Microsoft Azure AZ-801 — Section 8: Implement and manage Storage Spaces Direct Part 5

55. Configure cluster sets

Let’s talk about the concepts of cluster sets.

Now, a cluster set is a way of grouping clusters together so that you can fail over into different areas. Usually this would be done in like a large data center. So, a lot of your cloud providers, including Azure, actually builds these cluster sets through their cloud service. So, you would only really get to do this if you work for a very large environment with a lot of cluster servers. Unfortunately, it isn’t something I can show you with just a couple of virtual machines, but I am going to be able to show you a little bit of how Azure manages it to give you a better understanding of it. All right. The main thing here, I best this thing to do is look at this little drawing. They give you this diagram and you can see that. You’ve got a cluster set here and you have various you have these different virtual machines that you’re basically providing high availability for. You’ve got four different nodes per cluster with these virtual machines. All right.

The cluster set creates something called the Cluster Set Master, which is a service that’s going to be communicating amongst these different clusters. This service communicates with one of your cluster members here, which is called the Cluster Set Worker service, and it keeps track of basically these four sets of clusters as a whole and make sure that each cluster knows they’re running.

So, if these were like, let’s say, web servers that we’re providing like web and database services, and they were being load balanced between all four of these. What’s neat about this is you’ve got four load balanced instance of these virtual machines, instances of these virtual machines. But underneath each one of those, you’ve got you’ve got four different nodes in the cluster to keep that one virtual machine up and running. So, you can see the kind of underlying infrastructure that’s going on here. You’ve also got what’s known as the availability set, which is a huge important part when it comes to providing services for a cluster set.

Now, again, unfortunately, this isn’t something you can really you can really try out on just a single machine, a single couple of machines and get a good understanding. I will say that if you ever have to do this in the real world, this article will walk you through how to do it. And there are PowerShell commands that you basically have to do and fortunately there’s not any graphical tools inside of server for setting this up, but they do have these PowerShell commands that they walk you through how to do it. But let me give you an even better visualization. I’m going to go over to portal.azure.com. We’re going to click the menu button here and we’re going to go to virtual machines. And we’re going to click to create a virtual machine.

Now, let’s say, this was going to be a highly available virtual machine, right? That I’m going to be, maybe, it’s going to be some kind of web service or database service that I want to be highly available in the cloud. Microsoft has the availability options here, and we have this option here called availability set. All right. And if we look at the availability set, we can create a new availability set. And this is going to ensure that our that our information is going to be on different pieces of equipment in the Azure environment. In fact, if you think about it, like if you go let’s do a quick Google or Bing search on the term data center. All right. Let’s just get a visualization of what I’m saying. Let’s see. This is probably a good little picture right here. Let’s just look at this. All right. Let me zoom in on that. It’s not a very high quality graphic, but let me see if there’s a better. Better visualization of it. Yeah. There we go. That’s better.

So, imagine this was the Azure Data Center or any data center for that matter, really. You have these different equipment racks, right? Each equipment rack is running its own power, its own, and it’s got its own network equipment. All right.

Now, you’ve got to imagine that sometimes we have to reboot equipment. This would happen in the real world and a data center on-premises. It would also happen in Azure. And so an availability set is going to make sure that there are different there are multiple instances of your virtual machine running on different equipment in the Azure environment. So, if an equipment rack was to go down or have to be rebooted, it instantly fails over to that other version of your virtual machine.

For example, you have the you have what’s called a fault domain, and they’ll allow in Azure, they’ll allow your you to have up to three instances of your virtual machine on the different equipment racks that are in your Azure data center.

So, three instances will be in being basically synchronous, synchronously replicated amongst the different racks to provide us with redundancy. The other thing is you have what are known as update domains. And what that is, is that. It will. It will make sure that if you had let’s say you had five virtual machines that were being load balanced, it would only reboot one at a time. If you had ten virtual machines and you had left this set to five, it would reboot two of them at a time. But this just makes sure that your virtual machines are going to be stored on different racks. If you had if you had 30 virtual machines, you could actually have this availability set and you’d have ten instances of it or ten virtual machines on each rack because you’d have three fault domains. But when it comes to update domains, that’s a rolling restart scenario where the equipment has to be rebooted. Sometimes, the service has to be rebooted. They would they would be careful. The services would only reboot certain virtual machines, so you wouldn’t have all your virtual machines being rebooted at the same time. Again, if I had five virtual machines, it would it would reboot the first one, then reboot the second one, then the third, then the fourth and the fifth. Right? It wouldn’t reboot them all at the same time. All right.

So, that’s the idea of an update domain. If you just had one update domain and you had, let’s say, five virtual machines and all five of them would get rebooted at the same time. So, this is the idea of an update domain.

