KubeVirt KubeVirt Summit, 24 Feb 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Building great VMs with common templates

Description

Common templates are covering most of the nowadays operating systems. Users can easily create e.g. Windows VMs, without complicated settings.

The presentation covers:
- What are common templates
- Which operating systems are supported
- How common templates work
- How to use them

Attendants will be able to know what are common templates and how to use them.

Presenter: Karel Simon, Software Engineer, Red Hat

A

Let's go for it, okay, so our next session um for our next session, we have carol simon from red hat, who will be talking about templates, a mechanism to build great vms, so you can build the ends or you can build great vms. Let's see what is this about? Thank you carol and all yours.

B

Okay, thank you so today I would like to talk about common templates and how to build vms with them.

B

At first, let's go over topics which you can find in this presentation.

B

Firstly, we will talk about what common templates are and the key factors which defines templates. Then we will talk about which operating system are supported, and after that we'll have a brief introduction. How common templates are working and at the end, I will show you how you can use them.

B

So at first, what are command templates? Common templates were created to handle common usage and scenarios of vms.

B

The idea is to reuse and enhance the concept of openshift templates and its parameter expansions, and also to reuse the keyword to be on spec structure.

B

There is one limitation that common templates are only working on openshift. That's because kubernetes currently doesn't have support for templates, but we are looking for solutions how to enable them on kubernetes as well and now two key factors.

B

There are three key factors: one of them is operating system. This allows to ensure that emulated hardware is compatible with guest os. Then there is workload. I will talk about that later and size.

B

These sizes vary the required memory and the number of cpu cores.

B

Here you can see part of the large rail8 server template. You can see that the template consists of information about operating system, like name supported versions.

B

This information is used only for or via web console openshot.console and at the bottom you can see a part of the definition of virtual machine. The vm definition is missing here, but it will be shown on a later slide.

B

Common templates have few requirements before you start working with templates. You have to install few other projects into your cluster at first. You will need keyword. Of course, then you will need cdi.

B

Cdi is a persistent storage management add-on for kubernetes template validator, it's a keyword, add-on to check the annotations on templates and to reject them if on valid and twitch ctl, where ctl is required only when you are working with templates via command line. If you are using web console, you don't need it.

B

And now, let's go to section how to install command templates.

B

There are two ways: how to install common templates first way is via web console.

B

Just look for an open shift, virtualization operator in operator up tab. uh You will select, for example, the latest version, just click on install and the command templates will be installed for you.

B

uh The second way how to install command templates is via hyper converge. Cluster operator just run the commands which are in this presentation and common templates will be installed as well.

B

Both ways of these installations, I will install all components required to run command templates into your cluster.

B

And now, let's jump to support operating system. Currently we support rel, santos, fedora windows, 10 windows, server, ubuntu and open source operating system and, of course, prs which add. Support for new operating systems are welcome.

B

And now how common templates work templates are generated from meta templates via ansible to load all necessary information about the operating system? We are using libos info, which contains info about, for example, minimal memory requirements and, as I said in previous slide, each template has a desktop server and some of the operating system have high performance workload.

B

The difference between these workloads are the desktop. Workload. Has a tablet device which other worker loads doesn't have desktop or server should be used. When you would like to have a maximum flexibility and high performance workload, uh you can use it uh when you would like to have, for example, better performance, because it has. I o threats, feature enabled by default.

B

Templates have some useful features like validation rules, and I will talk about them in the upcoming slide.

B

Then you can in template. You can set, for example, the cpu topology, like number of socket cards and thread network devices and storages.

B

Before you use the template, you have to create a data volume which will contain image with operating system. You want to run example of the data volume is in the slide.

B

After the data volume is created, you can create the vm from a template and after that cdi will start cloning. The origin latter volume and after the cloning is completed, vm can start using the clone of the original data volume.

B

uh As I said before, template has some good nice features like validations.

B

On this slide, you can see an example of validation rule. I have to mention that uh if you would like to use validation, uh you need keyword, template validator installed in your cluster.

B

Otherwise it will not work, and if you can see this rule is checking the minimal requested memory, it's checking if it has value higher than 1.5 gigahertz if it doesn't have or if the requirement is lower, this vm will not be created. The template validator will deny that.

B

You can, of course, create your own rule, which will then, for example, check number of cpu cores or, for example, if the disk volume is using sata or twit io.

B

On this slide, you can find part of the vm definition from your template. As I said before, you can set a cpu topology like number of socket cars threads. You can now set the requested memory. You can set input devices, for example, tablet in this case, network devices or.

B

Volumes and now how to use common templates, uh there are two options: how to run vm from templates. First one is via openshift web console.

B

In this case you can just go to workloads, tab click on virtualization and then go through the vm creation. Wizard, as you can see in the picture you can select uh from many operating system.

B

uh Webconf web console will then create a vm which is based on the selected template.

B

And second way: how to create a vm is via command line uh to create a vm via command line. You have to just run the oc process command. Then the name of the template. Then you have to specify the name and the namespace of the original data volume and then just run oc apply.

B

uh This will create the vm. Then you have to start the vm by where ctl start and the name of the vm and then after it's booted, you just run the ctl console and the name of the vm, and you can work with this vms with any other vm.

B

And it's all from me thank you for attending and if you have any question or.

C

Comment: hey there, we actually did have a question in uh slack chat um which is um sachin. I I don't know. Your name sorry uh wants to know if the openshift virtualization operator can be installed in vanilla, kubernetes, or should he be using the hyper-converged operator.

B

You can use both issue is that on kubernetes or on vanilla kubernetes, the templates will not work, so this feature will not be no not be there.

C

Okay, uh will the templates work on uh this is from will the templates work on okd.

B

uh Yeah, they are meant primer primarily for the openshift, slash okd, but in the future we are working on the on adding support for kubernetes as well.