Microsoft Azure AZ-801 — Section 10: Implement and manage Storage Spaces Direct

63. Create and configure storage spaces

Now, Microsoft includes a feature that’s been around for a few years now in server. That’s meant to give us better control over having multiple hard drives and volumes on those hard drives. This feature is called storage spaces. This is again, the ultimate goal is to basically allow us to have more control over our drives.

Imagine you’ve got a server, you’ve got multiple hard drives in that server and you want to set up volumes on it. The problem is that what are you going to do? Are you going to set up like if you had three drives, you’re going to set up three different volumes with three different drive letters? Wouldn’t it be great if you could just connect all that together? And then wouldn’t it be great if you could also provide redundancy for that? Well, you can.

Now, a lot of you probably are aware of hardware based RAID and all of that, but this doesn’t require you to have any special hardware. You just need the drives in the server and then you can actually connect them all together into what we call a storage pool. From there you can… Let’s say, have three drives at 127 gigs a piece, which is kind of what I’m going to demo here. You know, you can combine all those together and then you can take all of that space that’s in the storage pool and you can make virtual drives that sit on top of that. And these virtual drives can be combined to provide redundancy, mirroring or parity. And so this is what you’re really going to get out of this. It’s a great feature. It gives you more control.

Now, I will say hardware based RAID solutions and all that is generally the better IT-acceptable way to do things. But if you are especially in a smaller environment, you just got to get you want to get some redundancy and you don’t have a hardware based controllers or whatever to help you do that. Or, maybe, you’re working with virtualization where there’s, you know, you’re in a virtual environment, not a hardware based environment, and you need to provide some redundancy. This is a great way to do that.

So, the first thing I’m going to do is I’m going to add some more disk to my server to do that. I’m going to right click my server here in Hyper-V and go to Settings, and here’s my iSCSI controller. I have this disk I created previously that my data disk, I’m actually just going to remove that and now I’m going to add go right here. I could use that disk, but I’m not going to create some more. So, I’m going to click Add from there, I’m going to click new dynamic expanding and we’re just going to call this Storage Space Disk1. Click to create it. It’s going to be 127 gigs. That’s fine. All right. And I’m going to create another new disk. I’m actually going to create three. All right. So, we’re going to go new next dynamic expanding. And we’ll call this Storage Space Disk2, 127 gigs. That’s fine. Finish. It’s dynamically expanding, so it’s not going to take up too much space. And then finally, we’ll do one more. All right. So, new next dynamic expanding. And we’ll call this number three. I’m going to go ahead next and finish. We created three disks at 127 gigs a piece. And so we’re good to go. We’re now ready to jump into our virtual machine.

So, here I am on NYC-SVR1. I’m going to go to Server Manager and go to file and storage services and we’re going to click on disk and we should see that our three disk are showing up. If you ever do this, ever add disk and it’s not showing up, just click the little refresh icon. It’ll take a moment. It should show up. But I’m going to go here to storage pools now, and here it is. It’s telling me I’ve got primordial. Primordial means it’s the original disk that’s on this drive here or that’s in the machine. So, that’s fine. And I’ve got three more physical disks that I could use.

So, from there, with creating a storage pool within storage spaces, I’m going to go down here where it says virtual disk. I’m going to click to create a virtual disk. So, it’s going to bring up this little screen here and then I’m going to click to create a storage pool. So, it’s going to bring up the storage pool wizard from there. I’m going to click Next. All right.

So, I specify a storage pool name and subsystem. I’m just going to call it Pool1. All right. Because as always, I’m so original with my names. All right. I’m specifying the NYC-SVR1. This is where it’s going to go. Click Next. And it says, “All right, you’ve got the three disk. Which disk would you like to make a part of this?” So, I’m going to make all three. But notice it gives me an allocation option. And so I can do automatic, which is going to let the system manage it. I can do hot spare, which is basically going to make it where I can make a disk, just a spare disk. It’s not going to be used unless I run out of space on these disk. So, it’s just a spare disk. It would use it if it needed to. Or I can do manual, which allows me to just manually decide, okay, I’m going to use this space or I’m not going to use this space. So I’m going to do automatic. Let the system handle it. That’s what Microsoft recommends anyway. And then we’re going to click to create. And at that point it is going to create the little storage pool. And we now officially have our storage pool created.

Now, we’re going to click on Pool1 here. And then to create, we’re going to create a virtual disk by clicking here on this link. So, all right, specify the pool you’re wanting. I’m going to create this virtual disk on this pool. All right. And then from there, I’m going to click Next. All right. It’s asking me what I want to call this virtual disk. So, let’s call this just kind of give it a name. I’m going to say important data disk. All right. I’m going to click Next. All right. It says, EnclosureAwareness stores copies of your data on separate storage enclosures (JBODs). Just a bunch of disk if you don’t know what that acronym stands for. Helping to protect your data. We don’t have what is known as enclosures set up right here. And this gets into having the right hardware to do this. I’m not going to expand on that right now, but I’m going to click Next.

