10 on Tech Episode 004: Ansible with Jonathan Frappier

Share with your friends


Today’s show features friend of the IT community Jonathan Frappier to discuss Ansible (and DevOps in general) and how leveraging tools like configuration management solutions and orchestration engines changes the way that teams work. As always, thanks for listening, and don’t forget that the transcript is available below the player if you’d rather read than listen!

And please be sure to subscribe to the show on iTunes, Google Play, or Stitcher!

In this show, we discussed:

Show Transcript
James: Hey, everybody, it’s James Green, welcome to 10 on Tech. It’s produced by ActualTech Media. Go over to www.actualtechmedia.com to check us out. We produce content for IT administrators, architects, and CIOs all around the world. Keep an eye out for our MegaCast events, and for our webinars and papers and that kind of thing. We’re also blogging over at ActualTech.io, so check it out.
  I’m fortunate to be joined today by Jonathan Frappier, a good friend of the IT community and of mine. I’ve been, Jon, listening to you for a while, especially about DevOps-y kind of stuff. You’ve been talking about DevOps automation, orchestration, configuration management for longer than a lot of people that I know.
  I’ve been listening to what you have to say about that for quite a while and I wanted to have you on the show to talk about what you’re interested in right now. I know that some of that happens to revolve around Ansible. I’ve been seeing a lot from you about Ansible, so I want to talk to you about that a little bit.
  First of all, welcome to the show. Tell people where we can find you online, how we can read your stuff, interact with you.
Jon: Sure. You can hit me up on Twitter. It’s where I’m most active, @jfrappier. You can check out my blog at www.virtxpert.com.
James: Cool. Thank you for being on. I want to start by asking you … Tell me a little bit about from two years ago, or however long it’s been, when you really started talking about DevOps. I think at the time it was a lot about culture and ideas. Over the last two years, it’s become a lot more of a reality. How does that look different now than it did then when you started talking about it? What sort of realities are you seeing when you talk to people about the way the DevOps culture and processes are changing organizations?
Jon: Well, it’s not changed much other than, I think more people are recognizing it now. I’d like to think back to my early days where I was a desktop admin or a windows server admin and I saw the shift going to virtualized environments versus all physical.
  I’ve been watching the technology community. I’ve been watching technology change and seeing this shift from unmanageable workloads, doing everything manually. It was just not sustainable, so something had to change. DevOps came along and I think more people are recognizing now that we need to change the work culture in order for businesses to succeed. It’s still, I think, really early on. I talk to people at VMUGs and other local events and they’re still, just now started to talk about it internally.
James: Yeah. That’s the big thing that a lot of DevOps pundits, for lack of a better word, are saying to people who want to know more about DevOps, is “Always remember, we’re talking about processes here. We’re talking about culture here. We’re not talking about, necessarily, a tool that you run or a program that you decide to follow. It’s a mind set. It’s a culture.”
  However, there are tools that make a big difference in the way that you do these kinds of things. One of them, I should say a category of them, is configuration management and orchestration. There’s a handful of them out there that are really good. I’ve tried a number of them and I know you have as well.
  In looking at those, I know that you have decided for your purposes, there’s a lot of value found in Ansible. Can you tell me just a little bit about Ansible? What does it have that you like? What are you doing with it? A high level, what’s it for?
Jon: Ansible allows you to define the state of some piece of technology. They’ve really come a long way since I first looked at them, maybe, it’s going to be three years ago now. They’re up to version 2. It’s a little bit different mindset than say, for VM or folks that are maybe used to PowerCLI or Windows folks that are used to PowerShell, and building these scripts that will make folders and set permissions.
  We’re not doing that in Ansible and in most configuration management tools. What we’re doing is we want to define the state of the object we want to be in.
  I say object, because Ansible supports not only Windows or Linux operating systems. I believe it was 2.1, they introduced support for a wide range of network devices, so things like switches and firewalls. You can actually define your configuration in Ansible and make sure that, that is applied to those devices on your network.
James: We are not talking about, probably, 100% of the infrastructure here, but a pretty large portion of all the different components that make up an IT infrastructure. Ansible is a tool that we can use to define their state. I’ve heard the phrase “declarative state configuration” used with tools like this.
