 Yeah, so our infrastructure's just not set up for that. Well, you can't do 5-9's availability on a 3-9's budget. Look, this is coming from the top. Can't just sit here and make excuses. Not much time. What's going on? Banes on the loose, and our crime-critical systems keep going down. This is what happens when you put a kid named after a bird in charge of I.T. Holy sap kinda, Batman. When your systems go down, you lose business. When mine go down, people die. Right. So what's the problem? Hardware failure. Some of these mainframes haven't been upgraded since the Adam West days. Good times. We need to hurry on it. Batcave defenses armed. Tracking systems online. Batman, I expected more from you. Your defenses are weak. My defenses were weak. So you should be good to go, Batman. You're fully clustered. Mainframes and x86 systems all upgraded to SUSE Linux. High availability extensions are working. No more downtime. Thank you, SUSE Master Builder. We can never repay you. Oh no, it was no trouble. No, I made some horrible investments last year. We can never repay you. Oh, OK. Eric. Eric. Bots? Guys, we've totally got this. Drive on there, that is a bootable drive that will image whatever you put on it, boot it up off of, with the appliance that Cameron is doing. It does have a little prompt. It says yes or no, do you really want to do this? So there's one level of protection from brain deadness. Just be careful, exercise caution. Yes, be careful, please. You can only use your powers if you're good. So does everybody have one yet? OK, everybody's got it. Good, good. Oh, I didn't even notice they were on the seat. So yeah, this will be fun. The little thumb drive with the cloud on it. So you can go and take that back to your room tonight or this afternoon. And deploy a cloud and start playing around with it. This will allow you to play around with it really, really quickly. So what are some of the reasons why you guys came today? So the comment there was there's too much theory going on. You wanted to see some actual practical implementation. Too much theory, OK. Anybody else want to make a comment? All the rest of you guys are open. Usually the comments are, you need a PhD or, you know. Yeah, you want to see how it's supposed to work. OK. All right. Has anybody here actually tried to do and failed at the, just doing the, I'm going to grab everything from GitHub and check it out and do it all. I'm raising my hand legitimately here. I failed at it, too. OK, that's most of the room, OK. All right, well, that's fine. I'm going to show you right from the very beginning on how to deploy your cloud with just a little bit of interactive slide deck here as well. And so we'll kind of go back and forth and we'll deploy a full cloud today, all right. Now, there are some things I'm going to show you that you can actually deploy a cloud much quicker using some nice tools that we've implemented with our product to be able to batch deploy your entire cloud in minutes, OK. So starting from this little thumb drive, you can plug it in, use it for VMware. You can run your cloud on top of VMware. You can run it on VMware Workstation if you really want to test around and play around with it. You can install it on your own hardware, just plug it in and boot from it, play around with it there in your own lab. We also have some different formats available. We have a VMX format. We have some KVM formats. So if you really want those other formats, we'll later on give you a location where you can grab those formats, OK. So from this thumb drive, installing the physical machines, virtual machine, and create your cloud platform. From a single image, we then have our admin node. And that's essentially what we're going to be creating here is our admin node. That admin node has all of the tools needed to deploy your entire cloud, OK. So let's take a look at that admin node. I'm going to switch gears away from this deck and actually just go to VMware. Can everybody see that? Hopefully that's not too small for everybody in the back of the room. Let me show you what I have here. It's eight gigabytes of memory, two processors. It's 20 gigabytes of hard disk. Ideally, in a production environment, you'll want a little bit more so that you can mirror down a bunch of software repositories that are required for your cloud, possibly later on to do updates and other various things in a production environment. I only have a single network adapter. Ideally, in the real world, you're going to want more. You could do this in VMware. I'm going to make it really simple today and just do a single virtual network adapter. But ideally, in the real world, you're going to want at least probably five, six network interfaces. And your hardware is going to come with that anyways. And then I have an actual ISO image of the thumb drive that I'm booting up in VMware. I could plug this guy in and point it to the physical if I wanted to as well. And it will work just the same. So let's go ahead and start this up. Wasn't fast enough. I'm going to boot from the CD-ROM drive, boot it off the ISO. This is what you're going to see when you boot from the thumb drive. And I'm going to select Install. And he will come up with loading the kernel. And from that point, he'll ask you, do you want to blow away your disk? Yes, question. Yeah, the only thing that's customized in there is I'm disabling