Transcript
you just create a new repo, point it at the URL, and once you do that, what you have available is an extension for the VergeOS node driver. As I mentioned earlier, that's what provides the functionality for Rancher to store credentials and provision the cluster against VergeOS. In addition to that, there's some UI extension stuff for Rancher also bundled into here. When you go and look at providers out here for node drivers, without that, VergeOS here wouldn't be on the list, but you can see it's available after you install that extension. You can, once that's also installed, you get this availability to create cloud credentials, and here we are alongside all of these other cloud providers, and this is just a basic Verge 26 and above API key and endpoint required, just to authenticate against the VergeOS cluster, and once you just have that small number of things, no massive tie-ups with kubectl or anything like that, you just go into clusters in Rancher here, click on create, pick your VergeOS provider, select one of your credentials, give it a name, your machine count, template VM. In this case, for demonstration purposes, I'm just placing it on an external network, but this is just a name network within VergeOS. No cloud in it is necessary here. The defaults will work just fine, and you click on create, and then you see the demo webinar cluster you're creating. Now I've seen, even locally, when I've done this stuff in the past, messing with EKS Anywhere or other hypervisor providers, it's like, it's at this point that it's just like, you know, five, 10, 15, 20 minutes later, what's going on? And there's not so much of that in this case. We can see right away that after clicking that, it's created these three VMs in VergeOS, and it's already began provisioning the cluster from just that minimal amount of input. Generally, this takes about five minutes, but the whole time you have this available in Rancher to look at it, firing up these VMs, waiting for the IP addresses to become available, and report back through the QEMU agent. And then once Rancher has knowledge of that IP address, it uses SSH, goes out to that VM and starts provisioning. It's a rather hefty runtime on these VMs. These are running kind of light with just two cores and four gigs of RAM, but, you know, obviously for a production environment, you'd be giving maybe a bit more resources depending on your needs. And the whole time this is running, in this case, just wanted to talk through it a little bit to kind of get across the point that during the course of a normal short conversation about the weather, you can provision a VergeOS cluster using Rancher. I did a lot of the work on the node driver, and as I was working on it, and it's part of the reasons that I also did the UI extension, is I had the node driver and I'm thinking, well, that's great, but now I've got to do all this kubectl stuff, kubectl as it's affectionately known, command line stuff to get it working, and that doesn't really run as well as a demo. So then I started my foray into developing a UI extensions for Rancher and was able to just get this operating with a little bit less friction. I wanted to mention too, that a couple of times I've mentioned kubectl, you can very well use that method as well with our node driver and with the integrations, this just provides the ability to do it through the Rancher UI completely. And generally, I think when you meant timeline, like how long does it take to deploy a cluster, even on my lightweight constrained lab with little VMs, with two CPUs and four gigs of RAM, I'm looking at about five to seven minutes for a cluster become available. It's also nice is that this is utilizing the vSAN and BurgeOS, so the deduplication, the rapid provisioning of VMs, all of that stuff kind of adds up to make this stuff move a little bit quicker and the limiting factor becomes Rancher itself, just moving binaries into these VMs over the network. And it's also Rancher, I think is an excellent way to get familiar with a lot of this stuff too. So I liked it for that. And this is all running on, I'm running my lab at home here on little minis forums with 96 gigs of RAM and a couple of terabytes of storage on NVMe, it's pretty amazing. So here we made it, we created a three node RKE cluster in six minutes. Now they're a little strained obviously, but it's out there and it's running. Right, and this is really just to show people how simplified it's become, how easy it is with the BurgeOS and the Rancher. And I want to at least demonstrate what that looks like. So really appreciate what you've done David and showing us live. Thanks.
