57:02 Webinar
Migrating and Managing Nutanix Workloads on Everpure FlashArray
Let’s move past the slide deck and get into the Nutanix and Everpure solution. Join two of our senior technical experts for a live, end-to-end demonstration of this new integrated solution.
This webinar first aired on January 27, 2026
The first 5 minute(s) of our recorded Webinars are open; however, if you are enjoying them, we’ll ask for a little information to finish watching.
Click to View Transcript
All right. Looks like we're live. Hey, good morning, good afternoon to everybody out there. Thanks for joining us today. My name's Don Foreman. I'm a senior technical evangelist for Pure Storage, and I wanted to welcome you to an
expert-led demo today on a very, very hot topic, which is migrating and managing Nutanix workloads on Pure Storage FlashArray. I'm sure that if you are have been close to Pure as a customer or a prospect or anybody out in the IT field, you've been following what we've been doing at Pure Storage with Nutanix in a joint engineering project that sees them using their Acropolis Hypervisor and
their management, leveraging our storage on the back end, which is definitely a shift in the market to some degree for them with the hyperconverged stuff, which is still a very relevant solution for them, but they are definitely expanding what they're able to do, and their partnership with us at Pure is a great example of what that means. I'm joined today by Kyle Grossmiller, principal technical product
specialist from Pure. Kyle, say hi to everybody. Let us know who you are and what you're doing here. Yeah. Thanks a lot, Don, and, and, and thanks everybody for joining today.
Looks like we got a great crowd. Looking forward to diving into some of the really cool stuff we've been working on. Yeah, I'm a product specialist. I've been with Pure, almost actually 11 years to the day. I think February 1st was my start date at Pure, so been a fun journey and a good time.
And yeah, looking forward to, to digging into it. One housekeeping note, there's a Q&A box in the bottom of your screen, so please use that for Q&A, and, whoever's not presenting Don and I will endeavor to answer those while we're going through the presentation. If it's a really good question, we'll save it to the end and, and answer it then.
But we'll, we'll save about 15 minutes or so at the end for, for Q&A. But yeah, l- I, I say let's, let's get into it, Don. We got a lot of really cool stuff to get into today. Yeah, we do. Lots of slides with really small point font too.
So I'm looking forward to reading those verbatim. Just kidding, actually. The slides are here for color and flavor really. So as we know, there are challenges in the market with virtualized infrastructure. We, we don't necessarily have to revisit this very much.
We know that there's pricing pressure that's been brought about based on Broadcom's acquisition of VMware, and I also think that we're just at a natural inflection point as it relates to what it means to manage into virtualized infrastructure for virtualized operating systems. And I think this disruption is going to be going on for a while as people take a look at
their existing contracts with VMware and their other providers, and really take a good look at what they might need to think about doing as they move forward, right? And I've said in the past that this move with, Broadcom buying VMware and what the market is seeing with virtualization, we're, we're actually seeing a democratization of technology here, that, y- you know, for the longest time in the virtualization space, the
idea was you can have any color as long as it's black, right? The old Henry Ford thing. And in the case of virtualization, it was always VMware, and it's an excellent product, but there is now an opportunity for other options to be considered in the space, and I think that inflection point is happening right now.
And what we've done with Nutanix and Pure together in this joint engineering project is a very, very good example of what that future looks like. So the good part is, is today we're gonna talk a little bit about what it means, what it looks like, why we did it, and then turn the keys over to Kyle and give real-world examples on what it does when you migrate workloads, manage them,
do the usual care and feeding on them through Nutanix Prism and things like that. So we've talked about it already. You know, what are we building? Well, what we're talking about here with the Nutanix solution is us combining our best-of-breed technologies together to provide an alternate virtualization platform.
On the hypervisor side and the VM management side, you have Nutanix, with its Acropolis Hypervisor, as well as its Prism Management, which does all of the managing of the infrastructure as well as the virtual machines themselves. And on the back end, we've got Acropolis Hypervisor talking to a Pure Storage FlashArray.
And the great news about it, and this recently announced, we are now supporting back-end array to not just be FlashArray//XL as well as FlashArray//X, we're also supporting FlashArray//C now as an opportunity as well. So for those second-tier virtualized workloads and things along those lines, a lot of optionality.
And the best way to envision what this technology or this solution looks like is if you're running an ESX cluster that has ESX hosts that are talking to third-party storage for shared storage for vMotion and all that stuff you do, same thing here. The Acropolis Hypervisor is talking to Pure Storage and using our volumes and our shared storage in order to provide the virtual machines themselves.
Again, we're talking about democratizing the idea of operating system virtualization. So VMware, while it continues to have a very good solution, this solution that we've done with Nutanix provides a very, very excellent alternative for people out there who are looking to either modernize their virtualization or take a look at what the future might hold for a more cloud-oriented or cost-conscious type of
solution.On the Nutanix side, they've been doing this for a very long time, and I used to work at Nutanix before I came to Pure, so I completely understand what their stack is all about. And for the longest time, in their hyperconverged infrastructure state, they were leaders of it. They continue to be leaders of it because it's a very viable solution, but they've
over time to really talk about their platform as really being the stepping everything like a cloud, whether it's on-prem, whether it's out on somebody else's heavy metal in some data center somewhere. The idea of Nutanix is they start with management first, and then the details of the virtualization and what you're managing become kind of the supporting
role to that whole thing. And that's an important thing to think about because as we continue to expect with less in the data center, which is the usual pace, Nutanix really emerges as somebody who, very much similar to Pure, believes that simplicity and elegance is really the secret to managing a very, very effective data center, and that's really what Nutanix
their side with their hypervisor, with Prism, and everything else that they've done in their ecosystem. And this choice of what we're talking about here, pairing up Nutanix with Pure Storage, really brings these kinds of advantages into the equation, right? There's simplicity, agility, and security.