the DHCP on that virtual network interface. It's actually a knadded network. So the question is, why do I have it as a custom network interface? So it's custom so I can, for one, disable the DHCP. And it is a knadded network. And the reason for that is that it's going to have a Pixie Boot server on there, and you don't want to expose that out to the outside world. Yeah, and on the host side of the machine, I could set up VLANs as well so I can get access to the public interface of the cloud and various other things like that. So kind of slow, this boot up part. Any other questions to this point? There we go. Destroy SDA. So this is what you get if you plug this into your machine and boot it up right now. So we are actually going to do that. And he's now loading that image onto the drive. That's what he's doing right here. Yeah, that's what we're doing right now. This is VMware Workstation that we're doing this in. Yeah, no problem. Any other questions while this is loading? Any good jokes? This is the painful part, really. Yeah, if you boot it off that machine and click yes on the destroy data, yes, your laptop will be gone. Yeah, you can put it in there while you're booted up and go and look inside of it and pull stuff off or whatever if you want to, but don't boot off of it unless you are feeling dangerous. Now we're going to get to a point where it's going to ask us a few questions just as part of the setup of the appliance itself. Actually, I would point out, Cameron, that with the file system on there, you can choose any size disk you want and it'll actually detect what the size of your system is and expand out to fill whatever that is. So you don't have to worry about choosing a storage to match up with the storage on the image. Yeah, you've probably missed the message going by the scrolling on the screen there, but it was actually extending the iNode table for EXT. So we'll go ahead and hit Next on this. You can choose your language settings and keyboard layout and then you can put in the password that you'd like to use for the appliance and then, of course, setting your host name and domain for this infrastructure and then setting up the network. Now this is really simple. We actually have pre-populated the network with some defaults that we like. These are actually defaults that we use in our documentation. So you can go through and you can actually change these. So if you don't like the IP addresses that are in here, feel free to modify that, how you like. I'm gonna leave it as is. And if you want to get an explanation of what those networks are and what they mean, then stick around for the next session. There's a couple of things I wanna change on the host name and DNS. Since this is under VMware, it routes through a .2 and not a .1 from the virtual machine side. So I do need to change that and then also change the default gateway to a .2. This is only inside of VMware for this case. And then you hit Next and it is writing all that configuration out to the disk. And now we can set up some time and NTP settings. I'm from Boise, so we'll set it for Boise. And then we'll hit Next. It'll give us some options to set up and configure NTP. It takes a moment to write this piece to disk. And then we'll go ahead and add an NTP server. We'll then select a public, that is okay. Write that config to disk. Now, most people skip by this screen, but it has some very important information on here on what to do next. We do have some documentation for this appliance. So if you do skip by this, you can refer back to that documentation. But essentially, the next steps are to run YAST crowbar. It's a module for YAST that allows you to configure the rest of crowbar, customize the rest of your network stack for all the various network pieces for Neutron and things of that nature. And then run a command called go. Go will essentially deploy crowbar onto the admin server. So we'll go ahead and finish this. And we'll load up the YAST crowbar module. And we do have a beta ULA in here. YAST crowbar. And now this, you can actually change the password for the crowbar user and a couple of other things set up some other repositories. I'm gonna focus on the networks. So this is where you can go in and modify the networking for your entire open stack setup. This is very important because if you don't get this part right, then you will have to redo this. And by redo, he means reinstall the crowbar. Yes, reinstall, exactly. So I'm going to change, again, the router should not be a dot one, it should be a dot two because this is running in VMware. And so that is actually the only parameter I'm going to change. However, if you do want your floating network to be much larger, you will need to modify that because it really only accommodates for, I guess around 50 IP addresses, which is not very big. So you'll really want to modify that for a production environment. And the next part is running screen and the go command. Now we put it in screen, I'm running it from the console here. So screen probably, it wouldn't really matter, but if I was running it from say SSH, the crowbar execution actually restarts the network and things and sometimes it disconnects and causes the whole script to actually terminate. So that's why we use screen for the command. It'll go through some checks and start installing some services and completely deploy crowbar and then we