 So this is considered a beginner's track, what the heck is triple O, which is the short version of O, O, O, or open stack on open stack. But we have an intimidating number of triple O developers here in the audience. So if you guys have a question that is off of my radar or deeper dive than this talk covers, we may be able to answer it anyway with the interactive help of the audience. I'll get to you owls all the way down in a minute. So triple O is a project aimed at installing, upgrading, operating open stack clouds, using open stack facilities as a foundation. It builds on Nova, ironic neutron and heat to automate cloud management at a large scale. And before we get started, much more, I am introducing a level of interactivity that psychologists say makes the talk more interesting. So feel free to tweet at any or all of these names and hashtags. So who am I? I'm a developer evangelist. I work within the RDO community, but also the triple O community. And I will reference RDO quite a bit, even though it's not in the title. So I'll introduce you very briefly to what RDO is. I work at Red Hat. I've been an ATC since last year. So I consider myself a baby baby. I'm Leander Thaw on Free Note, and that really is my email address. And yes, I am putting it out in public. So don't spam me, please. But that's how passionately I feel about making triple O accessible to developers and users and anyone else who wants to experiment with this particular installer. If you run into any problems, I really do want to get an email from you saying, hey, I tried to get on IRC and I didn't get the answers I needed. I joined the mailing list and I didn't get the answers. I'm feeling overwhelmed, and please help. Email me directly. Yes, I'm serious. I have glasses. So what is OpenStack? And since we're at OpenStack Summit, that will be a very quick couple of slides. RDO is not actually Rain's Dynamic OpenStack as much as I would like to say it is. And then triple O and the triple O QuickStart and what those are. I know everyone knows what OpenStack is, but can I get a show of hands for who actually knows what RDO is? I love you guys. And who already knows what triple O is? Get out. No, I'm just kidding. So my title used to be WTF. So every time I look up there, it kind of throws me off slightly because they changed it to what the heck because we want to be gentle. This is the marketing version of OpenStack. It's very pretty, very soothing. It only shows three projects as the main and then the horizon dashboard. Very happy and soothing, makes us all feel good, and like, yeah, I could try that. That looks really easy. And then this realistic version is actually quite a few years old and doesn't actually encompass every project that we have today. Triple O isn't even on here because it's an installer, but you've got Keystone and Swift and Glance and RabbitMQ is implied based on communication and it's the flying spaghetti monster of doom. But all that you really need to know about OpenStack is that it's an umbrella project and there are a lot of sub-projects within it, and so if you're looking to get started as an OpenStack developer or to contribute to the Triple O community, you might get really overwhelmed with that kind of massivity of it. And so one of the first things I say is, well, if you sample Triple O, you don't actually have to learn an entirely new set of APIs in order to use the Triple O installer because it's using OpenStack on OpenStack, so once you learn the Triple O API, you've learned the OpenStack API and vice versa. So if you've already sampled with the OpenStack API, you already know the Triple O API. And then RDO is exactly the same except it uses RPMs as the packages. So I say RDO a lot, so I thought that should be here. So it's basically for Red Hat Enterprise Linux, Scientific Linux, CentOS, et cetera, et cetera. So Triple O is OpenStack on OpenStack, and one of the jokes is the joke about recursion. I was told I wasn't supposed to walk, and that's very difficult for me, to stand still. So there was a professor giving a talk about the sun and how the planets go around the sun, et cetera, et cetera, and at the end of his talk, a sweet person came up and said, you know, that was a great talk, but you can start all of your questions the exact same way. That was a great talk, but they said everyone knows that the Earth is flat and is actually on a turtle, and the turtle flies around the universe. And the professor said, huh, well, that's an interesting point, and that's probably how I'll reply. That's an interesting point, but what is the turtle on? And they said, ah, ah, ah, it's turtles all the way down, and this is a demonstration of recursion. That's basically OpenStack on OpenStack. So it's literally a undercloud, which is the one that's the main turtle, and it has OpenStack on top of it. And that, except that we use owls instead of turtles, and then my dyslexia steps in and starts saying owls all the way down. If you ask questions at the end of the talk, I'll throw one of these at you. That old land plate right there, so you actually have to come up and get it. So why did we do OpenStack on OpenStack? Why did we do an undercloud that is used to maintain and deploy an overcloud? The fun answer, well, it is why not? Because there's even OpenStack on OpenStack on OpenStack, why not? But the actual reason was to address craft, to match resources, because the undercloud is specifically made to go, OK, you want a compute node? Well, here is the best hardware for that node. And for simplicity's sake. So if you want to maintain several clouds using a single cloud for any reason or for someone who wants to use a development cloud before pushing, for whatever reason, you can do this. And people have done this already. They've written their own scripts, but then you're not using the same knowledge for your deployment as you are for your maintainer. So this installer specifically allows you to not have to learn any new skills. You don't have to dedicate additional resources or brain time to a separate resource. So this gets into more specificity. The undercloud has a very specific set of requirements in order to give you every possible deployment opportunity and maintenance and words or monitoring. And therefore, there are specific factors that you have to have installed on your undercloud. But then for your overcloud, you could pretty much have the most complex cloud to the simplest cloud if needed, depending on what you want to do with your open stack instance. But it is important that for the host machine, you have 16 gig of RAM. Actually, 32 is better. 