Their product as far as HV is concerned is very, very battle tested as a hypervisor. Their management is well, above and beyond what many competitors provide as far as virtualization management is concerned, with all kinds of built-in capabilities like micro-segmentation is included when you use AHV. And the best part about all of it is that it's only going to get better over time with
regards to how their product is architected for security and how ours is also architected for security. High performance and scalability. Obviously, not everybody is lucky enough to have a virtualization space where the compute needs and the storage needs scale together, so adding nodes to be able to grow that
capability is perfect for well-balanced virtualized workloads. But many virtualized workloads are out of balance, meaning there are There's less compute needed and more storage needed. Very storage-centric environments, as data continues to grow the way it is. And that's what this really means with Nutanix pairing up with Pure is that with this option,
we now have the ability to scale, compute, storage independently of each other, which is very, very valuable for looking forward as to how you want to deploy and manage your workloads moving forward into the next 10 years or so. And then this is DeepReduce, and Kyle, you were right in the middle of all this when we did it. It This, this is not a, kind of a wink,
checkbox, loose type of list of requirements that we're like, "Hey, This is really I mean, these are API integrations that, that you were a part of with this that we shared notes on and we, w- we continue to, to wire together, correct? Y- yeah, absolutely. I mean, this is a several-year joint engineering project between Nutanix and Pure.
And, and honestly, I, I'm gonna feel a bit bad when we start going through these demos because i- in a sense, it's, it is that simple, right? It's, it's, it's really simple to use, and really we've endeavored to make everything that you need to manage done within the Prism Central and Prism Element ecosystem, which you'll see in a little bit here.
But, but as that saying goes, right, simple is hard. So there But I, I don't want to belittle the fact that there was a lot of, a lot of really hardcore engineering work that went into this integration. Yeah, and I think that's important, right? That simple is hard, but simple is also necessary, and we were born with that in our
DNA at Pure Storage. We often get into conversations, why have nerd knobs on our storage arrays and things like that, and Nutanix is very similar with regards to Prism and what it does in managing its virtualized environment. So yes, you're going to see a much simpler approach to things, but know that the hard work that was done behind the scenes was meant to do that by design.
So again, why us and Nutanix? Well, these are all of your benefits right here. Obviously, high performance and scale, we talked about that already. Granular storage control, we're gonna talk a little bit more about that as it relates to the difference in how we approached providing external storage up to
the hypervisor in Nutanix. And all of it really is meant to do the last three things, right? Preserve your existing storage investments, streamline IT operations, and enhance your security posture. We talked a little bit about how secure Nutanix is on their side.
As a matter of fact, they even, they, they ship their stuff with a DISA STIG as its standard, so out of the box they are as secure as they possibly could be for anybody who buys their solution. And Pure on the other side, we have all of the supporting technology for that as well. But preserving existing investments is probably one of the most interesting ones
because it was a tough one when I was an SE there, and you would sell somebody on the hyperconverged idea, and then they'd say, "Well, what do I do about all this storage here?" Well, you could use it for test dev or you could use it as a boat anchor or something like that. And, fortunately, with this solution, if you are a Pure Storage customer or you are considering buying Pure as a solution
possibility for you, especially if you are interested in what Nutanix can do with its Acropolis Hypervisor and everythingSo what is the Acropolis Hypervisor? Well, historically, and under the hood, it really is just a form of KVM. And the way the story goes is one of Nutanix's first customers wanted to use KVM, but KVM to them wasn't what they considered enterprise-ready, and its management was
challenging, as we know. And they came to Nutanix and they said, "Can you do something about this?" And Nutanix said, Sure." They threw some engineers in a room, they did what they needed to do, and out popped the Acropolis Hypervisor. And I think one of the things that I've read about this democratization of virtualization
is somebody did say that we are now entering the age of KVM, that the KVM something that people will seriously take a look at now since there are more choices out there. And I think AHV is a great example of what that means when it becomes the year or the decade of AHV, where- wherever you want to look at it.
You can see on the right there, it has really stood the test of time. It's been around for over 10 years, and you can see that everything that they've done with KVM plays right into what people are looking for in an alternative hypervisor. Integrated DR, micro-segmentation, we're gonna talk about Flow, enterprise validation, performance enhancements.
A lot of the things that Nutanix has poured into the Acropolis Hypervisor is meant to make it a very, very viable alternative to what VMware was able to do with ESX, and it's battle-tested. It's been out there for over 10 years. So what does it look like with this integration between us, right?
You can see the key capabilities over there as to where things line up. But on the left, let's take a look visually what it looks like. In the case of the Nutanix solution with Pure, what we do is we take the Nutanix cloud infrastructure cluster, and this is a cluster that is not operating as a hyper-converged cluster.
This is simply a collection of Nutanix compute-only nodes that are running the Acropolis Hypervisor. And as you know, Nutanix has this concept of a controller VM, which is the CVM that you see in the gray box there. That controller virtual machine is still in play here, that the Acropolis Hypervisor still
connects to the CVM, and that CVM out the back has a channel over NVMe, over TCP, to the FlashArray. And by doing that, there are some real advantages. First off, the CVM is still very, very critical for reporting, for stack performance, and things along those lines with Nutanix, so it needs to be in, still in play.