can get into the web interface. Once that's complete, we can actually do a batch deployment which we'll show you here in just a moment. Let me go back to the deck while that's actually configuring there. Any other questions up to this point? Yeah. I am not running SMT on this image. This particular image does not have SMT installed either. There is an alternative image out on susastudio.com that does have SMT. So for those of you who are not familiar, SMT is called a subscription management tool. That's a free SUSE tool that caches down your patches and updates locally so you don't have to go back and hit R servers every time you want to pull patches. Yeah, that makes sense. So you'll want that for a production environment because that will give you all the repositories that you're going to need to maintain your cloud environment. Okay, backwards there. So we have our admin node. Currently crowbars being deployed. From there, we'll actually deploy our control node and then our compute node, our storage nodes, and then configure the rest of our infrastructure. So we're going to use a batch in this deployment today. The batch will allow you to deploy your cloud in minutes. And in this particular case, it'll probably take about 18 minutes to fully deploy using the batch. Yeah, so the batch script is actually written in a YAML format. So it's really easy to read and it has the capability to actually go out and export. So if you have a configuration you've already set up prior and you want to duplicate that and create another cloud environment, you can actually do an export of your configuration into the YAML format and then reuse that again. So it's very convenient for replicating your OpenStack cloud. The orchestration tool behind all that is crowbar. If you're not familiar with crowbar, crowbar is a tool that is built on top of Chef. And so Chef is doing a lot of the work in the background. So it's sending commands to Chef to go out and do work for the nodes that it's managing. Let's go ahead and go back to, it's almost done. It's transitioning to ready and then we'll be able to log right into it. Who here's familiar with crowbar? Nobody, okay, well this will be new for you then. So crowbar was an open source project. It was originally started by Dell and they were on it for a while. They transitioned away from it when they started doing partnerships with other companies that use other deployment frameworks. And so at this point we are the primary maintainers of the crowbar project. So some of the things that you run into with crowbar, such as if you have to change the networking, you have to rebuild the whole thing, that was kind of a legacy decision that we're in the process of trying to fix. But overall it's a really good and very flexible tool that enables you to get things done quite quickly. All right, it's almost ready. Any other questions at this point? Okay, he's telling us he's ready. Just add a YAML wrapper around so you don't have to do this chef for a code Ruby. Yeah, you're not writing, for most deployments you're not doing any code in Ruby. You may, if you have to have some sort of customized deployment, may have to alter one of the recipes that are in place right there. But for the most part it's just point and click through the graphical interface that Cameron's got up right there. You'll basically, you get this set up and then you network boot your, pixie boot your systems against it. It'll auto discover them and then you can start doing things with them. In the case of this demo, Cameron's gonna batch it all out in a script but you can very quickly and easily do it from drag and drop as well. So this is crowbar. This is the dashboard for crowbar. And all that's managing right now is just the admin node. That's not really special. So let's go ahead and start up a compute node and a controller and get these all running from here too. Alice. Now ideally you want to pixie boot. Don't do that guy. These nodes. It will actually hit the crowbar node and pixie boot off of that. So the admin node is set up to pixie boot in the entire environment. So from a bare metal perspective, it's set up to manage the entire infrastructure. We'll do the same on this one. Yes. It is a full-fledged Chef server. Yeah. So if you're familiar with Chef, you're probably going to be able to dig right in. From a support perspective though, you don't want to go doing too much with it just because you want to remain and not invalidate your warranty kind of stuff. Yeah. Talk to us before you do it. Should we make our next? So it's all designed around around managing your cloud infrastructure. So if you're doing things like tweaking how the cloud is going to be deployed, those things are fine. You may want to tell us about it ahead of time so that we can make notes in your file so that when you call into support, they don't freak out. But I would not go in and add in stuff for like managing stuff outside of the cloud and you're trying to do your whole company off of the Chef server on here. And luckily none of our engineers are in the room so I can say that. What? It's all on tape. It's all on tape. So what I'm doing here is I'm drilling down to the compute node. I'm going to edit the compute node. I'm going to give it an alias. The aliases are actually used for the batch deployment. I'm going to call it compute one. And then I'm going to