Now, I’m not going too deep into this in this particular lecture. I’m kind of just giving you an overview of that. But all of this stems back to the concept that of understanding how our data centers and all of that are going to deal with this type of situation. Right. Where we’re having to deal with, you know, this idea of cluster sets and all of that. Again, you could set that up in your own data center. You’d have to, of course, use a you use the PowerShell commands, but you could you could build that sort of thing even in your own data set.

56. Configure Scale-Out File Server

Let’s talk about the concept of a clustered Scale-Out File Server concept.

Now, the Scale-Out File Server is just an alternative to doing the iSCSI method. So, when I set my cluster up before I did that, I went through the iSCSI way of doing it, which is pretty easy way to do it. And the great thing about iSCSI you don’t really have to have like a third server to do it with the Scale-Out File Server. You need to have like a third file server. That’s not part of the cluster, which I don’t have. I’ve just got these two servers You could for extra credit if you wanted to add a third server. If you got enough memory on your Hyper-V box, you could add another server, join it to the domain and you could try this out. But the idea of a Scale-Out File Server is instead of using iSCSI. You’re just going to use a shared folder to store cluster information and all that in it. That shared folder is an SMB based shared folder. So, it’s the same thing as just any other shared folder in Windows. SMB is server message block. All right. You still got to have pretty decent connectivity with that server. So, the two or more cluster servers that are going to communicate with it do need to have a low latency. But once you basically set up that other server, you can create a shared folder. And then you can go right here into your cluster. You can go to roles and you can go to configure roles. Click Next. Choose File Server. Click Next. Choose Scale out file server for application data. And that’s the main thing that you’ve got to remember here. Scale out file server for application data. So, we click Next and then you give it a name. What do you want? What do you want it to be called? We’ll just call this NYC-FSVR, right? And you would click Next and then at that point you would click Next and you would click Finish. Now, you actually do have to have the other server and all that stuff set up in order to configure it and use the correct IP address for it and all that. So, it’s just going to fail. If I try to do this right now as I don’t have a third server. But again, you want to try that out. You can, you can if you’ve got enough memory and stuff to play around with. If you want to just see it with your own eyes, you can. But remember, we used in this environment here, we used iSCSI, which works just as good. And it’s, you know, it’s a few more steps, a few more clicking around to have to do. But I’ve already shown you how to set that up. And we used iSCSI when we can figure this, so I’m just going to remove that. But that’s the idea of a scalable file server is an alternative to iSCSI, but it would just be a shared folder located on a file server.

57. Create an Azure witness

Let’s talk about using Azure as a witness for our quorum.

So, might already be familiar that we can right click here on Failover Cluster Manager and we can go to More actions and we can go to configure quorum settings here, cluster quorum settings. And if we do that, we can say select the quorum witness and we have this option right here called Configure a cloud witness. And we basically can we can create a storage account in Azure. And use the storage key. And then from there, establish an endpoint with that, which is just a connection with their cloud service that will allow that storage account to act as a witness in Azure for our call. And so what’s great about that, too, is if I had my two different failover servers in different locations or whatever, as long as I have an Internet connection, we could use that Azure service as a witness. So, that’s the idea of setting up Azure as a witness.

Now, if we want to do this, we will open up our web browser, go to portal.azure.com. I’m going to click the menu button here and I’m going to go to storage accounts. All right. Storage account is a service that’s going to host storage type services inside of Azure. We’re going to click to create storage account. And so we’ll just create a storage account. This video is not really trying to explain the intricacies of storage accounts. We’re just going to create one real quick. I’m just going to call this storageaccountrg for my resource group. All right. And I’m just going to call this elpmystorageacctwitness. All right. That’s going to be the name of my storage account. Storage accounts need to be all lowercase and they have to be unique across all Azure. I’m just going to go with Standard, and we’re just going to do LRЫ, which is locally-redundant storage. And I don’t think there’s anything else I really need to configure here.

So, I’m going to click Review. It’s going to run a validation test and then I’m going to click to create and I’ll go ahead and pulse the recording while that’s getting created. All right. Once that’s done, I’m just going to click Go to Resource and that’s going to bring me into my storage account. All right. And I’m also I’m going to want to copy this storage account name. So, we’re just going to right click that. We’ll go over here. We’re just going to paste that into the storage account name.

The next thing we’re going to need is the key. So, we’re going to scroll down here and go to access keys. All right. We’re going to go right here for our first key. Key1. We’re going to click Show this key, and we’re going to copy that clipboard and we’ll jump back over here and we’re just going to paste that in. All right. And at that point, we’ve now officially got everything we need to create our cloud witness.