But the good part is, is on the back end between the CVM and the Pure Storage array, we've taken the Ethernet approach because Ethernet is very prevalent in the Nutanix space, and we're big believers in Ethernet as well. So when you look at the back end of the CVM talking to the FlashArray, we're using NVMe over TCP. It's recommended 25 GB/s as far as
your network speed is concerned. Some minor possibilities to do 10, but aim for 25 on that. What's different from the HCI model here, obviously, is that Pure takes on the storage functions. So all the things that you would expect Nutanix to do from a storage layer,
data-at-rest encryption, deduplication, reduction, things along those lines, we're actually managing into it on the Pure Storage side of things. And the best part about that is, is that we're merging the best-of-breed capabilities together into this whole thing. There's a joint support model.
You'll be able to manage into the Nutanix side with your Lifecycle Manager. You'll be able to manage into Pure's FlashArray updates through Pure1 and things along those lines. So there's a lot of familiarity here, but at the same time, when we put it all together, we insisted on a lot of simplicity as well.
Real quickly, and Kyle, I'm gonna have you talk a little bit about this if you don't mind. One of the really amazing parts of this joint engineered solution is really our choice in how we wanted to deliver storage up to the hypervisor for the virtual machines. The easy way would've been just one big fat data repository, right? One big data store that we would be able to throw virtual hard disks in, but we took a
different approach, right? We, we really like the vVOL concept, so talk through what this means that, that they're seeing here. Yeah, 100%. So on an AHV node, you have one or more VMs running on it, obviously, and those VMs are comprised of vdisks. The vdisk has a data volume, you know, you're probably used to, and then it also has,
kind of under the covers, a metadata volume. And now why that's important is that to use, Nutanix change block tracking Prism Central and Prism Element. So that's kind of the secret sauce under the covers that enables us to leverage FlashArray snapshots, but also use Nutanix CBD CBT, excuse me, in tandem, to,
to be able to move forward and backward in time. And that also applies to vdisks, ISO CD-ROMs, TPM modules, so, so all of those kind of disks that are utilized by Nutanix VMs are used here. And to your point, gives you that granular per volume control, right, over each of your VMs, and the ability to kind of manage it at a per VM level.
You know Or excuse me, at a per volume level. You know, maybe I, I need a certain database volume, but I don't need that entire VM. So this opens up the capability to be to use that kind of per volume control, within your Nutanix ecosystemYeah, and I think the awesome part about sticking with change block tracking on the Nutanix side is that it doesn't affect their integrations for DR and
stuff with, with all of their partners, like Veeam and Haiku and all of those So this was the best way to integrate the two of us together without non-disruptive what the existing vendors brought to the table as far as external integration capabilities are concerned. So it, it really is a win-win scenario for what we did as far as how the back-end storage
is presented to the hypervisor itself. So Kyle, here we go. It's time for you to, to jump into the demo seat. I'm gonna hit play here and- Yeah I'm gonna let you do your thing on- Burn my keys what this really looks like. Awesome. So- Yeah, g- channel your inner geek.
Yes, exactly. So bear in mind we've got FlashArray stood up, we've got Nutanix stood up, and now we're just going to connect the two of them. So the first thing I'm doing here is I'm creating a new concept in Pure called a Pure realm, right? So a realm you can think of as kind of like a federated environment on the FlashArray, and that's to basically limit access, right,
between the Nutanix stuff and the other stuff running on the FlashArray. So I've created the realm via s- I'm doing all this via CLI. You can do it in the GUI as well. I fat-fingered the pod name there. But we create the realm, and now we're creating a pod, right?
A pod is a collection of hosts, volumes, and other things that's inside of this realm. One of the really cool things we can do with a realm is we can essentially, as I mentioned earlier, federate access, right? So what I'm doing here is I'm actually creating a management access policy for a specific user, on the, on the, on the FlashArray.
So I'm creating this admin policy, I'm assigning it to this realm, and I'm creating a specific Nutanix user that only has access to this realm and this policy. And really, this is all you gotta do with respect to setting up the FlashArray. So here I'm creating this new user, Nutanix PSTG user. And then next I'm assigning it to this access policy.
Oops, and again, I messed that up . So forgot my double hyphen there. So I've created a user and I've assigned it that policy, giving it that username. Now I'm logging out, and I'm just gonna make sure that I can log in as this new Nutanix, specific user that I've created.
Now, when I log in here, really what I'm just checking is to make sure that I can see the realm that I created that I want this user to have access to and nothing else, and that's confirmed here. This array actually has multiple realms on it. And if we go into the GUI, we can see it all here.
Another thing you can do optionally is I can assign a quota. So I want this Nutanix realm to only have 80 TiB space so it doesn't overrun of the array. Now I'm gonna hop over to Prism Element, that's where we set this up, and there's a little bit of networking config that we're gonna do first. I'm gonna create a virtual switch.
As Don mentioned earlier, we're using NVMe over Fabric TCP. That's what we support with this. That's a modern fabric that we're using. We recommend using jumbo frames, so I'm setting it to 9000 MTU. And what I'm setting up here is, ActiveCluster with MAC pinning, right?
So I'm gonna set two physical NICs, for use with this. I'm just gonna go down the list here of my five, five-node cluster. Five is what we need for a, default cluster. And I'm gonna create the virtual switch. So obviously skipping ahead a little bit here just in the interest of saving time.