  The difference between, like you were saying, with something like PowerCLI is we tell it to go do exactly this. I know what end I’m trying to get to. With something like Ansible, you’re going to say, “I want it to look like this.” Then it’s Ansible’s job to go do whatever it needs to do, to make sure that it looks like that.
  The cool thing you’re saying about Ansible is, “It does that for lots of different parts of your infrastructure.”
Jon: Absolutely. Windows, Linux, there’s a lot of different community contributed modules. They’re called Extras Modules if you’re looking through the Ansible documentation. Which is actually really well written for something that spans so many devices and types of operating systems. They’ve really … It covers so much of that … You’re not far off if you’re following the Ansible docs. There may be one or two steps missing, ’cause you can’t get every use case, but so much of it’s in there.
  You can manage databases. You can manage Apache or iAS patches, restarting machines, making sure service levels are started. All of that’s in there. Defined network ports … How do you want … You want the network port up or down? Do you want it at 10 gig or 1 gig? You have control over that across your entire, most of, your infrastructure.
James: That sounds a little bit daunting, just the scope of everything that you can control. I’m sure when you started, you looked at a particular area and said, “I can start here.” How do you recommend that people approach moving into a configuration management sort of mentality rather than all of the one offs. They way that we’ve been used to doing things. Where’s a good place to start? Is there one?
Jon: Don’t start where I did, which was trying to change the world overnight. That’s a easy way to get frustrated, I think, with anything. With Ansible, you can do very simple things.
  You’d install Ansible on some Linux virtual machine or host and run that from there. It uses what it calls playbooks or roles to do that configuration definition. It can be as simple as one thing.
  For example, there may be a service that you need to restart often. Maybe you’ve got a buggy application, or you just want to ensure that, that service is always running on a machine. You could write that one task in a playbook and run that as needed. You could set it to run as a chron job, so that it’s running every hour or 30 minutes, whatever it happens to be. It will ensure that … For example, I was testing with Windows earlier and anyone that follows me on Twitter saw that, “Don’t use DNS as your test service, if you’re going to set it to stop and you’ve only got one DNS server in your lab, because then nothing will work anymore.”
  Lesson learned there, don’t do that. Obviously, you want to [crosstalk 00:08:38]
James: Lab, right?
Jon: Yeah, lab … One DNS server doesn’t happen often, but of course, I nail it on the head. I was like, “Oh, yeah, DNS that’s important. Let me make sure I can stop and start that. Oh, yeah, it stopped and now I can’t resolve anything. Awesome.”
  Pick a thing. Go through your help desk ticketing system. Go through your bug tracking system and find a thing that you’re doing often or that’s really important to do. Define what state you want that to be in, whether it’s a service, whether there are file permissions or a configuration file that you want to be present, and start there.
  Get some small wins and then other people on your teams may see how easy it is to do that. That will allow more collaboration, more communication, some of those actual DevOps-y things that we want to do. Get people talking together to take a look at what you’re doing and how it works and then you can start slowly changing the world versus trying to just do it overnight.
James: Sure. I know that the documentation is very good so that’s probably one good place to start exploring and see what you can do. What other resources would you recommend people check out to start digging into Ansible?
Jon: They also have a community site, AnsibleGalaxy, Galaxy.Ansible.com, where people can contribute roles and playbooks that they’ve written to do with things. If you’re trying to get a webserver set up, you don’t necessarily need to do that from scratch. You can go and find roles that other people have published.
James: Awesome. How ’bout you, where can people go to read more about what you’re writing about Ansible? Connect with you on twitter @jfrappier. Do you publish anywhere else? I know you’re at the VMUGs sometimes. What other events are you going to be at?
Jon: Yes, so you can find me at most of the New England area events, so Boston VMUG for one, the New England VTUG events, especially the fall and winter warmer ones. We hold those at Gillette Stadium. Sometimes we get to spy on the Patriots playing. You’ll find me there. Definitely hit me up on Twitter. I love chatting with people and figuring out things that they might have going on, to help them out or just keep learning myself.
James: Awesome. Thanks for being on the show, I appreciate it.
Jon: Awesome. Thank you very much.
James Green

James is a Partner at ActualTech Media and writes, speaks, and consults on Enterprise IT. He has worked in the IT industry as an administrator, architect, and consultant, and has also published numerous articles, whitepapers, and books. James is a 2014 - 2016 vExpert and VCAP-DCD/DCA. Follow James on Twitter

No Comments

Post A Comment

Web Analytics