Now, I’m going to click Next. And all right, checking everything else is everything looks good so far. We’re going to click Next. And it’s now going to go through the process of trying to actually perform the link.

So, as you have successfully configured the quorum settings for the cluster and we’ll click Finish and there you go. We’ve officially set up the storage account as a witness.

Now, I will advise you, I would recommend eventually you’re going to want to delete that storage account because it is it is taking up. It is going to cost money to keep that storage account in Azure. So, you might want to keep that in mind. So, we can do is we can just go right back over here, we can go to More actions and go to Configure Cluster Quorum. All right. We’ll click Next. And we’re going to say use default quorum. All right. We’ll click Next. And next again. And that switches us back. And then the last thing I’m going to do is just delete that storage account because I don’t want it costing me money. Well, it’s costing Azure credit, right? You have if you went through and set up the free trial thing, then you’ve got some Azure credit. But we don’t want to waste that.

So, I’m going to go to Resource Groups here. And we will go into the storage account resource group that I’ve created and I’m just going to delete the resource group, which will also delete the storage account that’s inside of it.

So, we’ll do that. Put the name of the storage account in here. Delete. And it may take just a moment, but eventually you will get a message saying that the that it was created successfully. And should be able to even refresh my screen and just see that everything is gone. But ultimately, that gives you a good walkthrough, I think, on setting up an Azure Witness and then also how you can just kind of clean things up afterwards there.

So, yeah, so as you can see, resources not found, so it’s no longer finding the storage account. If I go to resource group, that resource group that I created should be gone and it is. So, we’re going to go.

58. Visualizing floating IP address and load balancing for failover clusters

I’d like to now go over the concept of a floating IP addresses and load balancers in regards to clustering.

So, you can utilize load balancing equipment for load balancing different cluster sets. So, for example, let’s just add some servers here. And we will lower the font down a little bit here, too. There we go. We’ll just call this a node. So, this is. We’ll say we have three servers here. And then let’s just copy that. And we will have three servers over here. All right. We will call this Cluster A and we call this Cluster B. All right. And we have a storage replication going on. You should understand those concepts. And so what we can do is we can use a load balancer.

A load balancer is a piece of equipment that you can purchase that directs traffic. Right. But there is a slight problem. And we’ll take a look at what that problem is. So, if traffic comes in, the load balancer is going to load balance to this cluster first; and the next bit of traffic that comes in, it’s going to send it to this cluster, right? Next comes in, it’s going to there. Next comes in, it’s going to go to their right. And the idea is if the whole cluster was to go down, it would just keep sending it all to Cluster B, right. But what if the load balancer fails? All right.

So, you have an IP address. Let’s say, that if this was an Internet based scenario, you would have like an IP address that would be, you know, maybe, a public address. It could be a private, though, if it was an internal based situation. All right. But what if the load balancer fails? Well, no more traffic is going to be able to communicate with either the clusters. Right. So, the way around that is to purchase another load balancer. So, we have two load balancers. All right. And so the idea, though, of all of this is we do what’s called a floating IP address. All right.

So, what is a floating IP address with load balancers? You are allowed to set up this thing called a floating IP address, and the IP address will be active on whichever server is up and running at that time. It will be active on that server. So, you’ll have an active load balancer and you’ll have a passive load balancer. This will be our passive load balancer. All right. Right here. So, what happens is you have connectivity. Like so. All right. Both load balancers can communicate with the equipment and the load balancers can communicate with each other. Load balancers will actually use heartbeat messages as well. And pretty much any load balancer equipment you get, a fibre or whatever equipment you buy, it can support this, by the way. And so, the idea is for some reason, if this load balancer dies, the heartbeat messages stop. And guess what that IP address is? That floating IP address is now going to be over here as the active load balancer so traffic can still flow in this load balancer gets control of that IP address and traffic can still flow in and B, load balance between all the clusters. All right.

Of course, you can also when you configure the load balancer, you can configure a fallback scenario so that when this guy comes back on line, start sending a heartbeat messages again. Then at that point he becomes the active. All right. But that’s the idea of load balancer.

Now, of course, we’re not really getting into configuring a load balancer because we don’t have equipment or any of that. But in the real world, that’s basically what you would have to do. You’d have a piece of load balancing equipment and you would, you would configure all that. All right. And you can actually use an Azure Load balancer as well. If your company connects directly to the Azure Network using like ExpressRoute or if you connect to the Azure Network using like a VPN Gateway or something like that. You can use an Azure Load Balancer also.

But traditionally, this would be done if you were going full blown with on-premises side of this, then you would use something like a fibre or one of those companies. All right. Hopefully.

Now, that gives you an understanding now of the idea of floating IP addresses and load balancing.