So my vSwitch has been configured, active-active with MAC pinning. That does transparent failover on the host side. And now I'm gonna create a couple of Nutanix interfaces, right? So this is NVMe over Fabric A. So we're using that kinda classic dual VLAN a- approach for this.
VLAN A, so I'm generating an IP pool so that I actually get, IPs on the NVMe over Fabric TCP network that I'm using here. So there's my IP r- first IP range. I'm going to select that. And then we can see that this is important, right?
There's this external storage flag that we set. I'm matching the MTU to 9000. And then I'm just gonna click Save. Takes about five minutes to build this virtual switch. Now I'm gonna repeat it quickly and I'm gonna build my B VLAN, right, A, B VLAN, so that I
have that kind of baked-in failover. You could use more than two VLANs if just kinda depends on your environment and what you're using. But generally two is, two VLANs is what we recommend for the minimum. So I'm gonna go ahead and click Save here.
I have created my two interfaces. I've created my virtual switch. That's kind of what you need to do in the back end from a networking perspective within Nutanix Prism Element. And now I'm gonna go and set up external storage, within Prism Element.
So I'm gonna go attach external storage, select the FlashArray. We can see there's my virtual switch and my two interfaces I created earlier. I'm just gonna give this the array name. I'm gonna give it the VIP for my mgmt. This is my Nutanix user that I created on the FlashArray at the start of this demo, and I'm
just gonna enter my credentials here. And then I'm gonna create that realm and that pod, that I entered in at the beginning of this demo as well. So there we can see that, everything that I built at the start of this demo. Takes a couple minutes for it to do this.
And then everything that I need it to do, with respect on the FlashArray, is automatically configured by Nutanix Prism Element. This is that kind of deep engineering work that I mentioned earlier. So what this does is it creates my Nutanix CVM objects on the FlashArray. We can see there's my FlashArray showing up, under a STaaS as a storage container.
There it is as external storage. I can click into itOpen up my storage management here, and then if I actually click back into the realm and then into the pod, we can see that it's starting to populate with, again, all the CVM stuff, and then this is where all of my Vdisks, my metadata disks, all those things I talked about earlier will live, on this thing.
So I know that was fast. We have a support page that actually shows this in a little bit more detail, a little slower, so I invite you to, to check that out as well. Um- Que- question for you on the realms thing, Kyle. Yeah. Y- you know, it, it's a brilliant move obviously because we're talking multi-tenancy-
Yeah for the different clusters themselves. Since we are using realms, can you clarify if we will manage and monitor into QOS between realms? Like, if, if one cluster starts to overrun the other, will we ratchet it back, or is that something that's on the roadmap, or...?
Yeah, it's, it's more a roadmap. The, the QoS right now is, is primarily capacity, but we will be extending QoS, and it's kind of soft limits on the QOS today. You can do volume limits and all that good stuff, but it is softer limits. But yeah, we're always looking at ways to kind of optimize, the performance especially around
this, for sure. Yeah, and I think, y- y I think the thing, too, that we need to state here, too, is that this is 1.0, right? We're, we're- Yeah looking at 1.0 of this thing, and there's a very aggressive roadmap ahead, not just to additional capabilities, but additional things that w- are, are being supported on the Nutanix side, right?
They just announced the Kubernetes solution they have is supported in this environment as well. So my advice would be to keep a very close eye on whatever feed you're following as it relates to this solution because the tweaks and enhancements are coming fast and furious because of everything everybody wants to see in this thing.
So from a deployment consideration thing as far as initial release is concerned, these are the three Nutanix nodes that are supported right now. I know some of you out there are a little surprised that the Supermicro OEM version of the Nutanix hardware isn't on here. But the idea here with regards to these three being out there is that these are
three v- th- these are the big three vendors that everybody is looking at next-generation compute stuff anyway. So obviously Cisco, we're able to do it either with the Cisco UCS stuff on its own or we now support a FlashStack version that runs Nutanix as well. We just announced that as a Cisco-validated design.
I believe it was last week we announced that. We're also able to work with the, the Dell models as far as PowerEdge is concerned, and then HPE with the ProLiants as well. The big thing I would say here, and the bottom right-hand corner there's a great hyperlink as it relates to what the compatibility list might look like for
hardware, because I think those things are going to change over time as well. An important note to, to see as a common thread, by the way, with HPE and Cisco, you will need an M2 RAID 1, at least 480 GB. That's to be able to boot the hypervisor and things like that inside of the node itself. But once that's there and it's configured through Foundation, it's pretty much autopilot
from there as far as hardware management is concerned. Out of the gate, too, there are some supported topologies to think about. Remember, we're only talking NCI compute only here. You can't just take a FlashArray and plug it into an HCI cluster that is being managed by Nutanix.
This is meant to be a viable, almost apples for apples alternative to what an ESX cluster using third-party storage looks like. So these deployment scenarios, these solution topologies, shouldn't come as any surprise. There's the chance of a one-to-one, obviously, with a Nutanix cluster relationship to a Pure Storage FlashArray, or you can have multiple Nutanix compute-only
clusters talking to the same FlashArray as well. In that case, each cluster, as you saw in Kyle's demo, would get its own realm and be able to manage into its storage that's dedicated to it as the cluster without seeing the other realms itself. Disaster recovery and business continuity, that is coming directly from Nutanix as well.
Right now, the only supported replication is Async. I know people are wondering when synchronous is gonna happen, when Metro is going to happen. Those are all things on the roadmap, I promise you, but for us, getting the solution out the door and being able to get people to start using it, we opted in with just Async as the only option right now, and it's being managed from Nutanix.