16 will be very, very slow. You need to be able to SSH to the vert host as root without a password. It needs to be running Ansible and the vert host needs to be running a recent red hat based Linux distribution, CentOS 7, REL 7. But actually, I want to say CentOS 7 is the strongest, I could be wrong. You'll notice I didn't say Fedora. And why is that? We've kind of, there we go. We've kind of, because Fedora is also released every six months and OpenStack is released every six months, it's a little bit more difficult to keep them stable and what not as the releases are coming out. So if it's your first time to use Treplo, I would recommend going with CentOS, but once you are having fun and more comfortable, then by all means dive into Fedora, of course. The part that I specifically help with is this quick start. Right now the documentation for Treplo and Treplo.org is fairly thorough. And you can go in and you basically say, okay, I'm installing on a bare metal or I'm installing on a vert system, I'm doing REL, I'm doing CentOS. I'm with that REL system, I'm registering to RHN or I'm registering to a satellite system. You can get very specific. You can add Seth, you can add all kinds of basically factors. And the documentation gets very specific to what you're doing. It's still pages and pages of bash commands. And with that in mind, that tends to be a stop place for a lot of people is you're going to have typos, you're going to be intimidated. So we tried to take that out of the equation and this quick start script does just the under cloud. Cuz it's basically you install REL or CentOS. You prep your under cloud, you install the under cloud. You set up your images, you register your hardware to the under cloud. You run introspection, etc, etc. And then there's the deployment of the over cloud, etc. And this quick start pretty much does the prep from the, after you've installed CentOS 7 or REL 7, from there to the prepping of the under cloud. Which takes about three pages of commands and squishes them into three, which is nice. So one of the things that we ask people to do is to test this and see if it works for you. And if you run into issues to jump on to free node and tell me I'm Leanderthal or email me. And let me know what error message you got or etc, etc, etc. So that's my little passion piece right there, is whether or not the quick starts working. But we'd love your feedback as well. Like one person I spoke with, actually quite a few people I've spoke with, they're not developers. And so they are completely uncomfortable with bash or running commands at all. And they would like to have a GUI click click, where they just click next and it all happens. And I said, that's a great idea, I'll get right on that. In my copious spare time, but we would love those ideas because those are things that especially this week, we are working on design summits. We are talking about what the future of triple O is going to be. And if enough people were to come to those sessions and say, I want that click click interface, then we would make it happen in our copious spare time. So this is the kind of boring page, but it has all the links that are the most important. And it won't run perfectly. There's often lag between documentation and new releases and we're on Otaka. So, yeah, and as a matter of fact, there's going to be a triple O demo tomorrow at 12.15 at the Red Hat Community booth. If you would like to come actually see an installation run, we would love to answer questions there. And I would love to answer any questions you have here. Thank you. You can just not ask questions and I'll just give you a sticker, I promise. So the question was, what can you do with Ceph deployment in triple O? I will admit that that is something that is very strongly been implemented over the past six months, and that I have not touched it. Cool. And that's been implemented, tested, and is now available on the latest version. Yes. So, yeah, we've got active contributors in here, and I'll admit that I've been less active over the past six months. So I'm very happy Stephen is here. This is Stephen Hardy. Yeah, and we'll see you on FreeNode, triple O. Are there any other questions we have triple O developers in here? You can actually help us write that guide. Yeah, email me and I'll get you the direct link. Yeah, of course. And that's very recent, actually. We had someone as recently as two months ago request this specific scenario and it was not production ready. It is now, based on networking and whatnot, but the documentation isn't finished if you want to test. Yeah, the people tend to use the instances of the over cloud for testing purposes rather than having separate clouds for their over clouds. You can ask as many questions as you want, as long as we have time. No pressure. Well, if you would like to come up and get an owl sticker come on up and if you would like to go to the Ardio party, if you're already registered, it's at the Barcelona Princess Hotel across the street. Thank you very much.