Troubleshooting in Ansible
In this post I am about troubleshooting of ansible playbooks. Ansible provides several ways of managing playbook. Some of them are: log configuration, defining some options such as --step, -v and defining running management in playbook using such modules as fail or assert
Log file configuration
Unfortunatelly, logging is not configured in Ansible by default. However, we can configure it.
There are two ways to activate logging:
- Set
log_pathparameter indefaultsection ofansible.cfgfile
cat ansible.cfg
[default]
log_path = /var/log/ansible-logs/ansible.log
It is recommended to use logrotate to manage the logs if they will be written in /var/log directory.
- set environment variable
$ANSIBLE_LOG_PATH
Managing problems
--syntax-check - checks the syntax in the playbook
ansible-playbook myplay.yml --syntax-check
--step - using this option we can step to a specific task in a playbook
ansible-playbook myplay.yml --step
--start-at-task - with this option we can start execution of a playbook from a specific task
ansible-playbook myplay.yml --start-at-task
We can debug execution of a playbook using -v parameter. There are 4 levels of verbosity: -v, --vv, --vvv, --vvvv
--check - option executes a playbook without making changes to the managed host. This option we can define by typing -C too.
Moreover, we can control --check option individually.
- This task will be alway run in checking mode:
- name: always run in check mode
shell: hostname
check_mode: yes
- This task will not run in check mode even if
--checkoption is defined in ansible-playbook command
- name: always run in check mode
shell: hostname
check_mode: no
--diff - option shows changes made to template files.
ansible-playbook myplay.yml --check --diff
Troubleshooting using modules
script module executes a script against managed hosts. This script have to locate on the control node. Executing fails if script ends with the non-zero code.
tasks:
- name: Run my script
script: run_my_script
uri module helps to check some http services
tasks:
- name: check my API
uri:
url: http://myapi.home.lab
return_content: yes
register: answer
- fail:
msg: 'Service is not up'
when: "'UP' not in answer.content"
stat - collects information about files and folders
tasks:
- name: Check if /opt/app/hello.py file exists
stat:
path: /opt/app/hello.py
register: answer
- name: hello.py file does not exist
fail:
when: answer.stat.exists
assert module we can use alternatively to the fail module. Here, in order to customize the message we can use either success_msg or fail_msg options.
tasks:
- name: Check if /opt/app/hello.py file exists
stat:
path: /opt/app/hello.py
register: answer
- name: hello.py file does not exist
assert:
that:
- not answer.stat.exists