So, from there, we have an option of just doing simple volume. All right, we can do mirroring or we can do parity. All right? So, simple volume is just going to create a volume on the disk and it’ll let you use the max amount of space that you want. But the problem is you don’t get any redundancy with a simple volume. You can do mirroring. A lot of people know that as RAID1 (redundant array of independent disk), or the older acronym which is redundant array of inexpensive disk, whichever one you prefer, but ultimately mirroring will require you to have a couple of disk. All right. In order to do that, if you’re just going to do what’s called a two-way mirror. Now you can do a three-way mirror. And if you but you do have to have at least five disk. Five disk in order to do that. So, two-way mirror is going to require you to have two disk, three-way mirror is going to require you to have five disk and you’re going to get a lot of fault tolerance out of that two-way mirror. It just simply means it’s going to replicate all your data to both disk simultaneously. If one disk fails, you’ve got another disk going and then three-way mirror, obviously you could lose two disks and still be in business.

Optionally, you can also go with parity. Now, with parity you have to have three disk to use parity and with parity you can also kind of do the same thing. You can provide extra redundancy. If you lose a disk, you’re still in business at that point. If you want to do a two-disk failure system, you can have if you have up to seven disk, you could lose two disks and still be up and going. So, a lot of people would call that RAID 5 and RAID 6, RAID 5 being just what we call striping with parity, which was a three-disk system. And then, of course, you could do RAID 6, which was in this case you have to have seven disk in order to handle that. All right. It’s a little different than hardware based raid, though. The rules are a little different with hardware based RAIDs. You may say, “Well, John, that’s not the same as hardware based, because hardware based didn’t require all that.’ All right. And so it is a little different for hardware. I’m going to go with mirroring and we’re going to click Next and then I can do thin or fixed provisioning.

So, what’s the difference here? Well, if you remember dynamically expanding disk, it’s the same idea in Hyper-V disk. Thin provisioning is going to make it where it’s not going to use up all of your space. It will use it as needed and not actually consume the full amount of space.

And then fixed provisioning is going to go ahead and just use all the space. So, fixed provisioning is going to give you the best performance because the problem within provisioning is as it grows, it kind of slows things down as it’s growing. So, to get the best performance, you’d want to go with fixed provision. But in a lab environment like I’m in, thin provisioning is good.

So, I’m going to go with thin provisioning and then I’m going to click Next. From there, I’m going to specify what size disk I want. I’m going to do 60 gigs and then I’m going to click Next. And at that point I’m going to click to create. Okay, so we’re letting that run and we’ve now officially got it. We’re going to hit Close.

So, then the next thing is it’s going to ask you to go ahead and set up your volume that’s going to be on that disk. So, I’m going to click Next. All right. So, there it is, important data disk, which is a 60 gig virtual disk that’s actually spanning three actual disk in a mirror format. We’re going to click Next. And then the volume size is 60 gigs. We’ll click Next, specify the drive letter. How about I do the letter M for Mirror? All right. We’ll click Next. And then which file system do you want? I’m going to go ahead and we’ll say Important Data. All right. Click Next and create. And we’ve now created our volume on our disk. All right.

So, let’s just going through the process of formatting and all that. We’ll hit Close. And if we open up File Explorer, we should be able to see that disk is available. All right, so we can store data on it. So, now that’s one that’s one virtual disk we’ve created. If we wanted, we could actually go and create another virtual disk because we’ve got extra space, right? So we could go through here. We’ll just call it extra disk just for quick name. And then from there, maybe, we do simple or parity, whatever we want to do then or fixed next specify. We’ll do about 30 gigs this time. Next and Create. And it’s now creating that other. So, notice I’ve created basically a storage pool. And on the storage pool, I’ve created virtual disk that I can then go through and format. Right. So, this Disk5. So the system sees it as an actual virtual disk. 30 gigs will give it drive letter E. Next. And this time we’ll do NTFS. All right, change it up a little bit. Click Create. We are now creating our next volume here.

So, as you can see, very easy to use storage pools in regards to Windows Server. And we should be able to go and verify that this is there. And yeah, there it is. So, we’ve got our extra disk here.

So, again, very easy to set up this and just gives you a little bit more control over your disk. And I will say that in the real world, people generally like to lean towards real hardware rate because this is a software based rate solution essentially, so that if a disk fails, you can keep going and recover and of course you’ll get them, you’ll get messy, you’ll get a message if that was to ever happen. Of course, again, in the real world, I think most people agree hardware based rate systems are better, but this is a software based, redundant system that will allow you to set up, run an array of independent disk on a server without having to have any additional hardware other than the drives.

