57:36 Webinar
Pure Fusion Expert Demo: From Fleet Creation to Policy‑Driven Provisioning
Manual provisioning and reactive management can slow innovation and drain valuable IT time. What if you could manage your enterprise data intelligently?
This webinar first aired on December 16, 2025
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
Greetings, everyone, and welcome to this edition of the expert-led demo webinars. The, this one specifically, the, expert demo from fleet creation to policy-driven provisioning. My name is Mike Nelson, and I'm a senior technical evangelist with Pure Storage. And, my partner in crime today for, for this session, Santhosh, if you'd
like to introduce yourself. Yeah. Thanks, Mike. Good morning, good afternoon, wherever you are. I'm Santhosh. I'm the product marketing manager for Fusion. Hey, Mike, what is fleet here? Are we Pure Storage getting into trucks and vehicle management business?
Yeah, yeah, pretty much. No. Yeah, we'll be talking about Pure Fusion and how, you know, fleet management, c- specifically around policies and workloads, pre-presets. We'll be talking about, you know, things like, remote provisioning, all kinds of fun stuff. So, yeah, it's, it's, We're get- we're getting more into, the, the, the single con-
controlled data plane, if you will. Yeah. I'm glad we clarified that. Sure. So welcome to the presentation. So we are redefining the storage, in a, in a rapidly evolving, data Today we'll explore the fundamental challenges of legacy storage and Pure Fusion
stands as the answer to modern day enterprises. And as I said, my esteemed colleague, Mike Nelson, will walk you through an in-depth demo of Fusion, showing how easy it is to manage, your data. Now, this is not news, right? Enterprise data is growing at an unprecedented rates driven by your AI/ML,
cloud-native apps, edge computing and, and the regulatory retention reqs. Like, what used to double every few years is now multiplying continuously, stressing your existing infrastructure and operating models. Most enterprise architectures were designed for predictable growth and static workloads, right?
Exponential data growth exposes scaling limits, operation bottlenecks and cost across storage, networking, and management layers. Um- So Santhosh, let me, let me, let me ask you a question here about that. Yeah. Now, do you think that that's really gonna start causing people problems coming up here in the next couple of years when they start evaluating their infrastructure?
What do you think about that? Yeah. Those You know, your systems operate in isolation, right? Creating your management silos, lack of unified scalable models for your workload deployment. There are too many to manage, your file systems, buckets in case of object
storage, right, instead of managing workloads and outcomes. Right? Right. Inconsistencies and incorrect configurations. A lot of mistakes, right? Right. There is, there is. And it, it really, it plays onto the, you know, moving into a new era, if you will, of
automation and AI, you know, the, the, the new buzzword AI. Um- Mm-hmm and, and exploring, you know, new ways to do things a little bit differently. Right. And status quo is not an option, as you can see, right? 80% of IT budget today is just wasted maintaining outdated systems.
Have you seen this when you're talking to your customers, Mike? Oh. Oh, absolutely. Absolutely. I mean, I, I just got off a customer, conversation here a little bit earlier, one of the biggest things that came into you know, around the, costs that they're gonna start dealing with, that
third point you have there. Because they are bringing in so much AI, they're bringing more AI into their infrastructure- Right and they're trying to, to understand all of the, the increases that they're gonna have to absorb, from a cost perspective. And a big chunk of that is gonna be storage.
So I, I definitely hear a lot of this, you know, inside- in the conversations having directly with customers and partners. Mm-hmm. And the answer to this is our, our own Enterprise Data Cloud. It's a new paradigm for managing your enterprise data built around a unified data plane and intelligent control plane, right?
This architecture supports a simple self-service, scalable consumption model and is designed to support all workloads and workflows, right? In- incorporates all your CI/CD pipelines, Kubernetes, and all the latest and greatest around policy-driven automation. Anything you want to add here, Mike?
So what do you Yeah, if I could, what do you, what do you mean by workloads and workflows? What, what, what do you, what do you mean by that really? Yeah. So workloads is, is your, It could be a file block or object storage or even your application workloads. So where do you want to place in, in, in your entire storage fleet?
That's ex- That's essentially a workload in the context of an array. Right. And, and workloads is about Yeah, go ahead. Yeah. So when you talk about workloads, you could be talking about like a SQL DBA that wants, you know, a, a place to put the databases or Oracle, or even if you get down to the, you know, when we're talking about AI, being able
to put those, that as a workload somewhere, or VDI, you know, virtual desktops. That's, you know- Right still a, a workload that has to be distributed across, you know, your infrastructure, but specifically your storage environment. Mm-hmm. Right? I think that, you know, for Pure's doing what the EDC as we call it, the Enterprise Data Cloud, it, it really is a, i-
it's a, it's a concept. It's a, it's a, it's a building blocks, if you will, right? It's not, it's not something that Pure sells. It's, it's, it's actually a, a platform, a vision moving forward, is putting this cloud together and, you know, combining all of these components that we have.
But the center of it, you know, when you get to the center of things, you're talking about the, the Purity operating system, but you're also talking about Pure Fusion, and Pure Fusion plays a big part of it. Right. And that's how the journey to intelligent control plane looks like, right? It involves unifying your fleets and remote management and defining
your presets and workloads. Preset. What, what exactly is a preset? At, at its core, Pure Fusion alongside your Purity informs part of your data intelligent control plane, delivering a common operating environment and an Evergreen architecture, right?