So the CVM is actually taking care of the replication, the change block tracking, and all of that stuff to the other Nutanix cluster. One of the advantages I think is really good about Nutanix managing into the replication possibilities is that your alternate location that you're replicating to doesn't have to have a FlashArray.
Now, we'd love you to have a FlashArray in your disaster recovery location, but if your CFO and your CIO and everybody are budget, you can just have a dedicated Nutanix cluster over on the other side operating in HCI mode that is simply taking the Async replication from the primary, which is tied to the Pure Storage FlashArray itself. We talked about change block tracking still being in place as far as the CVMs managing
into that, and that's the best part. If you are looking at Nutanix from an, an integration perspective with third-party providers, you can see all of those that are down there support Nutanix from a disaster recovery and business continuity from a change block And again, just like everything else in this solution, keep an eye on things because these
options are going to expand and change over time.All right, Kyle. So now I think we're talking about migrating, right? We're, I think we're at the demo where we're gonna take some VMs from somewhere in an old ESX thing and, and use Nutanix Move, correct? Yeah, exactly. So we are, As we mentioned, you know, yeah,
migration obviously is a huge component of, of Nutanix and, you know, given as you, as you mentioned earlier as well, a lot of the upheaval with, with Broadcom, thought it would make sense to kind of showcase how exactly can one move from VMware specifically, to AHV, with this joint integration. So just, yeah, keep in mind that we've got a, a VMware vCenter set up.
We'll have a few VMs that we've kind of targeted. If, if you wanna hit play, we can, we can get this rolling. And if you're not familiar with Nutanix Move, I mean, it's actually a really great offering. It's, it's very easy to use. This is another case where I was kind of struggling to come up with enough to put into
this demo, but I figure there's probably enough folks on this call who are, or on this webinar, who are not familiar, with Nutanix Move that it would be worthwhile to give just a quick demo of what it is and really just how easy it is to move stuff around. Yeah. One of the cav- couple of caveats I wanna throw in before we hit play on this. Sure.
The first one I learned a couple of weeks ago, which I thought was really cool, is that Nutanix Move, which is included in the functionality of Prism, not just does the conversion of the VMs over into the format it needs to be for AHV to run, but it also has the ability, and Kyle, I know you know this, to pull over NSX rules into- Mm-hmm the Flow part of Nutanix, which is the micro-segmentation part.
So when you're thinking about a migration from older legacy to new, you have to be able to take those network things with you as well, and Move is able to do that, convert the NSX stuff to the Nutanix Flow stuff, which I thought was very interesting. The other caveat I wanna make sure everybody understands is, yes, we're showing you Move and we're showing you how it works, but everybody out there managing virtualized data
centers knows that switching from something from one hypervisor to another isn't like the Raiders of the Lost Ark thing, where you swap the bag of sand for the gold idol, right? There has to be a lot of planning, there has to be a lot of consideration into workflows, security, and things along those lines. So don't let your brain oversimplify what you're seeing.
Yes, the Move part is very, very straightforward, but it's only one piece of a bigger pie. Exactly. It's all yours, Kyle. Thanks, man. And, and, yeah, as we can see, so I've got source and I've got target, right?
Super straightforward. So I'm going to my source VMware environment. I've got about 10 VMs here that I've targeted for this demo migration. One of the neat things is that we don't really care about the data store, right, that any of these VMs are on. They can be on virtual volumes, they can be on NFS, they can be on VMFS.
That doesn't surface within Move. The only thing that surfaces is y- the, VMs themselves. Now, of course, the array that I'm moving off of needs to have connectivity to move. That's an important caveat here. But, you know, besides that, that's it.
And then it We can see there too, as I was setting up the migration plan, really the only thing I need to do on the AHV side is say, I wanna use the FlashArray container, which the storage container, which if you recall, is what we set up in the previous demo video. So now I'm just clicking down, grabbing these 10 different VMs, that I wanna move over, via Move, and then I'm gonna select, my target network, which in this case I'm using DHCP.
You can optionally use a test network, you know, if you wanna make sure everything works. There's a couple of different preparation modes. There's a manual mode where if you wanna It's a really critical VM and you wanna keep your hands on it, you can run this prep script, and to basically run PowerShell to install Nutanix guest tools. But then there's also an
automatic mode as well. In this case, I'm gonna use automatic. My target network is DHCP. And then I'm just gonna enter my credentials, for both my Windows VMs as well as my Linux VMs.
Now, there is, of course, a way to granularly select individual VMs. Maybe I've got different usernames and passwords for each of these. Maybe I wanna do DHCP on some, but static IP on the other. So it gives you that kind of granular control on a per VM basis. But this is simple and fast, so I'm just gonna go quickly here.
Can set priority for your different VMs. You can assign Nutanix category and tags, after they're migrated over. You can change your target properties. And then again, you can, you know, individually control these VMs. Maybe I want one VM to be high priority, something else I want to be lower priority.
So you can kind of go through and individually set these up as well. So as I click through, it'll validate all of the passwords, for these VMs, and then really that's it, right? So you can do minimum of one and up to 100 VMs per migration wave. You can also deploy multiple Move appliances, which, you know, if you
stuff, you can do that here. Move today uses front-end array IO. It does not use Purity replication, so that is something to be aware of. You know, so as you're going through this, you'll wanna maybe do a test migration, make sure that this migration doesn't impact running workloads.
