Hyper-convergence has been the focus of many networking infrastructure discussions for a while now. And with the new partnership between Nutanix and F5, it’s a great time to talk about how you can get started reaping the benefits of integrated systems. Over the next few weeks, I’ll be discussing the technical details about how to use the F5 BIG-IP system and related technologies (e.g., iRules, iApps, etc.) in a hyper-converged infrastructure. At the same time, my colleague Frank Strobel will be talking about the business benefits of F5’s partnerships in this space; his first blog post focused on Nutanix.
First off, let’s talk about hyper-converged infrastructure and how it’s different from traditional and converged systems. Traditional IT infrastructure is silo-based, usually with dedicated hardware appliances for networking, software, storage, etc. Converged systems streamline operations by consolidating server and storage hardware to provide a shared compute infrastructure. Hyper-converged systems take that consolidation a step further, using a software-centric architecture that integrates compute, storage, and virtualization on commodity hardware supported by a single vendor.
And that’s what Nutanix provides: a turnkey hyper-converged appliance that scales quickly and easily, allowing organizations to support any virtualized environment—without the inefficiencies of overprovisioning or complexity of separate networked storage. The Nutanix management and operations solution is called Prism, and it provides centralized management of infrastructure and virtual machines, along with rich operational insights, and one-click remediation of common issues. But when you deploy enterprise applications using a hyper-converged infrastructure, you still need to provide the application services your users demand. That’s where the BIG-IP system comes in.
So let’s jump into it. In this post, I’ll walk you through installing BIG-IP Virtual Edition on Nutanix using the VMware ESXi hypervisor. I know that many of you already own a F5 BIG-IP appliance to support your application infrastructure. While this post will not focus on how to leverage your existing investment, we will provide some information about this specific use case in a later post.
Ready? OK. Here we go.
I start by logging into the Prisim Portal
I started with a blank slate. This Nutanix box is only running 4 VMs at this point: the vCenter appliance and the three controller VMs. I know that this is not a qualitative measurement by any standard, but I was impressed by how fast the system was — from provisioning to the actual startup of the VMs.
First, you need to download the F5 BIG-IP .ova file. Just point a browser to https://downloads.f5.com and either log in or create an account, then follow the instructions and log in.
After you’ve logged in, click on the Find a Download link.
Click on the BIG-IP-v12.x/Virtual Edition link.
Click on the Virtual Edition link.
Read (You did read the EULA? Right?) and accept the EULA.
Next, you will need to select the “hypervisor flavor” of the download you want. In this case, I downloaded the BIG-IP-18.104.22.168.606.ALL-scsi.ova and the BIG-IP-22.214.171.124.606.ALL-scsi.ova.md5. You will notice that the Description field for these downloads indicates that these are for ESXi, as opposed to XenServer, Hyper-V, or KVM.
Don’t forget to check the .md5 value of the .ova file to verify the file’s integrity. Once you’ve done that, make sure the file is available either as a local file or on a web server so that you can reference it during the deployment process.
Next, you need to switch to the vSphere Web client and log in.
I’ve pre-created the 4 dVS networks to which we want to attach this BIG-IP HA pair (Management, Internal, External, HA). You can name them what you want or connect them to already existing networks; I just chose these names to make it easy to visualize the four separate networks. You will notice these network names later in the post when I describe which networks to assign to each interface.
Once you’ve logged in, go to the VMs and Templates view. I created a folder called “F5” into which I will place these new BIG-IP VEs simply to keep the Web Client organized.
Right click on the F5 folder and choose Deploy OVF Template… then select the source .ova file you downloaded earlier. In this example, I chose to browse for a locally stored .ova file.
Here is another EULA to read and Accept. Then click Next.
Provide a name for your VM. I chose “BIG-IP12-01” (even though this screenshot was taken before I changed the name.) And the F5 folder was already selected.
There are several configuration options available: 1, 2, 4, and 8 CPU options, each with a corresponding 2 GB RAM per CPU). This allows you to size the VE for the modules you will provision on it. In my case, I chose the 2 CPU/4 GB RAM option because I’ll be using only two modules. You can find info on how to size the memory and disk requirements for various modules in the BIG-IP Virtual Edition manual.
Make sure that you choose a particular host on which to install BIG-IP #1 and a different host in the cluster to install BIG-IP #2. And make sure you enable anti-affinity rules so that the two BIG-IPs are not running on the same hypervisor host.
For more details on this, see the relevant section in the vSphere 6.0 Documentation Center.
I selected the shared default storage container that was created via the Nutanix console to deploy the VEs, rather than the local storage associated with this specific hypervisor host. Selecting the shared datastore allows the F5 Virtual Edition to be vMotioned at a later date for maintenance purposes.
Here is where those four networks I created come into play. I matched the F5 interfaces to the Port Groups that were created for those purposes— you can see them in the picture below.
Review your selections. Once you are happy with them, just click Finish.
Notice that we have two BIG-IPs installed. I simply followed the steps above and named the second VE differently from the first. I also chose to place it on a different hypervisor host.
You can see these two BIG-IPs in the vCenter web client.
And you can see on the Nutanix Web Console that I’ve deployed them to different hypervisor hosts. Here’s how those New BIG-IPs appear in the Nutanix Web Console > VM > Table page:
And that’s how you install the BIG-IP system on Nutanix using the VMware ESXi hypervisor.
Stay tuned. In the coming weeks, I will be taking this a step further and describing how to use these newly deployed BIG-IP Virtual Editions to support an enterprise application. To learn more about what F5 is up to in the integrated systems space take a look at our dedicated web page.
In the meantime, check out this post on the Nutanix blog about how Nutanix and F5 simplify, scale out, secure, and accelerate VMware Horizon View.