RHV vs OpenShift Virtualization

[Chetan_Tiwary_]

RHV is a good choice for organizations that need a traditional virtualization platform with a lot of features and flexibility. It supports a wide range of hypervisors, including KVM, VMware, and Xen. It also supports a wide range of workloads, including virtual machines, containers, and bare metal.

OpenShift Virtualization is a good choice for organizations that are looking for a platform that is specifically designed for containerized workloads. It is built on top of Kubernetes, which is the most popular container orchestration platform. It also integrates with other Red Hat products, such as OpenShift Pipelines and OpenShift Service Mesh.

OpenShift Virtualization allows you to leverage existing VM investments while moving towards a cloud-native future. It delivers a consistent, efficient, and secure platform for deploying both traditional and next-generation applications, helping you future-proof your virtualization strategy.

https://access.redhat.com/support/policy/updates/rhev 

Refer OpenShift Virtualization here : https://www.redhat.com/en/technologies/cloud-computing/openshift/virtualization 

Courses to refer RHV : RH318 : https://www.redhat.com/en/services/training/rh318-red-hat-virtualization  with exam EX318.

OpenShift virtualization : DO316 : https://www.redhat.com/en/services/training/do316-managing-virtual-machines-red-hat-openshift-virtua...  with exam EX316.

[diac]

Hey everyone.

I have been a RHV user for a few years now and I was pretty happy with it.
When I heard RH was killing RHV in favor of Openshift Virtualization...
Let's just just say I was more than a little bit ticked off.

With RHV you could build simple, relatively low cost, functional and highly performing virtualization environments with nothing more than a couple of bare metal servers and a NAS.
Implementing the same solution with Openshift Virtualization is a whole other ballgame, minimum 3 nodes, way more expensive and quite a lot more complicated.

RHV was ideal for small to mid sized companies that had no need to invest heavily in container infrastructure.
And yes, such relics will continue to exist.
Now that RHV is gone I'm starting to see customers move to Oracle Linux Virtualization, based on ovirt.

And then there is performance.
I must admit I was a bit shocked when I read this report by Principled Technologies.
https://www.principledtechnologies.com/Vmware/vSphere-7U2-vs-Red-Hat-OpenShift-4-9-0322.pdf
Just do yourself a favor and browse through it...
I am very curious to know if RH has commented on the report or published a report to counter this one.

I'd love to get your take on RHV vs Openshift Virtualization.
In my mind these two solutions should have co-existed for years to come, to target a wider audience.

Best regards,
Sigurdur Bjornsson

[Fran_Garcia]

Red Hat had 4 virtualization solutions, and maintaing all 4 doesn't make sense technically or commercially:

- KVM on RHEL

- Red Hat Virtualization, based on oVirt

- Red Hat OpenStack

- Openshift Virtualization

For the smallest usecases (two servers), you can still use RHEL+KVM+Shared Storage (eg NFS). This set up is included by regular RHEL and supported with your regular subscription. You still have a (limited) Web UI as part of the latest RHEL9 developent in Cockpit / RHEL Web Console. This blog is a bit dated, but you can get a taste on how that would look like: https://www.redhat.com/en/blog/managing-virtual-machines-rhel-8-web-console

Re oVirt: Red Hat has been calling for community collaborators to help with development with the oVirt project for years, both before and after the RHV product announcement. Unfortunately those calls for actions have not been followed by any user/group of users/developers of other commercial distributions of oVirt. Red Hat's priority with this transition has been to ensure the community built around oVirt was left in a good place to continue the development. Unfortunately, it seems no one wanted to pick the torch. Anyhow, the oVirt project has all their development resources now in public infrastructure (Github and others), so if anyone wants to continue the development, they are able to do so. https://blogs.ovirt.org/2022/02/future-of-ovirt-february-2022/ 

Re the "study", "This project was commisioned by VMware". I don't think there's an official response to everything anybody says...

