DO410 Ansible and Ansible Tower training notes

April 6, 2020 Updated Dec 18, 2025 - Thomas Jungbauer - 5 min read (1052 words)

This article was written April 2020. Product versions, commands and screenshots may have changed since then. Please verify against the current documentation.

Notes taken during Red Hat course D410 Ansible and Ansible Tower.

Ansible installation

make sure that libselinux-python is installed
Ansible 2.7 requires python 2.6 or 3.5

yum list installed python

windows modules implemented in powershell
ansible requires at least .net 4.0

Configuration files

Ansible searches for ansible.cfg in the following order:

$ANSIBLE_CFG
ansible.cfg in the current directory
$HOME/ansible.cfg
/etc/ansible/ansible.cfg

whichever it finds first will be used.

use

ansible --version

to see which config file is currently used. you can view/dump/see what changed with

ansible-config [list|dump|view]

Default modules

List all available modules via

ansible-doc -l

For getting help on a specific module use

ansible-doc ping

Ad-hoc commmands

To display ansible output on a single line per host for easier readablility use the -o option

ansible all -m command -a /bin/hostname -o

Use the raw module for directly executing commands on remote systems that do not have python installed.

ansible -m raw

Custom Facts

Ansible uses custom facts from /etc/ansible/facts.d/. Facts can be stored in .ini style or you can place executable scripts in this directory. The script needs to output JSON. Custom facts are available via ansible_facts.ansible_local.

Magic variables available

hostvars: variables defined for this host
group_names: list of groups this host is a member of
groups: list of all groups and hosts in the inventory
inventory_hostname: host name of the current host as configured in the inventory

Matching hosts in the inventory

Some examples on how to match hosts defined in the inventory

'*.lab.com': match all hosts starting with lab.com
'lab,datacenter': match all hosts either in lab or datacenter
'datacenter*': match all host and host groups starting with datacenter
'lab,&datacenter': match hosts in the lab and datacenter group
'datacenter,!test.lab.com': match all hosts in datacenter, except test.lab.com

Dynamic inventory

Example scripts for dynamic inventories can be found at https://github.com/ansible/ansible/tree/devel/contrib/inventory.

You can use ansible-inventory to take a look a the current inventory as json. This also works for static inventories.

Inventories can be combined. Just create a directory containing a static inventory and script to create a dynamic inventory, ansible will happily execute the scripts and merge everything together.

Debugging

The following might be useful when debugging ansible roles and playbooks

ansible-playbook play.yml --syntax-check
ansible-playbook play.yml --step
ansible-playbook play.yml --start-at-task="start httpd service"
ansible-playbook --check play.yml
ansible-playbook --check --diff play.yml

Ansible Tower

Notes on deploying and working with ansible tower.

Installation

System requirements:

at least 4GB of RAM
actual requirement depends on forks variable
recommendation is 100MB memory for each for + 2GB of memory for tower services
20GB of disk storage, at least 10GB in /var

Steps for installing:

download setup tar.gz from http://releases.ansible.com/ansible-tower/setup/
set passwords in inventory
run ./setup.sh

Authentication

Authentication settings can be changed under Settings / Authentication. E.g for configuring Azure AD authentication we are going to need

an Azure AD oauth2 key and
a Azure AD oauth2 secret

RBAC

separate roles for organizations and inventories
you need to assign roles to organizations and inventories

The Tower Flow

These are the steps to run playbooks against managed nodes in Tower:

Create an organization if required
Create users
Create teams and assign users
Create credentials for accessing managed nodes
Assign credential to organization
Create credentials for accessing SCM repositories (e.g. git)
Assign credentials to users or teams
Create a project
Assign Teams to project
Create a job template for executing playbooks

Ansible Roles support

If the project includes a requirements.txt file in the roles/ folder, tower will automatically run

ansible-galaxy install -r roles/requirements.yml -p ./roles/ --force

at the end of an update. So this could be used to include external dependencies (like SAP ansible roles).

Job Templates

Ansible playbooks are stored in GIT repositories. A job template defines

the inventory used for this job template
the project for executing this job
this connects the GIT repository used in this project with the template
the playbook to execute
the credentials for executing jobs
permissions for users / teams (e.g. admin, execute)

Tower creates jobs from those templates, which are ansible runs executed against managed nodes.

Fact Caching

It might be a good idea to use the tower facts cache. To speed up playbook runs set gather_facts: no in the play. Then enable the facts cache in tower.

In tower settings set a timeout for the cache
In job templates enable Use facts cache
Create a playbook that runs on a regular basis to gather facts, e.g.

- name: Refresh fact cache
  hosts: all
  gather_facts: yes

Inventory options

These are the options for creating inventories in Ansible Tower

static inventory defined in tower
importing static inventories via awx-manage
static inventory defined in git repository
dynamic inventory via a custom script
dynamic inventory provides by tower (e.g. satellite)

A special feature in Tower are so called smart inventories. A smart inventory combines all static and dynamic inventories and allows filtering based on facts. Filtering requires a valid fact cache.

Troubleshooting

Tower uses the following components:

postgresql
nginx
memcached
rabbitmq
supervisord

Useful tools

ansible-tower-service (e.g. status / restart)
supervisorctl (e.g. status)
awx-manage

Tower stores log files in

/var/log/tower/ (e.g. tower.log).
/var/log/supervisor/
/var/log/nginx/

Other important directories

/var/lib/awx/public/static static files served by django
/var/lib/awx/projects stores all project related files e.g. git checkouts)
/var/lib/awx/jobs_status job status output

by default playbook runs are confined to /tmp this might lead to problems with tasks running on the local system.

In case of a lost admin password you can use awx-manage to reset the password or create a new superuser:

awx-manage changepassword admin
awx-manage createsuperuser

Replacing the default TLS certificates

Ansible tower uses nginx to service it’s web interface over TLS. Nginx uses the configuration file /etc/nginx/nginx.conf.

To deploy custom TLS certificates used by tower replace the certificate and private key in /etc/tower. You have to replace

/etc/tower/tower.crt and
/etc/tower/tower.key

It might be a good idea to create a backup copy before overwriting those files.

Backup and restore

Of course backup and restore are done via ansible. The ansible tower setup script setup.sh provides a wrapper around these playbooks. Execute

setup.sh -b

to perform a backup. This creates a backup .tar.gz file in the current directory.

To restore a backup use

setup.sh -r

this restores the latest backup per default.

Things to remember

Workflow job templates
add autocmd FileType yaml setlocal ai ts=2 sw=2 et to .vimrc
use sudo yum install python-cryptography if there are many vault files to speed up ansible