allocate it. The allocation process installs the operating system. You do have choice between what operating system you can install. Right now, since I've already allocated it, his target platform is SLEZ11 SP3. Now it does give you choice between Hyper-V as well and SLEZ12. So if you wanted to attach to a Hyper-V environment, you could do that. It does deployment for Hyper-V node. So this will run through the install on that compute node. So going back to the dashboard, I have another node in there. This one's going to be our controller. We'll change his alias. Here's the dropdown I was referring to. Notice the other OS is in there. They're grayed out because I don't have the bits on the disk to deploy that. I could also change my intended role if I know what that's going to be. It's very helpful if you're doing this manually instead of batch because that way, whatever you select as control node is always going to be the default selection for any of the control node services. Yeah, absolutely. Yeah, for the Hyper-V. And the SLEZ12 ISOs are not on that image. So if you did want to do SLEZ12, you would need to download the ISO for SLEZ12 and put it into the admin appliance, okay? There's a specific location for that that's called out in our documentation. You would just extract it to the disk and that would be made available for use. This isn't super exciting here. So we're going to switch gears and talk about something else. So once we have a cloud deployed, there are various other templates in the cloud. You're probably familiar with heat and things like that. Heat can be used to basically orchestrate your applications in the cloud, deploying your instances and your applications out into your cloud. So deploy your OS, Apache, set up your networking stack, set up your storage and everything that's required for those instances in your cloud. And it's all orchestrated from the control infrastructure through heat out to your compute nodes. Once you have a compute node up and running, you can scale up, scale down that environment using that same infrastructure. If you're familiar with Solometer, that will allow you, if you're using that in conjunction with heat, it will allow you to scale up and scale down that environment. From that same infrastructure, that same batch deployment, if you go back to that admin node, back to crowbar, you can also do the same thing and have a load balance environment as well and be able to load balance your environment through heat and Solometer, scaling up and scaling down and adding a proxy environment front of, let's say your Apache web environment or your database environment. So you can make that load balanced in your infrastructure. Scaling up to multiple nodes. Let's see where that's at here. That guy is not going anywhere. That was the second one you did, so it's a little bit behind. So the really cool thing about this demo is how exciting it isn't. Right, because he's not really doing a whole lot of things. He's just kind of pointing and clicking in a couple places and it's doing what he needs it to do. If he was having to do some elite ninja hacking, that would be a problem. We do have images for the compute infrastructure too. So if you don't have the capability of doing pixie boot, we've got an image for that. And then you just use it. There's a Chef join, or it's a crowbar join script that joins that server to crowbar so that you can use it in the infrastructure. So there is another method besides just doing pixie boot. So if pixie boot's not an option for you, there's another method. And you could manually install Slez on it as well. Question? So have we ever run it on ESXi? And yeah, no, we've run it on ESXi. And it does work there, absolutely. VMware is a fully supported option for deployment. We support VMware Hyper-V as in KVM. And then we've got technology previews for Docker, which hopefully will not be technology preview for much longer. And trying to think if we have anything else. Nothing's coming to mind. You do? Be careful though, because you might have another pixie set up there. So you might have to isolate the network or something so that you're not trampling over somebody else. The plan was not to watch this stuff. I actually had it fully deployed this morning and for whatever reason, the networking stack on the host would not connect to it. So we're redoing it. It's fine. Demo without a net. So one compute node's done. This one's almost finished. He's going to boot back up here and run an auto-yast and NIT script. And then he'll finish up. Let's go ahead and take a look at Crowbar. He's come up green on one of them. Let me just dig into Crowbar just a little bit more for you. Crowbar has this thing called bar clamps. Bar clamps are basically a subset of software that it will deploy. Plugins. So we have the option to select open stack. And we have all of the suite of software available for open stack out here. And we can go through and start deploying all that. So if you want to do an HA infrastructure, you will absolutely want to set up your pacemaker first for your control infrastructure. And that's actually pretty cool because HA is something that can be difficult to get right. So in this case, it's completely automated. You can just basically drag a couple of systems together and say, I want you to be part of the cluster and just click the apply button. And that's