Does anyone have any useful or clever Ansible tricks that you routinely use in the course of your normal daily activities?
In training, I recommend the following in ~/.vimrc:
autocmd FileType yaml setlocal ai ts=2 sw=2 et nu colorcolumn=1,3,5,7,9,11,13
But use a good editor! (Atom,. Visual Studio Code, Sublime, etc)
Also, I keep my playbooks in separate directories (see screenshot), and every playbook is called playbook.yml in order to take better advantage of searching through my command history when using ansible-playbook playbook.yml and ansible-playbook --syntax-check playbook.yml
I always encourage my students to "tell the story of your playbook" i.e. articulate their playbook goals in point form. Those then become the names of your tasks.
In response to a VT class request, I created https://gitlab.com/rgdacosta/sync-work which uploads their class created playbooks to the cloud, so that they could download them from their physical machines. I opened vim, and told the story of my playbook in front of them, then noted the modules which I thought would best satisfy my task goals, then figured out the specifics offline. They enjoyed that!
Parametize your playbooks - keep them as dynamic as you can.
When creating your playbook consider creating the associated undo or rollback tasks and tag them accordingly.
Have a style guide that specifies how you do things, from using variables, to how you declare booleans, to the use of includes. Have guidelines as to how playbooks ought to be written.
Always keep your playbooks under version control!
And that leads me to my last point, always give examples which are not covered in the courseware - add value. Exposure to scenarios is important - add value!
I notice you have a different ansible.cfg for each of your directories. Do you find those settings change a lot from playbook to playbook? Right now, I have a main ansible.cfg that I use for all my tests (but I'm still in the experimental phase).
I'm a fan of this approach as well. Each of my roles as an "ansible.cfg" in the test directory [I keep each role in a separate code repository], and each of my playbook projects [again, separate code repository] contains one or more ansible.cfg Depending on the environment, the ansible.cfg can change quite a bit [Connection properties to Dev vs Prod environmments, etc.] I find that it is useful to have ansible.cfg local to my project. I can see where sharing a common ansible.cfg would be useful - YOu then have to worry about how to "share" that shared resource. I've done that in the past using symlinks, but I haven't been to thrilled about the "elegance" of the solution.
@ricardodacosta: Another vim option that's been suggested by one of the course instructors (I forget who, my apologies) is to use cursorcolumn (cuc) instead of colorcolumn. The cursorcolumn option will highlight the column your cursor is currently on, and may be a bit less cluttered approach. But yeah, no matter what it helps to have a solid text editor that you're comfortable with and that helps you with syntax and indentation.
For finding out what might run (just statically):
ansible-playbook --list-tasks production.yml --tags bareos,workstation
ansible-playbook --list-tasks production.yml --skip-tags nagios
ansible-playbook production.yml --list-tasks --limit windows
For better understand what is in inventory:
For better understanding how variables from host_vars, group_vars
ansible-inventory --graph --vars
Make lot's of sens for bigger projects. But it doesn't know anything about roles.
These are great tips - I think the "ansible-playbook" command is often under utilized. It has a rich set of options that allow a lot of different interactions with a particular playbook.
For clarification, what did you mean by "it doesn't know anything about roles"? Is that refering to the ansible-playbook --list-tasks option?