Obviously this integration we fully expect to improve over time and use FlashArray-based replication. But again, today this is what we've got available. So what this does, right, is it runs that PowerShell or that Bash script against your, your, your VM that you're gonna move, pun intended, and then it, actually creates a
VMware-based snapshot of that VM. Vvols-based snapshots are actually array-based snapshots, so that is a more quickly if you want. You can move that VM from VMFS to vVols if you're using it. But VMFS works just fine as well.
So what this does is it takes a VMware-based snapshot Whoops, think we jumped ahead there. Oh, did I do that?It might have I'm sorry It's okay Give me a sec. Let me get you- Yeah, no worries, man forward to where you were. That's, that's close. That's pretty good. That's good. Is that good? Sorry. Yeah, that's good.
N- it happens, man. No, no problem at all. So yeah, it's preparing this VM, and again, this is just using front-end Array IO. It can be on the same array, it can be going from o- one array to another. It's just important that they can all talk to each other via the move appliance. VMs that are in a segregated environment aren't gonna work.
Here we can see it's taking that VM snapshot. It's actually moving that snapshot over from point A to point B. And again, it's using that front-end IO, and it's grabbing each VMDK here. This, I believe, is a I don't know if this is a VMFS or a VVols-based VM. But, you know, based on network availability, size of the VM, you
that will kind of dictate how long it takes for the VM to move from source to target. Can see this array's reading a bunch of data and also writing a bunch of data. This is my target VM. Some VMs are on this VM, or excuse me, on this array. Some VMs are coming from a different FlashArray.
Again, so it, it doesn't care where the, where the VMs exist. It just cares that it can see them in vCenter. You can also do a couple of pretty neat things within move. As I mentioned, it uses VMware-based snapshots. The default duration is 10 minutes.
So after it establis- establishes that baseline snapshot, it will then take a new snapshot every 10 minutes and replicate that data until you're ready to cut over. You can also get into move itself and kind of see, like, what sort of performance am I getting, you know, am I saturating this move appliance. You can do stuff like increase the amount of vCPU and RAM for your move appliance to
potentially get better networking performance. But it's generally recommended to just deploy additional Move appliances if over a bunch of stuff. But anyway, you can get in and get some really good details there. We can see I, I did skip ahead here at this.
You know, it does take a little bit of time for this to get, set up. But now I'm gonna actually cut over this VM, right? So what it does to, for cut over is it will take one final VMware-based snapshot and then cut it over, for the, to, to be set up on AHV. So it takes that final snapshot, and then we hop over finally to AHV, so
final sync, seeding that data onto AHV. It powers off your original VM, in vCenter, and then it brings it over to AHV, which we'll see here in just a second. So now it's actually creating that VM based on that final snapshot data. And then it will Now we can actually click in from move and get into Prism
and see that VM itself. So there's that VM we just moved over. It automatically powers it on, and we're done. That's really all it takes to, to So again, for a free tool, this is really slick, in my opinion, to be able to, to move
this thing from, from VMware to AHV in an automated fashion. So while you were talking about it, I was going through the Q&A, and I did find one that I'd like to answer live, because it's been asked by a couple of people now. Are there any plans on the roadmap for looking at an array that already has ESX-based stuff on it a- and doing some kind of Zero Move Tiering that can
convert the ESX stuff magically to what we're trying to do with the granularity for the storage that we're presenting to AHV without the actual mechanics of move? Is, is that something that we're thinking about? Big time. Yes. Yeah. so, you know, we're It- since it's just
us in a 200-person webinar, you know, definitely something that, that we're on integrating more directly into move. So, so the answer is yes. I think it will start off as being a professional services engagement, but eventually it will be baked into the move appliance. Perfect.
Yeah, I, I know I get that question all the time- Yeah and I, I agree that it would be n- love, it'd be wonderful to be able to do this, like, without moving data around, if possible. Yeah. And again- Yeah, we, we hear you. That's definitely gonna happen, for sure. Yeah. And, and a couple of, just to answer a few of the questions as one group out
chose NVMe over TCP for a very specific reason. It's Nutanix lives and breathes in the ethernet space. We also believe that ethernet is a very strong future as far as storage protocols are concerned. We're not saying that Fiber Channel is a bad thing, just for this solution specifically,
because of the environment and the ecosystem that surrounds it, Fiber choice for us. We felt that the latencies, the throughput, the speeds of everything that you got with NVMe over TCP was a viable part for this solution itself. So real quickly, we're getting into the final parts of it all.
I mentioned this earlier, FlashStack with Nutanix. It was just validated last week as a certified Cisco-validated design. So if you are somebody who's looking for a type of almost turnkey solution as far as Nutanix is concerned with Cisco and Pure, FlashStack is your answer. Again, to answer some questions that are in the Q&A, if you are using Cisco, we are not
supporting boot by SAN for AHV, so you will need local storage on those nodes in order to be able to make it work. A couple of other things to think about here, and none of this is surprising. These are just requirements that you have to come to the table with to, in order to make it work. On the Nutanix side, you can see the
operating system and AHV requirements. On the Pure side, you can see Purity, arrays, which generations are supported. But all of those pieces, I think, are-Pretty expected when you're looking to implement something that's a new jointly engineered thing. We're not gonna engineer to the old stuff, we're gonna engineer to the emerging and, and
newer released stuff. One other caveat to note is that production environments, it's a minimum of five nodes on the compute side for Nutanix to be supported. You can do three, but it's only for test dev environments. You need five for the production.
Yeah. And then snapshots- And the reason for that is that CVM requires RF equals three, just as a FYI. So like, we got- Okay, so you guys are using a replication factor of three. Okay. Yep, yeah. We won't dig into what that means, but the fact you're using RF3 makes sense with five nodes.
