In our web series on leveraging virtual infrastructures for small business, today we will be looking at VMWare ESXi.
I will be splitting this series up into posts about workflows, basics and advanced. In the basics, I will cover installation and management of hypervisor resources. While in advanced, I will cover deploying more graphic, storage or computational intensive virtualization tasks.
Many small businesses are aware of the freedom and democratization that Virtualization has a brought to the business space. AWS (Amazon Web Services) has revolutionized the way startups and small business run IT. For example, Airbnb has an IT operations team of just five, but manages 150,000 overnight stays a night. They can do this through the power of Virtualization services provided by Amazon’s cloud services.
they’re growing in no time at all, they’re servicing customers like seasoned pros, and they’re thinking about resources without worry.
But you don’t need to jump onto the AWS cloud to gain these advantages. Some business owners today, embracing the switch from PC to Mac, have not left their PC software behind. They use virtual machines to run their legacy applications along side their Mac counterparts, using application like VMWare Fusion, Oracle VirtualBOX , or QEMU. What many do not realize is that there are free Server grade virtualization solutions that can do the same magic for their entire business, using much older hardware, and with the same easy of use.
Today we are going to take a look at VMware’s vSphere Hypervisor (ESXi) bare metal hypervisor solution.
ESXi installs directly onto your physical server enabling it to be partitioned into multiple logical servers referred to as virtual machine
Modern servers and workstation computers are often over resourced for their intended purpose. They spend most of their lives under utilizing their CPUs, Storage, RAM resources before being replaced a few years later for a more powerful machine. What vSphere Hypervisors allow you to do, is logically partition your computers hardware so multiple operating systems can more fully utilize available resources.
So don’t throw away your older desktop or laptop. In fact, you can combine different hardware together as the hypervisor only sees hardware as assignable resources to be allocated for consumption. A hypervisors can easily partition the same hardware resources previously used for one operating system into five or more operating systems running simultaneously.
We will look at downloading, installing and configuring the vSphere Hypervisor in another post. In this post, we are looking at a simple workflow to easily manage a single database applications using the hypervisor web interface. Here, we will be taking a development database and cloning it to production.
- Using the Web interface, shut down the development database.
- Using Action > Export download an copy of the Database application. This creates an .OVF (open virtualization format) file of the development database. Is saves two (2) files, the .OVF and a .VMDK (virtual machine disk).
- Create/Register a new virtual machine remembering to select “deploy a virtual machine from an OVF or OVA file”. Note: An .OVA is a single compressed file containing both the .OVF and .VMDK and may containing other files.
- Select a new name for the newly imported virtual machine. Example: “Database Production”
- Drag and Drop, or click to select both the .VMDK and .OVF you exported earlier.
- Select where you would like to store the imported virtual machine. Example: Datastore1.
- Select a network resource for the imported virtual machine. Example: Production network.
- Select either “thin” or “thick” provisioning for your virtual machine. Thin provisions means allocate disk space only when it it needed. Thick provisioning means pre-allocate all disk space to the new virtual machine.
- Wait for the virtual machine to import.
- Once complete you can power up the production database virtual machine.
You now have a clone of the development database as a new virtual machine, and a backup of the development database as an .OVF on your local disk.
Implications of the virtual machine workflow
The great thing about virtual machine is that you can repeat the cloning process as many time as you need. In fact there are many scenarios in which this process can be scripted, automated or modified to fit a wide range of business scenarios.
- A backup scenario for offsite disaster recovery.
- A continuous deployment strategy where, after testing, you can confidently deploy to production and safely roll back to a previous version.
- Easily setting up development and staging environments to test new ideas, products and integration.
In all of the above examples, we are confident that we have a backup of a working system that we can roll back and re-deploy if necessary. In each scenario the only physical constraint is the underlying physical resources our hypervisors can assign to the virtual machine. The rest is up to our imagination.
Piece of Mind
Often I tell my customers that hypervisors can offer piece of mind when it comes to business critical data. Users often don’t backup their data enough, and if they do need to restore that data it can take a while to get everything back up and running. You may need to rely on external IT, or finding your install disk and license keys. With virtual machines, you simply need to import a OVF.
Note: Using the free version of VMWare’s ESXi, you will need to shut down the virtual machine before you can export and clone a virtual machine. However with a more advanced vSphere license, cloning and migrating virtual machines can be done live, even over VPN to a backup machine offsite, in the cloud or at a home office.
Next in our series will be installation and GPU intensive virtual machines.