansible allows you to reference variables in your playbooks using the
Jinja2 templating system.
 Example: in a simple template, you can do something like:
My amp goes to {{ max_amp_value }}
 This is also valid directly in playbooks:
template: src=foo.cfg.j2 dest={{ remote_install_path }}/foo.cfg
 Variables might be defined in YAML file (e.g. vars/main.yml)

21
Ansible modules
 Ansible ships with a number of modules (called the ‘module library’).
 They can be executed directly (-m option) on remote hosts or through
Playbooks.
 Users can also write their own modules.
 These modules can control system resources, like services,
packages, or files (anything really), or handle executing system
commands.

22
Ansible roles
 Instead of creating giant playbooks with hundreds of tasks we can use
roles to organize tasks.
 A role breaks task into smaller more discrete units of work.
 A role is all the tasks, variables and handlers needed to complete the
unit of work.
 This allows a role to be completely self contained or encapsulated
and completely reusable.
 Example of role is the installation and configuration of NTPd service.

24
Where the hell is Spark…?

25
Spark
 Apache Spark is a fast and general-purpose cluster computing
system.
 Provide high-level APIs in Java, Scala, Python and R.
 Support a rich set of higher-level tools:
 Spark SQL for SQL and structured data processing.
 MLlib for machine learning.
 GraphX for graph processing, and Spark Streaming.

26
Spark in HA
 Creation of a Master node and several Slaves nodes.
 Configuration of the instances.
 Configuration of Spark nodes.
…

27
Hands-on Spark deployment with Ansible
 Clone the github repository
https://github.com/flopezag/ansible_spark_openstack
 Follow the instructions in README.md file
 Take a look into /vars/main.yml file.

28
Yeah, if you could just,
come to the dark side…

29
References
 OpenStack Client Command List:
http://docs.openstack.org/developer/python-openstackclient/command-
list.html
 Ansible documentation:
http://docs.ansible.com/ansible/
 FIWARE Lab Account portal:
https://account.lab.fiware.org/idm/

Thank you!
http://fiware.org
Follow @FIWARE on Twitter
30

Thank you!
http://fiware.org
Follow @FIWARE on Twitter

How to Deploy Spark Instance Using Ansible 2.0 in FIWARE Lab

More Related Content

What's hot (19)

Viewers also liked (6)

Similar to How to Deploy Spark Instance Using Ansible 2.0 in FIWARE Lab (20)

More from FIWARE (20)

Recently uploaded (20)

How to Deploy Spark Instance Using Ansible 2.0 in FIWARE Lab

Editor's Notes