The unified data plane supports your block file object storage with a high-performance, mid-performance, and low-cost option.All right.Yep. Yeah No, when you get to actually talking about Evergreen, the Evergreen architecture, being able to, to upgrade, your infrastructure without any fear, you being able to, you know, have things that go, go wrong.
Having self-service upgrades and non-disruptive upgrades. Being able to, scale your infrastructure seamlessly. And I'm, I'm, you know, not talking about, you know, any, any smoke and mirrors behind the scenes. This is legit. This is, you know, being able to scale, at, at
lightning speed, with just Flash. So yeah, that's I think Evergreen is, is really kind of core, to the EDC platform. And the hero of this movie is your Pure Fusion. So Pure Fusion is designed to unify your data and automate workflows. It offers your simplified data management, allowing you to manage a globally distributed
data with one consistent user interface anywhere it can exist, whether it's in the core, edge, or cloud. Right? And with workload automation, it uses policies for automated and workload out-provisioning, ensuring your better data protection, quality of service, and reducing mistakes.
Right? And furthermore, Pure Fusion provides your AI-driven workload placement, for balancing your workload deployment. So where does that- Yeah where does that, where does that workload placement come from, Santhosh? Where does that, come from?
So essentially, right, if you have a huge fleet of storage and you have the arrays, you don't know where exactly your workload is placed, depending on the performance. You want to move the workload. If, for example, an array goes, it has-- meets a threshold, we want to move a, STaaS part of storage into a different array.
So that part of this currently is done manually. What Pure Fusion does is based on a policy, which we call as preset, it decides where it needs to place, place the workload. Right. And it pulls that information from, from a, another p- p- piece of our platform called Pure1.
And, Pure1 has a something called capacity planner that it uses to, to, actually take a look at the, the fleet as a whole, to look at everything that Fusion is managing a, as kind of a single unit, and be to give you recommendations, of where that workload should land. Now I just want to point out one is, some folks are coming back and
we have dark sites. You know, what do we do about dark sites? We can't You know, we don't have them, you know, be able to talk to an outside service." But we've kind of solved that as well, right? We've added the dark site capability within Pure Fusion, so that connectivity
longer necessary. While it gives-- does give you, you know, more information, more analytics, it's, it's not required. Right. Yeah, the core of it, Pure Fusion simplifies your data management across three key areas.
One is fleet and policy management that drives sustainability and helps your total cost of ownership, right? And secondly, automation and orchestration, which are crucial for reducing your operational risk. And finally, the observability and AI Ops to increase your operational efficiency at scale. So right, M-Mike just mentioned there's a Pure1, website where a third-party solution
It's a, it's a Pure-developed, STaaS software that knows your entire fleet, and it can intelligently give, guidance to Pure Fusion to place your workload wherever it needs to be. Right. One of the things I'm interested in this slide, Santhosh, and I don't know if we can expand on a little bit, it's what's know, in the way of the compliance- Yeah you know, the compliance portion of this.
Yeah. Maintaining compliance is a big thing. Workload, mobility is another thing. Those are the two features that are gonna be, delivered over the next, couple of quarters. And, and yeah, it, it's coming. Y- you have anything to add there in terms of, how does it help customers?
Yeah. I mean, we're talking about looking at things like, you know, one of the, the problems with infrastructure is you have a bunch of infrastructure, that operates as a, as, as disparate units, and you want to get them to operate as a single unit. One of the things you have to be concerned about is, like, config drift.
And if you have things that kind of, ha- you have multiple people that manage these individual pieces of equipment, or infrastructures, and you, you know, maybe somebody changes something, that affects, the, the, the chain of, of, of things, chain of, of equipment, and so on, you know, you can, you can remediate that. And Pure Fusion will help you do that.
And Pure Fusion can actually do that today, able to remediate and, and, bring all of the, the arrays within a fleet within a, you know, a single set of configuration. So that definitely is something that, is useful from an infrastructure standpoint. Yeah. What you're saying is, as a pra- practitioner, Pure Fusion helps my life better, right?
So that I can do- That's right. That's right more of that rather than manual stuff. That's right. Yeah. That, that's awesome. Okay. And, y- you can join the community to learn more about how your peers are using Fusion in their environments, and there's a QR code.
We'll flash this towards the end as well. All right. So let's get into the demo, the exciting part where Mike will take us through how to do it. All yours, Mike. All right. And I'll just switch over
here for one second here. There we go. Okay. Hopefully we're all set. You can see that. Awesome. Okay. So I'm gonna start out here with a demo, and this demo is, actually gonna be talking, about, you know, how the,
Fusion experience works, from the start, all the way to the end of a successful, from, Pure se- presets and workloads, but also, how to do it not only via the GUI, right? Because we're not talking about just doing things via visual UI. We're also talking about being able to do things through automation, and the automation piece of this would be if you use Ansible, you use PowerShell, you use Python, any, you know
Or just, you know, have, have an interface, that communicates directly to an API. That is all possible with Fusion, because we are an API-first, company, and what we do is, pretty much everything that you can do within the UI, it's actually making an API call on the back end, to gi- you know, give you those results to, to do that function. So, in this particular demo, I do have, um-Apologize for not doing it.
Well, I do have three, FlashArrays, two FlashArrays and two FlashBlades. So we have the capability with Infusion now to, to combine, FlashArrays and FlashBlades. You'll notice I'm using a username of JJones, and a with a password, but we do require LDAP. And the reason why we require LDAP is because from a fleet perspective, we still have to be able to, you know, still have to be look at this from a holistic fleet point of
view, and be able to manage it from that standpoint. Now, when you think about it, you're logged into individual devices, and when you log into individual devices, you're only on that device. Well, with Pure Fusion, you're able to do it, across the entire fleet. So you log into one device, and you can manage the entire fleet from that single
So that's the reason for the LDAP. So if we go into an actual array, like I've done here, we look at the settings, and then we go to Users and Policies and we get down lo- a little bit further down here, we get to Array Management LDAP, and that's where that LDAP configuration exists, on each of our arrays.
