Welcome back. In this lesson, we look at using capacity reservations for worker nodes in your OKE cluster. As a reminder, the compute service enables you to create capacity reservations to ensure that compute capacity is available for workloads when required during critical events such as disaster recovery or unexpected workload spikes. Once you've created a capacity reservation in the compute service, you can then specify that capacity reservation when using OKE to define a node pools placement configuration, which will then be leveraged when scaling the nodes in the pool. Now, if your region has more than one availability domain and you're using all three AD's for your node pool, as in this example, then you'll need to create a separate capacity reservation for each AD. In fact, we'll look at this in just a moment in the demo. Specifying the capacity reservation ensures that the node pools worker nodes are created with new compute instances from that reserved capacity. In this example, our node pool has a total capacity of 60 worker node instances. Let's now take a closer look at how this is managed in the OCI console. Prior to specifying a capacity reservation for our cluster, one must first exist. I go to compute and capacity reservations. You can see I've already created three of them. I wanted to demonstrate real quickly how easy it is to create one. You simply click on create. I'll call this demo. I'm going to associate it with, in this case, one of my three availability domains. Of course, some regions only have one availability domain. Then in this section is where you decide you can either dynamically let it choose a fault domain or you can specify one. Then, of course, you specify one or more compute shapes. In my particular use case, I've been using the VM standard E21 with just one core and eight memory and then to specify some account of whatever should be appropriate. You can actually add additional shapes as well if you need to. For example, maybe I want to add a few more of these and allow for maybe up to two of those and so forth. Once you've done that, it gives you a review as how it's done and you simply create it. Now, for our demo demonstration purposes, what I've done is created three I intend to use. One assigned to the AD1 availability domain. We can see its capacity configuration is 420 of those. If I wanted to add another one I could or if I wanted to edit the configuration and simply change the count, I could do that as well or I could edit to make a change for this one by simply adding some more capacity. For example, I could add another virtual machine type and a specify account and add that configuration to our shape. Now we'll have both of those. Well, I'm going to revert that back in just a moment. Let's go back and look at the other two though real quick. The AD2, of course, is got a capacity of 20 of those shapes and the AD3 has the capacity of 20. I'm go back and fix the one that I changed and simply delete that configuration. Now that we've validated our three capacity reservations, let's head over to our cluster configuration. Show how you would then specify this when creating a new cluster. First things first, you can't use the quick create wizard, you have to use the custom create in order to make the specification. I'll launch that workflow and will simply select a VCM that already exists for us. Select one for the load balancer, select one for the API endpoint subnet. When I advance to now configuring my pool, I'm going to specify that shape that I decided we're going to be using. Let's go ahead and use the latest version of Linux. We'll go ahead and set up six nodes. Now if you'll notice down below is where we can do the placement configuration, of course, one for activity domain. The first one, assign it to the private subnet. When you expand the options, that's where you see the default is on-demand capacity. Here's where you specify the capacity reservation. Look for one of those. That's the one that's associated with this activity Domain. We'll add another one for the AD2. Over there I just created that demo one. We're going to be using this one and one more row to assign nodes to the AD3 specifying our capacity of 20. We click on "Next" and there's a summary of creating my configuration. Now, I'm not going to create the cluster right now. It would take a while, but you can see the overview of the configuration, the subnet, the capacity reservation, and so forth. What we can do though, is make a change to an existing cluster I won't show you how that would be done, I already have a cluster here that has one node pool. Let me slide down here to show you that I have one node pool. There's two ways you can make a change to add a capacity reservation. One would be to add a new node pool. And then with that new node pool, just like I did a moment ago, go ahead and define what that should be. Number of nodes, assign the availability Domains, use advanced options to do the capacity reservation and so forth. The other way to do it is to modify the existing pool I actually used different names right here. This was the one I had from before. The way to make a change there I noticed the placement configurations are configured the same, but it's an on-demand capacity. To make a change, instead of clicking "Edit", you click "Scale". Now, I could change the number of nodes or leave it the same. But here's where I can now go through. Right now I can go into all three availability Domains but if I look at each one of them, they're all set to on-demand capacity. Here's where I can make a change by temporarily increasing this now or leave it the same and go in and change that to a capacity reservation accordingly. Let me go ahead and do that real quick. As I make these changes and click on "Scale", it will automatically now add instances. Let's just do six nodes for fun and scale this out. What this is going do is going to take from that reservation those additional five that get instantiated. If we can wait for just a moment, we can then take a look at the nodes in that pool. Well, this is the one that was already running. The rest of these are beginning to be instantiated. I will just wait a moment for them to become ready. Now, that they're all provisioned and ready, I can go take a look at a couple of things. First of all, I can see my placement configurations up here with the capacity reservations. If I were to select one of these, for example, this instance, I can look at it from the Compute Service. And in that general information page and the Compute Service, I can look to see its capacity type is not on-demand, but capacity reservation and which reservation is associated with. If we wanted to navigate over and take a look at those capacity reservations, I can now see that out of that 20 capacity, two of them have been used. You can even navigate to see the other instances associated with that reservation right here. Notice it automatically puts them in two separate fault Domains. The other thing I wanted to point out to you is that very first instance that was already running. If I go and take a look at it, it will actually show me that as opposed to a capacity reservation, it still is listed as on-demand. The reason why it's going to be listed as on-demand for capacity type is because it was still running. If I want for it to count against that capacity reservation, I would need to go ahead and terminate this virtual machine instance and then let oke, provisioning the other one against the capacity reservation. That's it for this lesson on using capacity reservations for oke cluster node pools. Thanks for watching.
