Did you know you can do more than just “OR”-ing groups and hosts with the Ansible --limit flag?

# Target hosts that are both in group_a AND in group_b
ansible --limit "group_a,&group_b"

# Target hosts that are in group_a, but NOT in group_b
ansible --limit "group_a,!group_b"

But there’s more, like regex patterns and reading hosts from a separate file!

Regex patterns

Target hosts and groups using Regex pattern, matches group_aa, group_ac, group_ca, group_cc (must prefix with ~ to get processed as regex)

ansible --limit '~^group_[ac]{2}$'

Note that it matches any subsets by default, so the following would match group_a, group_b, etc:

ansible --limit '~up_'

Read hosts from file

Target hosts and groups in file (only works on ansible-playbook command, not on ansible command):

echo '
host_1
host_2
!group_c
~group_[ab]
# my comment
' > limits.txt

ansible-playbook --limit @limits.txt

This allows you to write a subset of hosts or groups to a file, without having to modify the inventory file.

Inside this file, you can still make use of the & and ! operators, as well as the ~ for regex patterns.

You can kind of use comments by adding # symbol, as you won’t have any hosts or groups in your inventory that starts the name with #. It’s harmless if it can’t match, and will log something like:

[WARNING]: Could not match supplied host pattern, ignoring: #my comment

Retries

There’s also retry files that Ansible can generate for you when a host fails in a playbook. E.g:

# Must be "True". Will not activate if set to "true" or "TRUE"
export ANSIBLE_RETRY_FILES_ENABLED=True

ansible-playbook playbooks/my_playbook.yml

Logs output:

PLAY [all] *************************************************************************

TASK [do something] ****************************************************************
ok: [host_1]
ok: [host_2]
ok: [host_3]
ok: [host_4]
fatal: [host_5]: FAILED! => {"msg": "error oh no"}
ok: [host_6]
	to retry, use: --limit @/home/yourusername/some/path/to/playbooks/my_playbook.retry

PLAY RECAP *************************************************************************
host_1    : ok=2    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0
host_2    : ok=2    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0
host_3    : ok=2    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0
host_4    : ok=2    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0
host_5    : ok=1    changed=0    unreachable=0    failed=1    skipped=0    rescued=0    ignored=0
host_6    : ok=2    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

Ansible created a new file, named my_playbook.retry in this case, right next to my playbook file, only containing:

host_5

Then if I want to rerun the playbook with only failed tasks, I run:

ansible-playbook playbooks/my_playbook.yml --limit @playbooks/my_playbook.retry

Read more

Documentation on --limit patterns: https://docs.ansible.com/ansible/latest/inventory_guide/intro_patterns.html#advanced-pattern-options

Documentation on retry files: https://docs.ansible.com/ansible/latest/inventory_guide/intro_patterns.html#patterns-and-ansible-playbook-flags

Config reference on RETRY_FILES_ENABLED: https://docs.ansible.com/ansible/latest/reference_appendices/config.html#retry-files-enabled

---

(Cover photo by Engin Akyurt)