pretty much your setup, aside from getting your stonet device attached and whatnot. Yeah. Yeah. So let me go ahead and see. We'll hit Create on this database. And there's not a lot of options to choose from. Although SUSE is not giving you any choice on what database you can use. The database we support here is Postgres SQL. So you could go into raw. And there's a few more options. Most of the time, you're not really going to need to do this. There are some other bar clamps like Keystone. So if you want to tie into an LDAP infrastructure or something like that, you may want to go into the raw mode and edit a few things. But for the most part, that's pretty rare. You're going to have to go into the raw mode and edit a few things for the configuration. We put a lot of options out there that are most commonly used. And we make some good choices for you on some of those options as well. So we make it easy there. And our controller and computer are fully deployed. Back on the admin server, let's just make sure that that template was deleted. He's gone. He won't conflict now. We do have, on that USB thumb drive, there's a couple of templates already pre-built for you. So you can go try those out. They're on Roots home directory. There are three files. You have drbd.yaml. You have nfs.yaml. Those two are for an HA environment. The drbd will set up a two-node cluster with drbd replication for the database and Keystone. The nfs will set up a two-node cluster using nfs as shared storage for your database for Keystone. So you could go with a three-node cluster if you wanted to on that. In a production environment, we recommend running three separate clusters. You would have a data cluster, which would actually have your database and Keystone running in that cluster. You would have a networking cluster. That cluster would run Neutron with a three-node cluster. And then you would have a services cluster, which would run all of your other services in your cloud. So let's go ahead and launch the one that says Simple Cloud. You do crowbar batch build, and then point it to your yaml file type this morning. And then off it goes. You can actually watch what it's doing from the web interface. You go to bar clamps and go to deployment queue. And it's going to commit a proposal here. And it will show you that it's actually deploying some provisioning various things within the nodes that are running in the infrastructure. If you watched the CLI right now, it's deploying the database bar clamp so it's putting the database out there. It's just going to go down through all those bar clamps and start deploying. It's deploying the cloud right now. So this will be up in a matter of minutes. And you'll have a full cloud deployed. Any questions? The networking node. This is a Simple Cloud, so this is a single controller infrastructure. So Neutron is going to be running on the controller with everything else. That I wouldn't recommend for production at all. But just for testing it out. This is really simple and easy. Yeah, question in the back. So he's asking if you can use crowbar for upgrades, right? And funny you should ask. But we actually have that built in to the infrastructure. Now you can attach Susan Manager to your environment to give you a much better and broader update stack. But we also have another option if you don't want to go with Susan Manager. We also have the option to deploy the Susan Manager client through crowbar. But we have this thing called Updater. It's just another bar clamp proposal. You can actually go in and edit that. Select patch or update or distribution upgrade and pull over the nodes that you want in that and apply the patches. It's worth pointing out that you have the ability to filter out patches that are going to require a reboot. So if you don't want to install kernel patches on a regular basis, you can exclude those. But upgrading from version to version is fully supported by Susan and has been for several since the Havana release. Good question. Any other questions? Yeah. So he's asking, what about SEF and other things that you might want to deploy in your cloud? So SUSE OpenStack cloud is set up to actually deploy SEF. There is a bar clamp for that. And you can actually go and create that proposal and deploy a SEF cluster using crowbar. So what that is going to do, though, is we have a product called SUSE Enterprise Storage where we're doing support for SEF. So this bar clamp is going to do that. Otherwise, you have the ability in the other bar clamps, say, for glance. You can select the rados gateway and tell it where your existing SEF cluster is, and you can join it up to that. So we're at the top of the minutes for this session. So let me just wrap up really quickly. It is SEA Bar Clamp Proposal Neutron. So it's setting up the network right now. It's got a few more bar clamps after that, and the cloud will be fully deployed. And then you can go in and start up instances and play around with it. So go play around with this with the cloud. I just gave you on the thumb drive earlier. And if you want updates to that, come up here. We'll scan your badge. And we'll send you update information on the latest version of this appliance. So we will be having regular updates to this appliance, making it better, improving upon it, adding more things to it. So if you want that, come up here. We'll scan your badge for it.