And this is the same across FlashArray and FlashBlade. So we need the policy mappings there to the groups that are, are in LDAP. You click the test button, you make sure everything shows up green, and you're good to go. Right? The key to that thing that we, the, that we,
connect between all the arrays is that base DN. So if that base DN is the same across all of the arrays, then we're good to go. Now, some folks are asking, "Mike, why getting into this type of detail?" Well, is an expert-led demo, right? This is an expert-led webinar.
I'm getting into the nitty-gritty. So I hope, I hope that's what you came for. So if we go into the storage and we look and we can see single array things today, right? You want log into a single array, you see everything single array.
But if you actually go out and create a fleet, like I'm doing here, I give the fleet a name. Creating a fleet is basically three clicks, believe it or not. So you create a, give the fleet a name, you click Create, that fleet is creating. Now I'm gonna go into FlashArray 2, I'm make sure that the LDAP settings are the same across, you know, between, both arrays.
I'm looking at it, and that base DN is, is identical. The policy mappings are good to go. And I click the Test button, and they should all show up green. All right. From there, I'm gonna go into the fleet, and you'll notice the fleet has been created on FlashArray 1.
Notice that this one says it's a fleet coordinator. The first array that creates the fleet is always considered the fleet coordinator. That doesn't mean that it is all-knowing, okay? Every All the information is, dispersed across all of the arrays in a fleet. It just means that this is the array that created the fleet.
So I need to go out and create a fleet key. This fleet key is good for one hour and it's only good for one use, and that's a security precaution. So then we go out and we give, you FlashArray//X, we enter the fleet name, and then we enter the fleet key that we generated on FlashArray 1.
And we click on Join a Fleet. And while that's joining, I'll go over to the FlashBlade. I'll log in as the same LDAP user. It can be a different LDAP user as long as it has read, read-only, permissions to, the LDAP. And you'll see on FlashArray 1, we have FlashArray 1 and 2.
FlashArray 2 is still joining, but we'll hop over and we'll make sure that FlashArray 3 is gonna join here. We'll create a new fleet key. Remember, it's gonna tell me, "You just created one. Are you sure you wanna create another one?" because it will nullify the, the previous fleet key that you created.
Again, all a security precaution. So we copy that. We click on Fleet. We join an existing fleet, just like we did before, put in the fleet name and then the fleet key. Once the fleet key is entered, we click on
Join a Fleet. Now we'll go back and we'll take a look and see what the fleet looks like. You will see there's FlashArray 1 and FlashArray 2 are both joined. Now we have FlashArray 1, FlashArray 2, and FlashBlade 1. So they are all connected to the fleet and joined.
All right, so we are good to go as is. We're ready to rock here. We click on Storage. If we go over to Hosts or Volumes, we now have the ability to look at a fleet view. So we have a local array and a fleet view.
If you click on Fleet View, you're gonna start seeing everything that's within, within the fleet. Same thing for volumes. We're gonna have a fleet view there as well. You will see this fleet view option available in several different areas.
While you're in here, you know, if you click on an object, it'll tell you everything about that object. You know, even if you want to search on these fields, the names, the size, things like that, you can search across the entire fleet. A lot of people have been asking that, to be able to search across multiple arrays. So now I'm gonna get into creating a preset.
If you create a free preset, you can upload presets. Presets are portable. They're just JSON files. So you have the ability to actually offload them, edit them, and then re-upload them. But we're gonna go through and create a preset.
Now, I am showing you something today that is brand new, that has not been released yet. So just so you know, you're getting the latest and greatest, but it may change by the time we release it, which is gonna be very soon. So we go ahead and we click on a preset type, a block, object or file. You can skip the compatibility check, which allows you to create something that doesn't
ex- already exist in your fleet from a, a, a array type format. We give the preset a name.And we'll give a preset of a name here. That name, go ahead. Yep, SQL. And we'll call it SQL Dev one, and the workload type. Now, here you can specify a type of
this particular one, we are gonna select, SQL. Now, you can from a storage class, we're to look at predefined or customizable. You'll see this quite a bit. Predefined means that when you deploy the workload, you can't change the things that are in the preset. If it's customizable, you can actually change
these parameters within the preset. All right, all these parameters we're gonna be configuring here. So I'm gonna select the FlashArray C as the, the storage class. You can give it a description, but you can also give it a workload tag. I'm not gonna do this 'cause this is a simple preset.
But a workload tag could actually be anything like a billing ID for chargeback or look back. It could be a department ID. Anything you'd like to put in there as a tag. We'll go to storage resources now. These are the actual resources you wanna deploy.
We're gonna call it a data vol. You do have the ability to change the naming pattern. I know some folks out there have really complex or simple naming patterns for their, for what they have. You can go ahead and put something in here, like I'll put data vol, and that will, you know, be a prefix to what is automatically put
as the volume name. So if you wanted to put like DB there or something like that, we have a bunch of options. If you click that plus key, we have, you know, you can take functions, you can take, attributes, all of those, and be able to add those to the naming pattern.
