An Ansible Project Setup that Scales

So I have been using Ansible for over two years now .. I use it for damn near everything — provisioning infrastructure, configuring nodes, deploying Web applications, testing whatever I can, and other ad hoc tasks (sadly, I’m still working on a “get me beer” playbook) ..  Long story short, it’s been a game changer .. Problem is, as my (and my team’s — hi guys !!) Ansible usage grows (50+ playbooks and 130+ roles) — so does my desire to organize it in a way that is scaleable ..

There are many ways you can setup your Ansible project (one, two, three, four, etc ..) — which is great !! That said, I love me some simplicity .. After trying out a few setups, I finally settled on one that works for me .. You can see it here ..

The view from 30,000 feet:

  • inventories
    • All of my inventory files, static and dynamic, will be children of the inventories directory
      • FULL DISCLOSURE: by design, I only have two inventory files — localhost and .. every once in a blue moon we will add additional inventory files (PoCs, targeted testing, etc ..), but we always fallback to only the two
    • Group and other variables
      • the group_vars directory can be relative to an inventory file, no matter where the playbook is .. since I like to add some structure to my playbook organization, it makes sense to put it here
      • variable files (i.e. Ansible Vault) that need to be explicitly loaded go in the vars directory
      • IN THEORY: you could also put a host_vars directory here .. but as y’all know, host_vars are the devil
  • playbooks
    • Let me explain my logic here .. In my work, I essentially perform five functions: 1) provision infrastructure, 2) configure infrastructure, 3) deploy to infrastructure, 4) test infrastructure (and other things), and 5) save the world (i.e. ad hoc tasks) .. Knowing this:
      • construct/*
        • since the line between provisioning and configuring infrastructure can get murky, I combined the two “functions” into one directory
      • deploy/*
        • guess what goes here 🙂
      • test/*
      • adhoc/*
        • the one-offs that don’t fit nicely anywhere — but here
  • roles
    • resource/*
      • these are the bits and pieces that make up a useful piece of infrastructure .. for example, we have resource roles for AEM, Bitbucket, CPANm, .forward, Java (1.6, 1.7, and 1.8), Apache, Netcat, Nagios, NRPE, etc .. all solid implementations .. all reusable ..
      • STYLE ALERT: we DON’T use Ansible Galaxy — directly .. we often refer to it for inspiration — but always end up using our own implementation
    • <primary>
      • this is a “completed”piece of infrastructure .. the sum of the (resource) parts .. the cherry on top .. for example, we have primary roles for user_web_api_server, internal_dns_server, aem_author_server, aem_publish_server, etc ..
    • action/*
      • this is what I “do” to my infrastructure that doesn’t maintain any real “resource” …. for example, we have action roles for silencing/unsilencing Nagios, deploying code, restarting application servers, taking data centers offline, etc ..

The view from 30 feet:

  • PROBLEM: “I want my playbooks to be able to refer to my roles in a way that is consistent and easy” .. Great idea !! The problem is, your playbooks and roles are gonna be all over the place — so you can’t take advantage of the “relative referencing” you can do in a traditional project structure .. Let’s also assume you don’t want to define an absolute path in the config file ..
  • SOLUTION: Hello bootstrap !! Take a path that will always be consistent (inventory_dir) and use it to define other paths with some regex magic .. When finished, shove it all into a external variables file ..

  • FINALLY: Refer to the bootstrap, early and often, in every playbook you create,  ..

Again, this project setup works for me .. “Me” being myself and 4 other sys admins ..

Anywho — I need to get back to work on that “get me beer” playbook ..


Qik-n-EZ: Nagios Status and Acknowledgement Links via HipSaint

This post is kinda dopey, but it might help one person out there in intertubes land .. That said, I have been using HipSaint for ……………….. 3 years ?? It’s great !! It posts Nagios alert information into HipChat:

I then:

  1. Log into Nagios
  2. Locate the host or service that is alerting
  3. Click the link
  4. View the details
  5. Acknowledge the alert, if needed

Y’all know I’m lazy though –right ?? I wish the HipChat message would just give me the links I want .. Well, now it can:

As you will see, I am simply appending the status and acknowledgement links to the Nagios “service output” .. I also use Ansible to populate variables such as:

  • HipChat token
  • HipChat room
  • Nagios hostname

Now I can be as lazy as I want to be:


Make Everyone Talk to Hal !!

Hal is my Hubot chatbot .. He’s awesome !! He gets me beer !!

hal beer me

hal beer me

He also does things like restart app servers, deploy code, and show me pictures of grumpy cats .. He’s so cool, I’ve started making non-humans to talk to him .. “Greg, what do you mean ??” .. Well, let me show you ..


  1. I have a Nagios server
  2. It monitors (allthethings)
  3. When the “logged in users” alert is triggered, Nagios sends a message to my chat service using hipsaint
    1. “logged in users” is a monitor I have that alerts me when more than 3 users are logged into a server
  4. I see the alert and the server in question
  5. I SSH into the server
  6. I type who
  7. I then determine if I need to care
    1. If not, move on with my life
    2. If so, dig deeper

The thing is, I have more than 1,000 1,500 2,200 5,000 6,500 active monitors .. That means Nagios can and will send many, many messages to my chat service — depending on the day .. So how can I make my life easier ??

