Skip to content

Latest commit

 

History

History
99 lines (77 loc) · 3.48 KB

README.adoc

File metadata and controls

99 lines (77 loc) · 3.48 KB

Overview

Bad Ansible is a deliberately poorly constructed ansible project designed to give students a basic project to cleanup, refactor, and improve. It represents a newcomer’s, aka newbie, approach to hacking together a functional playbook to deploy a three tier app.

The entry point, main.yml contains multiple good enough plays, and some redundant ones.

Goal of Bad Ansible

Set up an application that:

  • Sets up a frontend1 haproxy server which load balances requests to http://frontend1.<GUID>.example.opentlc.com.

    • Requests to frontend1 come in on port 80

    • Load balance requests to servers in the group apps, listening on port 8080

  • Set up one or more tomcat servers, the group apps, to:

    • Serve the included HTML content from ` /usr/share/tomcat/webapps/ROOT/index.html`

  • Install postgres on all servers in the group appdbs

Nothing else matters i.e. you can get rid of it!

Tips

  • Make sure you have read the Goals section above and understand the desired architecture

  • There are redundant plays within main.yml

  • You should be able to run your solution multiple times, not just once

  • You are free to restructure your solution in any way

Note
Any plays or tasks not related to the above are therefore redundant and can be removed.
Tip
Don’t rename any of the ansible groups: frontends, apps, appdbs - later labs require these!

Usage

  1. Set up the environmental variable GUID to your GUID

    export GUID=abcd
  2. Edited your ansible.cfg depending if you are executing on bastion or remotely

    • Pay attention to the value of ssh_args

  3. Configure your chosen ssh config file

    • Set the correct user

    • Set the correct key(s)

    • If necessary, e.g. executing from a remote control node, install the right key(s)

      Note
      You can easily test your chosen ssh config file e.g. ssh -F ssh.cfg app1 etc
  4. Execute your main wrapping playbook

 ansible-playbook main.yml

Structure

.
├── README.adoc                       (1)
├── ansible.cfg                       (2)
├── cleanup.yml                       (3)
├── haproxy.cfg.j2                    (4)
├── hosts                             (5)
├── index.html.app1                   (6)
├── index.html.app2                   (6)
├── index.html.j2                     (6)
├── main.yml                          (7)
├── ssh-bastion.cfg                   (8)
└── ssh-laptop.cfg                    (9)
  1. This README file

  2. Simple ansible.cfg

  3. Convenience playbook that removes the 3tier app - use to retest your work

  4. A badly written jinja2 template for HAProxy - HINTS

    • This should loop over the apps group

    • Perhaps the Tomcat port should be a variable?

  5. inventory file - referenced from ansible.cfg

    • Note the use of a lookup in line 2, if you prefer consider:

    • Hard coding it GUID=abcd or passing it in via -e GUID=abcd.

  6. Too many index.html files?

  7. Main entry point

  8. ssh configuration file for working off a bastion host

  9. ssh configuration file for working off a laptop or remote control node

Potential strategies

  • Divide and Conquer

    1. Break main.yml into individual plays

    2. Discard the redundant ones

    3. Wrap in a main.yml or similar using import_playbooks:

  • Role Play

    1. Identify the key roles e.g. common, frontends, apps, database etc

    2. Convert the relevant plays into roles

    3. Create wrapping playbook with multiple plays invoking roles on groups