“Happy birthday” to customers running VMs in OpenShift! It has been a year since OpenShift Virtualization was released as a generally available feature with OpenShift 4.5. Over the last year, customers have migrated key workloads, like business critical three-tier applications, comprising enterprise databases, middleware, and web front-ends. More organizations are adopting a containers-first development and deployment model, but are challenged when most application logic and data remains in virtual machines. OpenShift Virtualization enables seamless migration and integration of virtual machine-based application components with containerized application components, all on a single platform: OpenShift.
Customers in aerospace, automotive, and financial services have enthusiastically adopted cloud-native management for virtual machines and used it as their platform to develop, innovate, and be their most agile in increasingly competitive markets.
As we focus on developing and deploying hybrid applications in OpenShift, here are a few of the highlights of OpenShift Virtualization in OpenShift 4.8.
Embrace Your Superpowers
We have heard from many customers with applications that have grown into large accidental architectures, requiring lots of engineering expertise to keep running, while making small improvements. This complexity hampers innovation, resulting in incremental improvements, regardless of the demands of the business.
One customer who has leaned into OpenShift is Lockheed Martin. They have architected and deployed highly scalable applications like AI / ML frameworks, built and managed with a small but powerful team of engineers. See more details from the Red Hat Summit session: Lockheed-Martin Takes Flight With Amazing Superpowers.
While at Summit, we also had a great conversation about our current and future plans at the “Ask the Experts About OpenShift Virtualization” session.
Simplify Storage With Automated Configuration
Kubernetes has strong architectural abstractions for storage (CSI) and networking (CNI) that VMs in OpenShift leverage to provide persistent storage and access network services. Similar to what is available for networking, where node network configuration is easy with the NMState Operator, we have improved storage configuration with the introduction of storage profiles.
OpenShift Virtualization is designed to work with any available persistent storage. Some virtualization features, like live migration and VM snapshots, depend on storage features being enabled in Storage Classes and PVCs. To reduce complexity while increasing performance and functionality, OpenShift Virtualization can choose optimal settings when creating PVCs for virtual disks. AccessMode and VolumeMode can be chosen based on the storage type, and size can be padded as needed, to accommodate storage overhead. For now, this automation is opt-in when creating your own DataVolumes, and will be extended to other workflows in a future release.
Another improvement is performance of DataVolume cloning across namespaces is dramatically improved. Cluster administrators can create golden images using VM templates in a dedicated namespace; new VMs can be instantly deployed in any namespace within the cluster, accelerated by the cloning offload capabilities of the CSI storage provisioner.
VMs in OpenShift work with all CSI compliant storage; you can leverage existing storage solutions or deploy newer cloud-native storage providers, like Red Hat OpenShift Data Foundation. We have extended our OpenShift CSI certification end-to-end test suite with VM workloads; you can be confident that certified storage Operators from our valued ecosystem partners work with your VM workloads.
Accelerate Compute Intensive Workloads
One of the most compute intensive workloads these days (outside of High Performance Computing) is AI/ML and video rendering. These are almost always accelerated with dedicated hardware like GPUs. VMs in OpenShift can directly access GPUs for full access to these powerful cards. Workloads that may not be easier containerized, like implementation specific compute and rendering pipelines, can run in OpenShift on Microsoft Windows or Linux guests.
GPU resources are expensive, and accessing them with both containers and VMs within the same platform gives you flexibility to offer self-service technical workstations.
VMs deployed in OpenShift are a native part of Kubernetes ecosystems, and guest workloads are directly visible within OpenShift. If you are coming from a traditional virtualization platform, you will be comfortable with the additional VM workload metrics available in Prometheus.
More Flexible Configurations
We also added capabilities to dynamically add disks to a running VM, giving more flexibility to manage persistent data between workloads, currently in Tech Preview. As VMs are first-class citizens in the Kubernetes architecture, things like “immutable pods” make it a little more challenging to implement, but these improvements are completely native.
We enhanced high performance network workloads, by enabling live migration for VMs using SR-IOV. Seamlessly deploy, migrate, and upgrade your high performance workloads with minimal disruption.
Even Easier to Get Started
We have made it easier to get up to speed with VMs in OpenShift with new scenarios based documentation. Whether you are a cluster administrator or developer, it is easy to learn and be productive, even if you are new to OpenShift.
We are also excited for the Migration team, who recently announced that Migration Toolkit for Virtualization 2.0 is generally available. The Migration Toolkit for Virtualization, based on upstream Konveyor Forklift, makes it easier than ever to migrate and import VMs from other hypervisors to OpenShift at scale.
All versions of OpenShift are validated in the Microsoft Windows Server Virtualization Validation Program (SVVP), which ensures you are fully supported by Red Hat and Microsoft for workloads running Windows Server 2012 R2, Windows Server 2016, Windows Server 2019, and Windows 10.
OpenShift 4.8 with OpenShift Virtualization continues our aggressive delivery of capabilities for all of your virtual application workloads. It is easy to modernize your applications by leveraging the critical business logic and valuable data locked in your traditional virtualization platform, while using virtual machines in Kubernetes in a very cloud-native way. OpenShift Virtualization is included at no additional cost with all OpenShift subscriptions, regardless of whether you choose OpenShift Kubernetes Engine, OpenShift Container Platform, or OpenShift Platform Plus.
To learn more about what has been added in this OpenShift Virtualization release, see the OpenShift Virtualization Release Notes for details. You can see the bigger picture in the What’s New in Red Hat OpenShift 4.8
Find Out More
Upstream Project and Community
Red Hat is a leading contributor to KubeVirt, the upstream CNCF project that users and vendors are adopting to integrate virtual machines in Kubernetes. We are also seeing additional companies choosing KubeVirt as a proven key technology. Our upstream-first approach ensures everything is open, secure, and stable for even your most challenging enterprise workloads.
KubeVirt Summit - February 2021