How to speed up Ansible a lot

README.md

Introduction

I'm working with Ansible and many of servers every day, so let me show you, how can waste less of your time! :)

What Ansible do? 

• Connects, runs, check, sends back, processes it, runs, checks, sends back and voila!
• Connects... again... and again... and again...

What you can do to fix on node/remote server

Edit sshd config and disable UseDNS, it's mainly used only for logging and authentication. But.. only when you have set IgnoreRhosts in your config. So who doesn't know, should use it with clean mind. Because this has truly so far with security. :)

/etc/ssh/sshd_config

UseDNS no

Configure your ssh

Create needed directory:

mkdir ~/.ssh/sockets

~/.ssh/config

 Host *
  ControlMaster auto
  ControlPath ~/.ssh/sockets/%r@%h-%p
  ControlPersist 600
  PreferredAuthentications=publickey

Solve it on Ansbile side

Every additional connection will cost a lot of additional time.. :) The main feature in latest releases of Ansible is pipelining. Try to use this configuration in ansible.cfg.

Follow Ansible documentation (put it into ansbile.cfg)

• ANSIBLE_CONFIG (an environment variable)
• ansible.cfg (in the current directory)
• ~/.ansible.cfg (in the home directory)
• /etc/ansible/ansible.cfg

ansible.cfg

[paramiko_connection]
record_host_keys=False

[ssh_connection]
# Tuned ssh args
ssh_args=-C -o ForwardAgent=yes -o StrictHostKeyChecking=no -o ControlMaster=auto -o ControlPersist=600s -o PreferredAuthentications=publickey
control_path = /tmp/ansible-ssh-%%h-%%p-%%r

# By enabling pipelining you get realy better performance 
pipelining = True

# And the best what you can do if you have too many nodes... increase forks to 50 or 100+
forks = 50