Then we'll go down to quantity. Again, you'll be able to specify if it's predefined or customizable, but we'll create three volumes. And we'll put a provision size in there. We'll make these ten GB in size. Now, we could do a snapshot configuration and a replication configuration, but since this is
a simple one, I'll get into that a little bit later, we'll just go ahead and, and move forward. We could also specify QoS, which is IOPS bandwidth, but we're just gonna publish this as is. So this is our preset. You'll notice it's prefixed with ple- Evergreen//One.
It'll tell you everything about the preset. If we go to FlashArray two, we click on Storage, we click on the Workloads, you will also see that pre-- That you'll see I'm sorry, presets. You'll see that preset there as well. All right? Because the entire fleet
knows about all the presets. All right, we'll go ahead and deploy this preset now. We'll click on Workloads. There are no workloads currently. All right? We haven't deployed anything.
But if we click the plus key over here, all right, we're gonna be able to select this type of, of, of, of preset. Now, you can download the preset, all right, change the parameters of the preset, re-upload it, and then deploy it. So we're gonna click on the workload.
From here, we're gonna select that preset like we had defined, right? We're gonna click Next. Now we give the workload a name. Go ahead and, we'll just call it the SQL Dev. Yeah. Workload one, W one. And then we'll click Next.
And then now here comes the placement piece of this. Now, I do wanna show this as something because this is kind of important. When we look at the workload placement, we have a workload target. Now, the workload target from the workload placement could be any array within the fleet. You can specify a single array, you can specify multiple arrays or, you know, you can
specify the entire fleet, as you noticed when we did the dropdown. Now, when we did the dropdown, I select the top box, so it puts all of the arrays that are in the fleet in that, in that box. Now, you can pick a projection period. When you project-- put a projection period, you're s- you're saying, "Take a look at the
last X amount of time," okay? And also take a look at what could be happening with that fleet within a month, two months, six months, a year. And think about, you know, take the, take the algorithms and figure out where the best placement would be.
Now, if you-- This goes back to the Pure1 Workload Planner that we talked about in, in the, when we were using the deck, right? With the deck, we were talking about work- Pure1 Workload Planner that has AI built into it. It goes out, it knows everything about the fleet.
It sends back that information, back to Fusion, into this interface, and it tells you, "This is where we think you should place this workload," based off of the information you've given and the history, and forecast of the fleet itself. Now, if you are a dark site, it will attempt to make that call to Pure1. If it cannot make that call to Pure1, it will make its own call using just the fleet data
that it has, that it collects, at a fleet level. All right? So it won't actually use any outside information or any data gathering, anything like that or any, you know, extra AI functionality. It just uses the raw metric data that's within the fleet.
All right? So this particular demo is actually a dark site, right? It doesn't have capability of talking to Pure1. And I did that specifically 'cause some are like, "Hey, you know, if I do that out to Pure1 and it comes back, you know, it takes a while to come back, how long actually take to figure out how, you know, how, how I can do it without having, you
that AI capability of Pure1?" Well, it's pretty quick because it's constantly collecting that data, and it, it reads that data, you know, very quickly. So I'll, move forward with this. You'll see, you'll be able to select that projection period. I'm just gonna select one month and then get recommendations.
Now, I'm not speeding this up at all. This is actual live. This is actually doing this, you know. That's how quick it is, to go out and get that fleet data. It's gonna tell me that the acceptable, FlashArray two.
Now, normally, if you had multiple arrays that matched the workload type, the storage class, I'm sorry-The storage class of, FlashArray//X, so on and so forth, you would be those, okay, and they would show up acceptable, and they'd be ranked in order. All right? Now, you don't have to
take that recommendation. This is entirely up to you. Can you do it? Yes. Can you do it automatically? That capability will be coming where you can have it automatically do it.
But little bit of caution there, right? We've had people come back and say, "Yeah, that's not necessarily what we want to do." So we're being really flexible. We're giving you the opportunity to either do it or not. It's kinda like DRS, right, we have with, vSphere.
You have the ability to select whether you want it to be automated or not. And we're gonna offer that flexibility So it gives a recommendation, and then I click Next. All right? Now it's telling me this is actually what's gonna be deployed, where it's gonna be deployed, and I'll just
click on Deploy Workload. Now, once that deploy-- workload is deployed, you'll actually see it inside of the fleet view. You can see the workload there, all right? And we're on FlashArray//One.
It is Notice the context is FlashArray two. It deployed it there, so we'll go to FlashArray two. We click on Workloads. You can see there it is without the context because it, that is the array And you can see everything it created.
It created three volumes of ten GB each, you know, and if we would've added any of the things like workload tags, you know, the pro, protection groups, all that other kind of stuff, it would had that as well. Okay? So moving on, I'm gonna show this actually from a CLI.
Now, if you don't know that, Purity, which is the Pure operating system that is, runs our platform, it allows you to run CLI commands, right? If you have the ability to do that in your environment, be able to run, direct SSH, you know, connection to your arrays, I know some of that's possible, sometimes it's not possible.
But we do have the same type of commands that you can run directly at the command prompt. Now, why would you wanna do this? It's because, number one, some folks, from their automation standpoint or even from their, you know, if you, you know, have your NOC or operation center managing this, they may wanna do everything via the command line.
Totally respectable, and that's why we provide this, because, you know, we need line capability. So you can do everything that we just showed, via the GUI in the command line, you'll be able to do it via the API, which I'll show right after this. So from the command line, we're just gonna look at the Pure preset, and I'll, look at the
workloads here. I'll do a hyphen H just to show you what's, what's possible, right? This is where you can create, delete, list, rename, update, upload. You can do everything that you saw, excuse me, in the, the GUI. Okay? And it gives you a lot of flexibility.