64. Understanding storage spaces direct

I’d like to now talk about a feature called Storage Spaces Direct. So, storage spaces direct is a capability that we can use in regards to server that will essentially involve the concept of storage spaces, which has been in server for a while now, and we can utilize those storage spaces to combine a bunch of disks together that can then be used in term with a cluster, right? So it’s what they call a software defined storage solution. All right. You can combine up to is pretty much as many drives as you want and what’s called a JBOD configuration, just a bunch of disk. And you can have a cluster of physical servers anywhere from two to up to 16 as part of this.

So, you can take all these disk, you know, let’s say Server1 has three disk in, server two has five disk, and you can connect them all together. You can get all eight disks together. They can be different sizes, they can be different types, hard disk drives, solid state drives, whatever, and you can combine them together into a storage pool, basically. And then from there, that storage pool can act as like a Scale-Out File Server file share Cluster Shared Volume for your failover cluster.

So, your failover clustered servers can use those drives as a single virtual disk. And it can it can span all of those drives and be any size you want. And because you’re using storage spaces and all that, you get redundancy with it. You get fault tolerance, automatic fault tolerance. So, a disk could fail or two disks could fail. As long as you understand the concept of storage spaces, which you should, then essentially you’ve got you’ve got redundancy, you’ve got fault tolerance and it’s easy to scale as well because you could add more disk pretty much any time you want and throw it into the pool and it’ll extend the pool out. You got more space. All right. So, it makes it very easy to do this. And the other nice thing about it is when you store space direct in regards to a cluster, it’s always going to go with the fastest disks first for storage. So, it stretches out to the slower disk based on performance as it needs to. So, those slower disks may not even be doing anything until you really start filling up your space. So, the usages of this is you can utilize this for your clustered solutions, you can use it for file storage, so you can store files, databases, things like that, or it can be used for virtual machines. So, you could have guest operating systems that are sitting on the storage space, direct volume, so you can kind of mix and match just like you would any other type of cluster where you have a virtual disk or whatever it may be. All right. Utilizes a bunch of features to make this happen. Failover clustering, obviously the Cluster Shared Volume ability, which means you can link the server multiple servers to the same disk for the scale outside of it. You use a server message block if you’re going to store data in a shared folder. And then of course, the biggest, most important feature is storage spaces. Storage spaces is really what makes this work. Of course, you can probably guess that by its name, right? Storage space Direct. So, this kind of give you a little bit of a visual here. If you look, you can see at the bottom of the screen you’ll see you have local hard disk drives and some solid state drives and you have three different physical servers in this case and three different physical servers have their own disk. Right. A couple a couple of hard disk drives and solid state drives in each. And basically they can be different sizes, different speeds. You can combine them all together into that storage pool. All right. And then what happens is when you connect your cluster into all this and set this up, you get it uses what’s known as a software storage bus technology that Microsoft created for this that allows the communication to occur between all of the different disks as part of the storage space. And then from there, as you move up to the top there you see the storage space is spanning it all. You have that clustered shared. They mentioned ReFS volumes for Resilient File System, and they’ve got a server message file share there that is \\FileServer\Share.

So, they’ve got essentially a Scale-Out File Server of scale out file share, go in their SMB server message block. And so they can they can have that data stored on in those different folders there. It could be you’ve got multiple folders there storing different things involving your failover cluster. The key benefits here, it’s easy. It’s simple to set up. You know, going through the clicks for this is takes about 15 minutes max. It works with Windows Server. It also works with their cloud based Azure Stack HCI technology. You get high performance. They tell you it can you because you can support all the different types of drives, solid state, all that. They tell you the storage space can exceed 13.7 million IOPS. That’s the input outputs per second, right? You get automatic fault tolerance, right, with the storage spaces. The more disk you got together, you can have parity and all that going. You have resource efficiency. The system will manage the efficiency and work in regards to your server to try to control CPU memory, all that.

So, it’s got a monitor and management capability on there, speaking of which, with management capabilities, you also can utilize the storage quality of service QoS controls, which essentially means that we can control bandwidth prioritization if we want, of traffic flowing in and out of the various servers. Oh, and speaking of which, within saying that one of the great things about storage space is that you don’t really have to you don’t have any special equipment here. Really. You literally could have three servers with X amount of hard drives inside of it and just a cable connected to a Network Fibre cable or Ethernet, you know, just some kind of Ethernet based connection. It doesn’t even matter. You’re not connecting the servers together with any kind of special like iSCSI cable or anything like that. It literally is just connected to the network. And then finally, scalability. They’ll let you have up to 16 servers and 400 drives. So, you’re taeing a lot of drives and you can have up to four petabytes worth of space. So, you’re taking a lot of space.

So, that is the idea of storage space direct. It’s a nice little technology. It sits right on top of the storage space technology that we’ve had in server for a while now.