Yep. Real quickly we're gonna talk about snapshots in Nutanix and what it means. So we're doing managed snapshots, which will be done through Prism Element. Again, we want administrators working in Nutanix Prism as much as possible with FlashArray, while an important part of this, is a very, very supporting cast member of the
overall solution. So Prism is where we want everybody to be. And you take advantage of the CBT, the change block tracking, and all of that stuff when you do manage snapshots themselves. Which I know for some people who are in the Pure Storage world wish they could do it through our interface and what we do with them, but we need that change block tracking for the other
things that Nutanix is relying on from an integration perspective. And, and yeah- So real quickly, Kyle- just give us a little time, give us a little time there too. That, this, this story will change in the future, so Yeah, abs- absolutely. And, and I think, Kyle, let's do this real quick video that, that shows- Yep what that
looks like, because I'm sure the Pure huggers, the ones who love our stuff, looks like when somebody else does snapshots that our arrays are involved in. Yeah, and this is a, a very purposefully, simple demo. But just, you know, if you're not familiar with Prism, Central Prism Element, this is just to showcase, you know, how my operations would look like within Prism.
So I apologize for, you know, people who are Prism experts. This is very basic. But I just wanted to show that, yeah, the whole point of this, all this integration we've been working on, is we can see there is our FlashArray container. That's actually a FlashArray C60, so we do support the C.
That this is really how simple it is to, to do all this stuff, right? So you really shouldn't need to get into the FlashArray at all. And with respect to snapshots and recovering snapshots and cloning, do everything in Prism. If you take no other lesson from this that, that is it. So in this case, I'm just building a Windows server to 2022.
Again, this is really just business as usual, right? From within Nutanix Prism. I'm just selecting that FlashArray storage container that we can see there that we set up in the initial step, right? That's really all there is to it.
I create the VM, and that's it. And again, in the interest of time, creating a snapshot, I go to data You can use protection groups or protection policies, recovery policies, all those things. And if I click on recovery point here, that's really how I create a snapshot. That is making FlashArray API calls.
So that's actually creating a FlashArray snapshot. Just again, just to kinda prove out that this works, I'm just gonna create a dummy text file here. Again, this is very stupid, very simple but I like to keep it simple. I'm just gonna call this file delete me.
This is on a drive, on a VM running on top of the FlashArray. This is this VM. I'm going to my recovery point, which we can see here, and I'm just gonna revert it, right? And again, this is on the back end copying that volume snapshot, or excuse me, that snapshot to a volume on the FlashArray and overriding it, right?
So this is, this is done here. So now if I go back to the VM, it does power it off, you know, 'cause I am and restoring it. So what I'll do here is I'll just quickly power it back on, and then when we get back into the VM, we can see that, we can launch the console here, and then if I go back to that disk, we will see that it is
Give me a second to log in here. That text file I created, again, just to demonstrate, you know. Imagine that text file's a virus and, and my VMware- Ah has been taken over by a ransomware, right? I can, we can just kinda prove it out. Yep, so there we can see that, that file's gone.
And now last thing, I'm just gonna create a new file here. We'll just call this clone. I'm gonna clone this VM. Again, this is all using FlashArray APIs to update metadata pointers to a new VM. Very simple. You can do this live from the running VM,
which is what I'm doing here, or I from that recovery point that I took. I'm just gonna name this 2, 1 clone. Click next. We can see there's my FlashArray storage container again, and I'm gonna clone it. Again, this takes a matter of seconds.
This is great for very large VMs that I wanted to do test dev instances of. If I click in on this on Prism Central, we can actually navigate directly to that clone VM, power it on again, and then we should see that, that text file called clone, that I created. And we can see this is the clone VM.
It's got all of my disks that I used. And when I launch the console here, we'll log back into it. Again, this is, for, for those of you who are familiar with Prism, I apologize for how easy and basic this is, but again, the, the important thing here is I'm doing everything that I-Usually do in Prism from a daily VM management perspective,
as, as I normally would. There's nothing fancy or special here with respect to the FlashArray. It just, just works out of the box. So if I go back to my F drive, there's my clone volume I created earlier. Perfect. That was a quick demo.
I was trying, man. Yeah, we're, we're getting running short on time, so gotta roll down. Well, it's not like Y- you know, like it's, it, it's one of these things where it's obviously we've worked very hard to make it simple and elegant. And I, I think it's important to note too that, like, we, we really haven't
Nutanix administrators work with things, right? They're, they're able to use all of their tools in everything that they do, it's just that the data is on a different location as opposed to- Yep in a hyper-converged infrastructure type of thing, which is, in my mind, the best of all worlds, right? Because we, we bring a lot of value to the parts where we need to bring value, but we
don't disrupt how Nutanix people do what they do, quite honestly. Real quickly on a, a question that came out of that demo and then we'll, we'll with the rest of the slides, not many left. Somebody asked, and, and this is a roadmap item I'm guessing, is w- w- when we've got deployment scenarios, we've got a 1touch relationship, we have multiple clusters to a
single array. Are we looking on the roadmap at multiple arrays being accessed from a single cluster, a single Nutanix cluster? If I wanted an X and a C, let's say. W- we are. It is, and I'm th- this is more of a question for Nutanix, to be honest, 'cause today they only support a single
external storage container. So- Right I'd, I'd have to defer to them in terms of when that will be on their roadmap. But it is, I mean, obviously something that we would like to do, but that is ultimately a, a question for, for Nutanix today. Yeah, and it's an awesome It, it really does make sense, right?