And everything, everything has help, so no problems there. You'd be able to, you know, we also have the CLI guide, which is on every FlashArray and also via our support site. So from this standpoint, we're gonna look at Pure preset, we're gonna look at workload, and I'm gonna actually list the workloads that are available.
See, that workload is there. It's available, right? Now, if I wanted to deploy this, I would do, obviously, a Pure preset would download it if I wanted to do that. I'll show, my actually downloading it to a file.
All right. It wants a context, which is the fleet name. Remember that. I'm doing this, I'm stepping through this so you understand what's, what's, actually required, right? So the context is a fleet name. That's what I'll give it. And then you have to actually give it, the,
preset name, right? Obviously. And it's just the preset name here. Yep. See? I even, I even messed up when I did this. I was like, "Yeah, well, I thought I had to put preset name." No, you don't. It just wants the name. So for this standpoint, I'll go ahead and put
the SQL-dev one. I messed up on purpose, so I could show that, but you'll see here, the JSON file comes out, and you can download, so you copy/paste it. You can download it to a file, and you can edit any piece of this and be able to re-upload it to any fleet, actually, or, you know, the same fleet.
It's, it's just an easier way to, you know, share, presets and be able to also, edit them, quickly, without having to use the UI. So from this standpoint, I edited that, that preset. Now I can go ahead and I can, same thing, download it. I'll open it up here real quick.
I'll just open it with Notepad. Yeah, that's fine. And I'll go in here and, you know, change something here. We'll go ahead and f- and change the FlashArray name down and workload type. Yep, there's thin provisioning size.
I'll change, from FlashArray two, I'll change that to Pure1. I'll change the, the, uh-- I'll go down to storage class and change that. Well, I'll change the name first. Yeah, let's change that to prod, and I'll put that on a storage class of X.All right.
Okay. So, you know, we'll tune in, we can change the volume. Let's change that to Yeah, we'll make five of them, and we'll change the vol one to And we'll put it there and leave that to vol one. No, we'll leave the vol one the same. That's fine. Okay.
All right. Let's move up. And we'll remember to put prod in there just to make it clear for everybody. Leave the size, workload type. Go ahead and save this file. And I'll just change the name, so I have an extra copy of it.
I'll change that to prod, so I know it's the one, new one for prod. And click Save. And we don't have to have the fleet name there, right? In case we want to bring it into other fleets and things like that. I'm basically showing this vi- from the simplicity of being able to, you know, edit
these, offline. So here I go ahead and do an upload preset, and browse for the file. Go to the downloads, click on that one I just edited, click on open, put a give it a name. And we'll call this the same thing. And then I'll click done.
And now you'll notice that my preset is there, the prod one. So now I have the ability to deploy a workload off of that. It's that simple, right? And you can see that it actually made all my changes there, put in the prod data vol and, changed my resources, the quantity five and all that.
Cool stuff. All right. There's the workload. There's the presets. Now, I take a look here in the CLI, and I look at the workload, and I do a deploy of the workload. It's gonna ask me, you know, I wanna know, I gotta put the preset name in, I gotta give any
parameters if I want to change any parameters. Because remember, it was either prescriptive or, a- or customizable, right? If it's customizable, you can add that dash parameter and add any parameters want to it. Like, you wanna change the number of volumes, you wanna change, if it's, you know, the QOS numbers, anything like that.
So we'll go ahead and put in the, the preset name, and we'll, we'll create the workload. And call it workload two, W2. So it goes through, and when it creates a workload, it'll actually report back, that, that there's the created, it was created, and it's ready.
If we go back and we look at workloads, and again, we look at fleet view, now you can see that that workload was, is showing up there. We clicked on it. There are your five volumes, at 10 GBs each and all the different, text changes I made. So in this, part of the demo, I've switched over to some different arrays.
And that's from a, just from a strictly a versioning, standpoint. What we wanna do here is we wanna take a look at this from a standpoint of, of automation, and we wanna think about it from a automation standpoint where you have these multiple workloads. Now, we have a very simple one here, one called, VMFS, and then we have a SQL one.
Now, if you look at the VMFS one, it's very simple. It just has storage resource, just like I created previously. If you take a look at the SQL one, the SQL one is much more complex, and it has a of volumes to it, you know, and we add snapshot replication or snapshots and replications to it as well.
Now, the important ones we're looking at are quorum, MSDCC volume, the temp log, and then we look at the snapshot configurations as well. We change the sizes of the disks. And, when we look at the snapshot configurations, one snapshot every hour for one day and, one snapshot every one day for one week, and this periodic replication, we're
actually replicating it to another array. Now, this is all contained within a preset, okay? And what I'm gonna do here is I'm actually gonna bring this into Ansible. So I don't know if a lot of you work with Ansible, but it's basically just creating a playbook inside of Ansible to deploy, this, this workload, to create this preset.
I'm sorry, to deploy the workload. All right? So that's exactly what we're gonna do, is we're gonna go through and make sure that we're eliminating human error, and we have everything exactly the same, you know, templating, if you will, across, you know, any, every workload deployment that we do.
And by doing this via, you know, automation mechanisms like Ansible, if it was, you know, a Terraform or a Vetaria or anything like that, you have the accomplish those goals and, and having this as a repeatable, you know, a SafeMode type of deployment. So from this standpoint, we'll go ahead, I'm gonna take a look at the, actual YAML here.
