KubeVirt Tech Demos, 25 Feb 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Demo: Hyperconverged Cluster Operator

Description

This is a demo of the Hyperconverged Cluster Operator, a meta-operator that deploys various components to create an opinionated KubeVirt deployment.

This demo was part of one of the sessions of the recent KubeVirt Summit. A full recording of that session is available here: https://youtu.be/6Jxbt1SzLRE

A

Hello, everyone and thank you for joining in this demo, I'm going to show you how cubit and all of its supporting operators, such as cdi for storage, cna for network and so on, all can be easily deployed in monitored and maintained by the hyperconverged cluster operator um on a okd cluster or openshift container platform.

A

Cluster hcl is not yet available in operator by default, but will be soon be there. Until then we can create a custom catalog source by navigating to cluster settings, global configuration operator sources and we create a new catalog source. Let's call it hcl catalog source and the image url of the container image is this.

A

We are publishing the most recent index image here and we're creating it once we create the catalog source.

A

A pod is being created from openshift marketplace and this processing the index image and make the package of the operator available at operator.

A

Now we can navigate to operator up and search for cube, filled keyword. You can see it here. We can see general information about the operator and description clicking on the install button.

A

We choose the most recent update channel, which will soon be changed to stable and the operator is installed on the specific name space on the cluster which is governing converged. This operator is singleton, meaning it can be only installed once spell cluster, because it's watching all the virtualization resources within the cluster. We click on install now. The first stage of the installation begins in the resources are being deployed by olm, and we can see, um we can navigate to cubic type converged a namespace, and we can see our operator is installing.

A

And the resources that it's installed are listed in the install plan objects under component. This is what going to be installed by our hco package. This is mostly crds and service accounts and role bindings, and also the csv object, which is a spatial object containing all the deployments and other resources that was defined to be deployed.

A

This are the initial deployments, mainly uh operator pods, that are in charge of watching the lcl's and acting accordingly.

A

Once we see that all pods in stage 1 have been deployed and the status is succeeded, we can proceed to creating the hco custom resource and there is the form view or the yaml view um we can set some feature gates. These are the default values and we can click on create once we click on create the custom resource is being populated with conditions telling what is going into cluster. Currently ssp is progressing and in we can see that a bunch of additional pods have been created which are deployments and demand sets and hco created.

A

This custom results for all the underlying component, such as q, word scdi,.

A

A network add-on, config, ssd and so on. You can see that all well created within a minute ago. So since we created this, the custom resource.

A

So um when all the pods are going to be ready, the hcl operator will be ready as well, and it will signal the csv. The installation is completed, pausing and waiting.

A

Okay, now the hco is going to be ready and the installation is succeeded, as we can see here.

A

If any disruption will made to the cluster and for a component, it will be reflected here in the conditions and now that the installation is completed, we can move to the virtualization page, which is shown after the installation and create a new virtual machine using the ui, and this is all I wanted to show. Thank you very much.