Here’s an easy one: ask Hal who’s on a server ..

hal whos on

hal whos on


hal whos on

hal whos on

My stack is HipChat -> Hubot -> Jenkins -> Ansible .. That means I can damn near do anything I want, all from my chat client ..

Remember what I said earlier — about making non-humans talk to Hal ?? What I did was create a Nagios event handler that sends a message to my chat service using HipChat CLI .. Therefore, I AM NOT asking Hal who’s on a server, it’s NAGIOS WHO IS doing it ..

nagios hal whos on

nagios hal whos on

It doesn’t stop there !! You can create scripted Splunk alerts as well .. Before you know it, you will be making (allthethings) talk to Hal ..

Qik-n-EZ: Multiline Notes for Nagios Service Definitions

OMG !! It’s still hip to say that — right ??

Anyway .. While I consider myself to be reasonably intelligent, I still find myself doing dopey things from time to time .. For example: saying OMG .. Another dopey thing I have been doing for a very, very long time — are run-on sentence notes for my Nagios service definitions .. I am ashamed to admit how many times I have tried to “fix” this — but always errored out during the pre-flight check .. Yes — I am aware of Google, but I’ve never found anything definitive ..

And then it hit me like a bolt of lightening — Bash style line breaks ..

So obvious, yet so elusive ..

Qik-n-EZ: Collect AEM Thread Dumps and Email via Ansible

So one of our AEM nodes was freaking out the other day .. No, not the election results .. Some code was deployed to it that had runaway processes, thus gorging itself on CPU and memory .. EEEKK !! What to do ?? If you’ve been around AEM for awhile, you know how we love our raw thread dumps .. That being said, I really dislike the process on how to obtain them:

  1. Log into the node
  2. Get the Java PID
  3. Execute jstack and output to a log file
    1. Repeat every 10 seconds for at least one minute
  4. Compress the log file
  5. Share the compressed log file
    1. Typically via email

So how can I automate this ?? Easy !!

Create a Bash script that will generate the thread dumps for you ..

Create an Ansible playbook that will execute the script, compress the log file, and email it


Ansible & ServerSpec, Sitting in a Tree, K-I-S-S-I-N-G

We’re not dumb .. We don’t live in a cave .. As infrastructure “developers” — we KNOW we should be writing tests to validate our “code” .. Buuuuttt — how many of us are actually doing this ??

FEAR NOT !! Because it’s really not that hard if you break it down .. Think about it:

  • I have a node
  • I “do things” to that node to make it “more awesome
  • I validate “those things”
  • I move on with my life

It’s the “I validate” bullet we need to focus on .. So then, here we go:

Pick a Testing Framework

Sooo many choices .. Here are some popular ones:

Get it on the Server

OK .. Here we can have a lengthy beer talk about WHAT framework is most supreme and HOW to best execute it (locally vs over-the-wire) — but it’s Friday — and my appetite for prolonged discussions about personal opinion is nil .. So let’s just agree to put ServerSpec on a node with the intention of running tests locally .. Here’s an example of a role that will do just that ..

Write Some Tests

We’re not splitting the atom here — we’re just writing some simple role tests that can validate our intended work .. I mean, look how easy this !!


Get Those Tests on the Server

In order to avoid repeating yourself, this is where you have to get a little creative, in order to avoid repeating yourself .. I need to find a way to get a role’s tests onto a node .. “Hey Greg, haven’t you heard of the synchronize module ??” Yeah, yeah — but I want to be DRY .. I want one solution that works for all roles ..

Eureka !! I’ve got it !!

I’ll create a role who’s sole purpose is to “copy” a referring role’s tests to the target node, and then declare this “copier role” as a dependency of the referring role ..

Run the Tests

OK .. So I lied a little .. You know that role who’s “sole purpose” is to “copy” a referring role’s tests to the target node — I added more to it .. What can I say — it was lonely .. The role will also double as a way to execute your ServerSpec tests, when you want it to ..

Wrapping it Up

Like I said — it’s beer Friday — so here’s the complete solution .. I hope this helps make automated testing “less scary” .. Before you know it, you will be sipping Mai Tais with Mizzy, discussing the effectiveness of unit vs acceptance tests ..

Deep Thoughts – Why No Automation Framework ??

Why don’t ops teams choose an “automation framework” for their “development” ?? Think about it .. How many Bash/Perl/Python/Ruby scripts does an ops team create and maintain in a year ?? 10 .. 100 .. 1000 ??

“But Greg, they do !! It’s called Chef/Puppet/Ansible/Salt/etc ..”

ZING !! Got me there .. But that’s not what I am talking about ..

Yes — ops teams use configuration management tools to MANAGE the scripts they create and maintain, but I’m talking about defining a framework in which ALL scripts would be developed in .. Good (even bad) dev teams do it all the time .. Come on, say the names with me — Spring, Struts, CodeIgniter, Django, Rails, Express, AngularJS, etc .. They’re everywhere !!

So what does that make “us” ?? Good ?? Bad ?? Lazy ?? Adaptable ?? Wouldn’t ops teams realize the same benefits as dev teams if they choose a framework for their development ?? That seems like a logical conclusion, yet, I don’t see the same requirement being demanded from ops teams .. Why is that ??

Sorry .. I don’t have answer for you — but the more I do ops, the more I am convinced there is an opportunity here ..