So this is an actual playbook in Ansible, and it's just provision You can say my prefe- preset name is VMFS. The context I wanna, you know, deploy it to, that could be, you know, multiple names there. It could be, get a recommendation, if I had a recommendation on there. The name of it, and then alsoThe host.
Now, notice I don't have, like I mentioned there, the recommendation. You can have a parameter in there that actually goes out and does the recommendation step first and then prompts you. All right? So I'll just go ahead and run this. From the Ansible playbook, go ahead it, and you'll see it comes back and
created the volumes, and it was successful. So we go back in to the arrays, flip over, and then we go ahead and take a look at, the workloads. And click Refresh there. There's the workloads.
There's the volumes. It is a, the VMFFS. It is the VMware cluster, and, they actually connected the host. So it does an automatic host connection, is different than what, you know, we used to do in the past. Cause Fusion allows you now to automatically
connect hosts, when you create the volumes. It used to be a couple step process to take the volumes and then connect them to the host. We can do that all in one automation step now. All right, so we'll take a look at this again. We're gonna go ahead and change this, and we're gonna change it over, to the SQL one,
the more complex one. We're gonna change a couple parameters in this, in this, Ansible work, in this playbook. We'll go ahead and change it to SQL Server cluster. We'll change this to a Windows host, so it now connects all those to a Windows host, all right, instead of the ESXi host.
So we'll go ahead and, and run this. Now, you'll notice in the tasks how it's going through and it's creating all of those things there. Okay, it created it. We'll go take a look at it, and then we're gonna go back to that screen for a second, but
I wanna take a look at this first, is you can see the SQL Server cluster is there. We'll take a look at that. See, all the volumes are created, all the sizes are there, and we even created the protection group, right, and the snapshot schedule that we set out and the replication, right? The replication schedule is It's, it's there.
It's not enabled because it's not enabled by default, but, now it's enabled, right? All of the-- Right, we can take a look at each individual volume. You can see they're connected to that host, okay? Everything that they-- we wanted to happen. Now, notice the time.
That's one-- the key thing I wanted to come back here for, is it only took eight point 0.06 seconds to do that deployment from a preset. Now, this is, you know, this is something that people will be like, "Yeah, it's, you know, it's probably sped up. It's, you know, it's not, it's not real." It is real.
Trust me. It is something that, you know, is, is very quickly, can be done. It's incredibly fast. Okay, so we're gonna go back, now, and what we-- one of the issues that when you have these fleet of arrays, it brings up something that we talked about in the
beginning of this, the webinar, was around having configurations across all of the arrays be alike. Now, you have, you know, things like the, NTP and the, the Syslog servers and the, you know, different configurations for DNS. You wanna make sure that they're all the same across all the arrays. Now, if you have somebody who, you know, manages it and say they
change something, right? And maybe somebody changes a DNS server, like we're gonna do here. We're gonna change the nameserver to instead of eight eight eight eight, we're gonna change it to eight eight eight nine. So now we are-- have what we call a configuration drift, or we're out of compliance, if you wanna say that as well.
So how do we fix that? How do we fix that with automation? Well, we can do that with Fusion, right? We can actually go through Pure Fusion, can actually bring the arrays back into compliance using, you know, our Purity APIs.
So we're gonna go ahead and we're gonna create a playbook that'll actually go through and configure all of the settings for the arrays that we want specified, and then to each array, okay, and it's gonna set them the same. Now, you'll notice in here we have a couple different files. We'll take a look at the array config.
The Array configuration is what we want all the Arrays to be. This is where we say the DNS, the name server, should actually be eight eight eight, not eight eight eight nine. All right? And then we'll put in the NTP servers. We'll put in the Syslog servers.
We'll put in protection group updates. We'll put in hosts if we have to change anything on a host from an IQN standpoint, an iSCSI, or a personality, anything like that, anything we have to change, right? We put that inside of this, this file, and then we de-- you know, put it through the playbook and it deploys it to the entire fleet.
So we've changed a couple things here. We're changing the snap at three PM, and we're making sure that, you know, we have the eight eight eight eight as the DNS. Now, we are also going to the Windows host here and from that standpoint, if we ever wanted to change the IQN, we could do that.
So we're gonna go ahead and take a look at the configure FlashArray. This is the actual playbook. This is where it's gonna go through, and it's gonna retrieve the fleet information, get the membership, and then it's going to hit every one of those areas that we want configured. It's gonna configure the DNS, the NTP, the Syslog, and it's gonna do
it across the entire fleet. Think of it as doing it in a loop, right? It's an automation. So it's gonna, it's gonna hit all those FlashArrays. And whatever we want in that playbook, it'll do. So here, we're gonna run that playbook.
All right. And from this standpoint, you'll notice it's going through and it's configuring DNS on each host. It'll actually sa-- show you that it changed one of the hosts. That one host where it changed it to nine, it changed it back to eight dot eight. All right? So that's exactly what we intended it to do.
We want it to go through and do that, okay? And that is a prime example of how, from a fleet management perspective, you know, Fusion can help solve those problems.Okay, all it requires, see, it changed it back to the 8 dot eight dot three dot eight, and, you know, everybody is happy at this point, right? Even protection groups, you know.
We, we put the protection group, we make sure the snapshot schedule is the same across all of the arrays instead of having one that might have been a little bit off. Same thing with the hosts, we make sure the host IQNs and names and everything are all the same. Right? So that is the beauty of Fusion and being able
to, not only do a presets and workloads type of thing, but also do a compliance, and being able to get everything, in line and avoid, the configuration drift. Now, here I'm gonna join another array to the fleet. And the reason why I'm doing this is because I'm gonna show how running will actually configure this new array.