Cause the path starts with the CVM from AHV, right? So the CVM needs to know where to go if there are multiple arrays, and it can only support one storage container right now in order to be able to do that. So- Yep I'd love to see it happen eventually, and I'm sure it will because 208 people are on here right now very interested in the solution.
So I imagine our aggression in that roadmap is going to be great. Kyle, again, those were great demos. Thank you for that. The move thing, to clarify as well, when you're migrating stuff from move from ESX to AHV, we're saying it's live, right? Like, the VMs are still getting hammered on the ESX side as they're getting moved over to
Nutanix, right? So the only glitch would be the cut-over point itself, correct? Exactly. Yep, 100%. So that there is, yeah, live is a relative term, right? But but yeah, there is, there, there is that cut-over phase for sure, so yeah.
But that does- Yeah, live-ish. We'll, we'll say live-ish Live-ish, but it does power off your kind of source VM, and it does automatically power on your target VM. So to the extent possible it does. So yeah, for those critical w- workloads, you obviously want to do a little bit more
hand-holding and, and pay attention, in terms of, of what that window looks for maintenance to move it from, from ESXi to AHV. Great. Well, so to, to finish things out, again, I, I think this was a great run as far as explaining what the, the interest you know, why we did what we did with this joint engineered solution.
And Kyle, your demos were spot on as highlighting really the simplicity and the elegance of what this can look like if you do implement it in your environment. Again, it takes some planning, it takes some consideration, but it's, it's a very viable battle-tested solution that is a real alternative if you are looking to migrate your virtualization strategy for what's ahead 10 years from now kind of thing.
One shout-out to do towards the end of the presentation here is I know the Q&As are still streaming in, and a lot of my friends behind the scenes are trying to keep up with everything. But what I would suggest is that if you did not get your question answered on this webinar, please join the Pure Storage community.
You can see the link there. You can see the QR code. Our community is highly active right now on this jointly engineered solution, and the best part is the questions that you put into the community are going to get farmed out or seen by people like Kyle, people like me, and people on the Nutanix side.
So there's a very, very joint approach to how we're supporting this in the More than happy to answer any questions in our community that is related to this or else as far as Pure Storage is concerned. So take a look at that hyperlink down there, bookmark it, and join the community if you have not already.
I'm telling you, it's well worth the time. We're very actively involved in it. And that's- Absolutely. And I think, yeah, and I think, Don, I think that the excitement around this is a testament to the number of open questions we didn't get to, so apologies for that.
We'll- It, it's, it's bound to happen we'll definitely send them out. I did, I did one, like, three weeks ago, the poor guy from Nutanix who was trying keep up with it, he got carpal tunnel syndrome from typing. So- Exactly I, I felt really bad, but the interest is tremendous. And my favorite part about it is the is not just tremendous, we sold our
before this thing even went GA. Yep. And we're just getting started. Like I said, this is 1.0 as far as the solution is concerned, and it's really going to be about us pacing ourselves in, in- Yep really, you know, really building out the features and capabilities that everybody's asking for.
Again, one of the things I didn't touch on with this, and Kyle, I know you know this, is if you look at Nutanix and you look at Pure, we have a very customer-centric pedigree. Our NPS scores are very high. We are very, very customer-focused on what we bring to the market and what we do. So merging us together and being able to do this joint engineered solution is a very
natural extension for both of us, quite honestly. And I highly recommend anybody out there, if you didn't get your Q&As taken care of, join the community. Go to our website, look at the resources that we've included with this webinar. And I highly recommend you reach out to your Pure account team, your Nutanix account team,
and your VAR partner to get deeper into what this solution is all about, because we have engineered it with the idea that it is a viable alternative in this swizzling of the virtualization community, and ultimately a very, very strong solution looking forward 10 years from now. So Kyle, any parting thoughts before we close off on here?
No, I appreciate the time, everybody. Thanks for the great interaction and the great Q&A. And, and yeah, like, like, to echo what Don said, thanks, thanks a lot for the time, and, and thank you all for joining. Yeah, it was absolutely great.
Again, I love doing these things, especially when 208 people all flood in with, with Q&A on it. So again, thank you everybody for joining us this morning on this expert-led demo, and we look forward to seeing any and all of you in our community and on future expert-led demos. This is Don Foreman. Thanks again, Kyle.
Thanks for your help, man. Great seeing you. Hope everybody has a great day out there. Cheers.
Don Poorman
Sr. Technical Evangelist, Everpure
Kyle Grossmiller
Principal Technical Product Specialist, Everpure
Let’s move past the slide deck and get into the Nutanix and Everpure solution. Join two of our senior technical experts for a live, end-to-end demonstration of this new integrated solution.
We will simulate a real-world deployment scenario, showing you exactly how to leverage the performance of Everpure FlashArray™ within your Nutanix environment. This is a technical "how-to" designed for architects and admins who want to see the plumbing behind the partnership.
Here’s what we’ll demo live:
- Connectivity and setup: A step-by-step connection of the FlashArray to the Nutanix cluster, ensuring optimal configuration for low-latency workloads.
- Seamless migration: The workflow for migrating active workloads onto the joint solution without breaking a sweat.
- Provisioning in action: Create a Virtual Machine and track the corresponding volume directly on the FlashArray management console.
- Advanced data protection: How to execute and manage high-performance snapshots for instant recovery and data mobility.
We Also Recommend...
Personalize for Me
My Updates
No updates at this time.