How to propagate changes to previously-applied tasks (e.g removing files, changing crons, etc)

Archives Ansible Project
1

Hello,

I have really enjoyed using Ansible to be able to completely template new servers that I set up, but one thing I have not yet figured out how to do in my workflow is deal with changes that occur after a playbook has been run on a server. For example, let’s say that I have a cron task that creates a cron:

  • name: create my cron
    cron: name=“My Cron version 1.0” job=/usr/local/bin/cron.sh minute=1

I run the playbook on my servers and this cron is added to the crontab, good. However, later I decide to change the name, the job, or some other attribute. If I do that, Ansible will add a new entry to the crontab but leave this old one around. One solution would be to add a state=absent task, but that involves duplicating or adding some legacy code to my playbook:

  • name: create my cron (old)
    cron: name=“My Cron version 1.0” job=/usr/local/bin/cron.sh minute=1 state=absent
  • name: create my cron
    cron: name=“My Cron version 1.1” job=/usr/local/bin/cron.sh minute=1

The best way to handle this that I can think of is:

  1. review changes made to playbook, determine if they change any pre-existing files, crons, etc
  2. if changes to pre-existing files are made, duplicate corresponding tasks and add a state=absent entry with the old names, run playbook to remove old data
  3. remove duplicate tasks, commit changes to playbooks repository

Is there a better way for ansible to handle removing old data from a previous version of a playbook?

Thanks,

Andrew

2

not really, since you could have 10 diff plays for adding cron,
ansible won't know about the other 9 when running 1, it is up to you
too keep track. I just normally create 1 off plays and put them in a
cleanup/ dir that i can rerun in case i suspect something was missed
or someone reran an old version.

3

Brian,

Thanks for the clarification.

Andrew