Remember, this is an array that is unconfigured. It has no settings, it just went through those screens. It is just joining the fleet, right? Or it has different settings than what the fleet has. So we're joining it to the fleet, and now we're gonna go back and we're gonna run that
script again, okay? And it's gonna go through and it's gonna say, "Okay, get fleet membership. Oh, I've got a new array. I've got to go through and configure all this." See all the yellow showing up there? I know it might be difficult to read.
Apologize for that. But it's actually going through and configuring that new array, okay? All via automation. Once it, gets through the, the script here, we'll take a look at the new array. Yep, there we go. And now we take a look at this is the new
array, and you'll notice NTP server is all set. Again, syslog, the DNS, which was blank before, and even protection groups, snapshot schedules, everything, matches, the fleet. And you can see now every, every, host has, every array has the same host. So kinda cool, right? You can see all that.
Now that's the end of the demo, all right? So we really what we would like to do now, is come back, and if you have any questions for us from a Q&A standpoint, this is the time. Please feel free to ask the questions. And, Santos, I don't know if you've been monitoring at all, if, uh- Yeah, I've been
monitoring you have any questions that are out there. Yeah. We have a very engaged crowd here. The first question, that Chris has is, "Are there any limitations to using SafeMode with Fusion?" there actually, there will be, right? Because SafeMode inherently, will lock down the array, right?
You have to be an array admin, number one, but number two, there are provisions to actually, you know, be able to work around, the Safe Mode limitations because you are an array admin and you give that, that capability. It's all based off of permissions, number one. Number two, it's based off of if you have Safe Mode, enabled for the entire array or you have
SafeMode, just enabled for, different, aspects of the array. So there's a couple different things involved there. And, to answer the question, yes, there are provisions that you can work. Fusion will work with SafeMode. Okay. Good.
And there are a couple of questions around security. So what they're asking is- Yes what about identity protector, identity provider, s- something like SSO? There's a related question, shouldn't, Fusion tie to Active Directory or LDAP? Yeah. So as it stands today, Fusion does
identity provider, right? And pr- and specifically it requires an LDAP provider. Okay, so we do require LDAP. It doesn't have to be Active Directory. It could be a Linux-based LDAP, it could be OpenLDAP, it could be anything.
But it does require an external l- authentication or identity provider, service. Now, we do have SAML support coming, so you'll be able to do SSO, you'll here in the near future, like in the next, you know, couple months or so. And you'll also, at some point, you know, be able to, go beyond that, and we're, we're taking a look at different avenues of being able to have, you know, local users actually
being able to manage your fleet. But from this standpoint, we, we really have to be cautious about that. Because from a security standpoint, when you look at it, you have a local user that may be identified on one array. They only have permissions to do things on a single array.
Do you really want them managing the entire fleet, right? So you, you really, you have to take all of that into consideration, and it's something we take very seriously. We do have a, you know, a, a new concept that came to, Purity OS here in the last year called Realms and multi-tenancy. So if you haven't looked into that, that's something that also plays into the whole
security aspect of, of how, the arrays are administered, and addressed, from a, just from a, a, a plain security standpoint, and from a, you know, a targeting standpoint, right? It, it how, how authentication happens, granular RBAC, things like that. And there's a related question to that, right.
In a DR scenario, could the lack of LDAP be an issue? Yeah. It will be for Fusion, absolutely. I mean, Fusion requires it. There's no way around that right now. It is just something that we have to have an identity provider, that actually, you know,
supplies that ability to, to look at every array. I mean, if we didn't have that, you know, it's not possible to, to If, if, if you change the permissions on one array, it won't reflect on the other arrays, right? So, it, it's just something that is necessary today.Yeah. And there are a couple of other questions on fleet creation.
There's one here from Phil. Is it best to create a fleet for arrays at a, at each site, considering potential configuration differences? So that's really kind of, you know, up to what the organization is all about, right? And how it's, how it's actually structured 'cause we have customers that are saying that
they have Say they have an example, of 10 arrays. They're saying, "Should I have all of the arrays within, you know, all 10 arrays within a single fleet?" Well, if you wanna manage them as a single, as a, just like a single Array, absolutely. But if you actually have different, you reasons for separating them, whether be
geographical, or from a compliance a governance, if you You know, some know have arrays over in, in the EU, which have different laws governing data access and data retention and, and things like that. You know, it all depends on your situation. Having all the arrays within a fleet is great.
You can move arrays in and out of a fleet seamlessly, right? You can remove them, you can add them. If they have workloads on them that are related to the fleet, you can move those workloads easily. But there's, I don't think there's really one answer to that.
I don't think there's an all-encompassing, you know, you should do it this way type of thing. Yeah. We do have a feature coming up here, that will be announced soon, and that actually brings something like grouping to, to the picture of Pure Fusion. We currently don't have that today, but we will have that, that feature, where we'll be
able to group the arrays within a fleet, if you have purposes for that. Again, getting down to an organizational level really. Yeah. And, and there's another related question to in moving data across storages. So I think you're, he's talk- probably talking about arrays.
