Containers and Virtual Machines Together on Red Hat OpenShift Platform

It is well known that the Red Hat OpenShift Platform is the best choice for enterprises to manage container workloads. In this blog, you will find how to use OpenShift Virtualization to deploy Virtual Machines and containers side by side to run your hybrid workloads.

Red Hat OpenShift Platform uses Kubernetes in its core and brings benefits in several areas of your enterprises, including:

• Consolidation of workloads
• Ease of management
• High availability
• Accelerated development processes

With the GA of OpenShift Virtualization in OpenShift 4.5, enterprises can benefit from running virtual machines next to your containers without overly complicating your application architecture. With OpenShift's declarative language, deploying a cloud-scale mixed architecture is easy.

You might wonder why do I need this? If you look at the market, new software architectures are designed and produced with microservices. Can't I keep OpenShift just for containers, you wonder? In the past, enterprises have invested heavily in virtual machine architectures and still have lots of legacy applications--for example, a windows .net application or an RDMBS-- that depend on them. Even many new applications may rely on these legacy VMs, sountil applications are modernized, they still need to be run in VMs. Legacy applications may also need to be divided into multiple functions and move to microservices gradually. To achieve this and progressively and incrementally move to more modern architecture, you need to run containers and virtual machines side by side.

With the Red Hat OpenShift platform, you get a single pane of glass to manage your container and virtual machine workloads. OpenShift provides you with a unified set of tools and processes, including CI/CD pipeline in a full-scale cloud platform, both on and off-premises, to run both workloads.

In a production environment, you will need a bare-metal deployment of OpenShift to experience OpenShift Virtualization. If you are experimenting in your lab, you may consider nested virtualization, but please keep in mind the performance impact of such deployment.

The installation of OpenShift Virtualization is quite easy and done entirely using operators. All you need is a Red Hat OpenShift Platform subscription. Operators handle the rest of the installation for you. Please follow the link below for detailed installation instructions:

https://docs.openshift.com/container-platform/4.4/cnv/cnv_install/installing-container-native-virtualization.html

After installation, you will see Virtualization enabled for your cluster.

You might wonder how this works in a container platform. To summarize, if you have used a Linux-based virtual machine platform, you would have seen one or more "qemu-kvm" processes running with the virtual machine name and a long list of parameters defining the machine's hardware resources. With the OpenShift Container Platform, we put this single process in a container with all its bells and whistles to provide you with the VM experience you are looking for.

KVM is a widely used virtualization technology and well matured for production environments. It is being deployed in many critical environments in all types of verticals.

Demo Time

Here in this demo, you will be deploying the well-known Guestbook application, but with a slight change. The master Redis node will be a virtual machine. The rest of the application components, specifically Redis Slaves (two counts) and frontend (three counts), will run as containers. It will also use the network, IP addresses, storage, and load-balancer from the OpenShift platform itself.

First, you need to upload the Operating System qcow image to the Red Hat OpenShift Platform. It will help quickly deploy virtual machines. You will upload it to a ReadWriteMany storage backend provided by the Red Hat OpenShift Container Storage.

Before you begin, you should have installed the OpenShift Virtualization operator on the platform and logged into your cluster with appropriate permissions.

Deploying this lab is easy, and you can find yaml files at: https://github.com/ansonmez/openshiftvirtualization

Please follow the link below to install “virtctl,” which is the command used to upload virtual machine images to OpenShift:

https://docs.openshift.com/container-platform/4.4/cnv/cnv_install/cnv-installing-virtctl.html

Follow the commands below to get started.

Create a new project for guestbook:

# oc new-project guestbook

Download the centos cloud image from cloud.centos.org:

# curl -OLJ https://cloud.centos.org/centos/8/x86_64/images/CentOS-8-GenericCloud-8.1.1911-20200113.3.x86_64.qcow2

Upload the downloaded image to OpenShift. Use the already defined ReadWriteMany storage with 11G disk space and with Persistent Volume Claim name of “centos8”:

# virtctl  image-upload \
 --image-path=CentOS-8-GenericCloud-8.1.1911-20200113.3.x86_64.qcow2 \
 --pvc-name=centos8 \
 --access-mode=ReadOnlyMany \
 --pvc-size=11G \
 --wait-secs=1800  \
 --insecure

Create a new data volume using the command below. Please inspect the “centosdv.yaml” file to ensure the parameters you want:

# oc create -f centosdv.yaml

Finally, create the Template to be used by users. Please go through the template to ensure you have all required parameters set:

# oc create -f guestbooktemplate.yaml

If you take a look at the template below, you can see that we define a route, a couple of services, a virtual machine, and a couple of containers using ReplicationControllers. You might have some idea by now, and yes, you could keep your virtual machine definitions along with your container definition in a source code repository and maintain it just like any other container workload.

Run the Newly Created Template

Now it is time to run the template. From the Red Hat OpenShift Platform web interface, you can click on Developer → Add → From Catalog → Search Then after you type “guestbook” in the search box, you will see the screen below:

Go ahead and order “Guestbook Demo - Multi tier with CNV” application. It will ask you to provide the namespace and Name of the application. Provide them according to your environment requirement. For example, Namespace = guestbook , NAME=testguest. See below:

Once launched, you can observe the pods and servers running from the OpenShift console or from the CLI as shown below. As you can see, “virt-launcer-redis-master-*” pod is the pod containing the Redis master virtual machine:

Make sure that Redis replicas are in sync with Redis master. Check the logs of Redis replica pods:

Finally, If everything is on track, you can get the URL of the guestbook application using “oc get route” command as shown below:

Type it in or copy and paste the URL in a web browser. and you will get access to the shiny new Guestbook application with a VM based Redis master:

Wrapping Up

This was only a sample application to get you started. But the potential for running a VM along with your containers are much more. As stated, it gives even greater flexibility to the application designers and developers. For the infrastructure operation team, it gives reduced complexity around managing the VMs and connectivity between different infrastructure platforms, including the security nightmare that you might encounter setting up such an infrastructure.

There are more example templates that you could explore in the developer catalog that you could use for tests or probably for hosting your application:

OpenShift Virtualization is GA and continuously improving, so stay tuned for features and experience it in your environment. Check out the release notes, and find more information on what's next and new here.