Ansible Playbook for System Updates

So lately I’ve been playing a lot with ansible. It makes administrating close to 20 servers a real breeze! Come Patch Tuesday, you never know if reboots will be required on a server. I took a bit of inspiration from this ansible script and tweaked it a bit to reboot the servers after updates if required… as well as performing an apt-get autoremove to clear out old packages (usually kernels) so my /boot partition doesn’t get full and wreak havoc.

- hosts: all
  sudo: yes
  tasks:
  - name: Update package list and perform safe-upgrade
    apt: upgrade=safe update_cache=yes

  - name: Check if packages need to be autoremoved
    command: apt-get --dry-run autoremove
    register: check_autoremove
    changed_when: False

  - name: Autoremove unused packages
    command: apt-get -y autoremove
    when: "'packages will be REMOVED' in check_autoremove.stdout"

  - name: Check if packages need to be autocleaned
    command: apt-get --dry-run autoclean
    register: check_autoclean
    changed_when: False

  - name: Clean up package cache
    command: apt-get -y autoclean
    when: "'Del' in check_autoclean.stdout"

  - name: Check reboot if required
    stat: path=/var/run/reboot-required get_md5=no
    register: sym

  - name: Reboot if required
    shell: reboot
    when: sym.stat.exists is defined and sym.stat.exists
    async: 0
    poll: 0
    ignore_errors: true

  - name: Wait for server to come back
    local_action: wait_for host={{ inventory_hostname }} state=started
    when: sym.stat.exists is defined and sym.stat.exists
    sudo: false

All that’s left is to run ansible-playbook -K doSystemUpdates.yml (or whatever you saved it as) and twiddle your thumbs!

This worked great on all of our Ubuntu production servers the other night. Hopefully someone finds this helpful!