Should a single host be connected to all arrays so, so that I can be able to move data? Yeah. So that's one of the things is do I Did you, did you like open up zoning? So zone everything everywhere? Yeah. Yeah, so it depends on, on, on some of There's
a bunch of factors involved. You know, we can watch, you know, Purity has a native ability to watch for, different types of, of, zoning changes, right? Different connection, changes. I won't go into too deep into details of it, but we have the ability to watch for that and
then, you know, alert for pathing errors, which we do today. Or, you know, have your automation automatically, it can trigger actual automation changing pathing for you. So there's a bunch of different things we can do here. Do we have people that like zone everything to everything? Absolutely.
Is that the best way to go about things? Well, depends on your situation again, right? Because, you know, depends on how flexible you are, and I'll just leave it at flexible because flexible can be interpreted a different ways from a security standpoint, I'll just leave it at that. And you know, if there was a, a magic thing
that we could throw at, at, you know, at, at the infrastructure in general and say, "You know, whenever you do this, we need move it over here and do it magically." we're TR- we're trying. We're trying to get to that point, and we've gotten to that point in some aspects but, but not all, so. Yeah. There's another question from Sebastian.
So this is related to fleets, right? We-- Can we use automation without creating a fleet? Yes, absolutely. Purity is based off of an, a set of APIs, right? So we have, the REST APIs, which are all available to the public via Swagger, out on
our code.purestorage.com website. And you can do that, all that automation, okay, from a, multiple arrays. You can do it to multiple arrays. You just don't have that fleet capability. You can specify multiple arrays when you talk to API endpoints.
You can do that, but the problem is, you can't do it from a holistic, you know, single point, type of thing. You have to address each array. So it, it's, it's, it's more cumbersome to do it via automation, but you absolutely That's what we've been doing for years before Pure Fusion came along.
And we have examples of that out on our Open Connect website as well, out on GitHub, on being able to automate some of those things via, you know, SDKs or Ansible, or Python, you name it. And again, you can go to code.purestorage.com and find out more about that. Yeah. You brought up the Ansible.
I think he has a follow-up question. Can Ansible connect using rotating keys only? Use-- Connect via what? I'm sorry. I did- can Ansible connect using rotating keys only? I'm not sure what that- that's a good question.
I don't know the answer to that, to be honest with you. I'd like to find that out though. If you give us your information, I certainly can follow up with you on that. But I don't know if that's actually It, it should be possible, because I, I don't see any, I don't see any operating system or API limitation around it.
Our API, you know, we give a token out. We give a, what's called an XAuth token, in return, and that XAuth token has to be used on the FlashArray. So I don't see, I don't see any limitation. I think- I've just never done it, so. Yeah, I think he clarified the question a little bit.
So he's asking, can we use without need for username and password? Meaning just have the keys and- Oh, yes be able to connect. Absolutely. Right? Yeah. Absolutely. Absolutely. We actually have two methods. Okay. We have API token with XAuth token, and we also have something called, we also implement
OAuth, OAuth 2, or you can use JWT tokens as well. Right. Awesome. And this is not a question, I think it's a suggestion. Can future trainings like this use industry standards like RFC 2606 and RFC 6761? I'm not sure what they are. I'm not sure either.
I mean, he's If you do a cut and paste, he's expe- expecting, generic names instead of specific names. I don't know. Yeah, I don't know. I don't know. All right.Um, one of the, the questions that I wanna talk about, is that, it, it comes up frequently is how do I get Fusion?
Yeah. Right? I think it's important to remember that Fusion doesn't cost you, a customer anything. Okay? If you have Pure and you are up to a revision level on your Purity operating system, you have Fusion. Fusion is actually There's no on off switch for Fusion.
Fusion is actually on by default. You'll see a, an option on the left-hand side of your UI called Fleet that'll show up, and that means that your Pure Fusion is on and ready to go. The only thing is, is that you actually, you don't have a fleet until you actually create one. So you can create a fleet, and a fleet can
consist of a single array. You don't have to have multiple arrays. You can have a single array within a fleet, to get that, the, you know, some of that functionality of the fleet. Is it viable to do that?
Absolutely. You know, and though, again, there's no SKU involved, there's no cost involved, it's just there. Yeah, I think the one question is what ver- version is that? It's probably 6.8, right? The minimum Pure Fusion version or the minimum Purity version.
The minimum Purity version on FlashArray is 6.8.1. Yeah. Okay? And that also includes the 6.9 LLRs, and 6 dot s- 6.7., dot I don't remember the, the last one of LLR on 6.7 release. And then on FlashBlade, I believe it is, 4.5.1? Yeah. 4.4.5? Yes.
Okay. It's hard to remember. It's been, it's been a little while. Yeah. Yeah. All right. I think we have run up, topped the hour. All right. Well, thank you much for joining.
I really appreciate it. Again, my name is, Mike Nelson, and, Santhosh, and thank you for joining me and, doing this, webinar. Thank you for everybody, for everybody for joining and, you know, hope you have a great day. Yeah. Happy holidays. Yeah.
Mike Nelson
Sr. Technical Evangelist, Everpure
Santhosh Siruvole
Sr. Product Marketing Manager, Everpure
Manual provisioning and reactive management can slow innovation and drain valuable IT time. What if you could manage your enterprise data intelligently?
Join us for a Pure Fusion™ Expert-led Demos Webinar:
- Walk through Pure Fusion configuration and fleet creation to securely federate arrays and gain one, consistent data management experience across your environment.
- See remote provisioning in action—manage any array from any array and provision storage anywhere via GUI, CLI, or API.
- Learn how policy‑driven presets standardize protection, QoS, and naming for repeatable, error‑free deployments—and get AI‑driven placement recommendations.
We Also Recommend...
Personalize for Me
My Updates
No updates at this time.