As for the future of on-prem virtualization: we can speculate as much as we want, specially if you take into account what other vendors are doing. IMHO some large on-prem customers will use virtualization; some will move to the cloud because it's still more efficient to do so, and many will start modernising their applications and benefit from a hybrid VM and Contianers platform. Openshift Virtualization can do both, well, and is being actively developed by a large community with multiple upstream collaborators, partners and users.

Cheers

Fran

[ipalagin]

I think Red Hat made a strategic mistake by killing RHV.  Many companies are looking to migrate off VMware these days.  But I don't see VMware admins start using OpenShift Virtualization easily, it's very different and requires thorough learning.  RHV would be much easier to adopt.

[Chetan_Tiwary_]

I dont think it is a strategic mistake because OpenShift Virtualization offers a path for infrastructure modernization, taking advantage of the simplicity and speed of a cloud-native application platform and aims to preserve existing virtualization investments while embracing modern management principles.

Red Hat sponsored a survey of 1,000 IT decision makers (ITDMs), backend developers, and software architects in the United States (U.S.), the United Kingdom (U.K.), and the English-speaking Asia Pacific region (APAC)  last year in october - to understand how organizations plan to approach modernization and migration and how they define application modernization success.

Some of the survey resuts were : 

1. 95% of respondents believe that application modernization is essential for their organization’s success. 
2. Companies plan to modernize 51% of their custom applications within the next year.
   75% of companies have already completed at least small-scale modernization projects, and 18% have progressed all the way to continuous modernization.
3. Application modernization describes many different initiatives, including improving continuous integration/continuous deployment (CI/CD) pipelines and data modernization.
4. Security, reliability, and scalability are the most cited reasons to modernize, metrics to measure efforts, and overall benefits of application modernization.
5. Organizations and individuals face interrelated challenges when it comes to modernization, and they’re addressing these challenges with new tools, application programming interface (API)-driven development, and training.

With flexible deployment and consumption options, Red Hat OpenShift supports a variety of use cases—including application development, AI/ML adoption, edge computing, and IT modernization initiatives—across environments.

And one of its component - Red Hat OpenShift Virtualization aims to simplify managing your entire infrastructure by providing a single platform for virtual machines (VMs), containers, and even serverless functions. This means you can standardize how you deploy and maintain everything using the same set of established, enterprise-grade tools. Migrating your existing VMs to OpenShift allows you to continue using those investments while gaining the benefits of cloud-native architectures. These benefits include streamlined operations and management, access to modern development approaches, and faster time to market thanks to features like self-service options and integrations with CI/CD pipelines. This focus on a unified platform aligns with a current market trend where organizations are increasingly prioritizing modernizing their existing infrastructure and applications over building entirely new ones.

So, I believe strategically the move to RHOV is really growth and goal oriented. 

[ipalagin]

Hi,

There are about 500,000 VMware customers in the world. Most of them would love to migrate off this platform because of the recent %200-%500 cost increase, caused by Broadcom.

Most of the VMware customers won't be ready to move their VMs to OpenShift Virtualization because of the learning curve and changes to other process like VMs backup and disaster recovery. RHV was much closer to VMware than OpenShift Virtualization, and it would be easier adopted, making profits for Red Hat.

Thanks,

Ilya.

[Chetan_Tiwary_]

@ipalagin  While it may not be immediate for many of them but I think  the ongoing advancements in Kubernetes and the Cloud Native Computing Foundation (CNCF) project ecosystem hold promise for the future. These innovations could pave the way for a practical approach to integrating traditional VM-based workloads with modern applications for example they can then be decomposed into microservices on containers over time, or maintained as VMs.

All I can say is future for virtualization on RHOCP looks bright : 

[diac]

My $0.02..

I am a Red Hat Certified Specialist in OpenShift Virtualization.
I had serious doubts about the product and I thought that getting certified would be a great way to confirm those doubts or to be proved wrong.
Long story short, as a sysadmin of 20+ years with Linux, Windows, Unix, VMware and RHV experience, I'd rather not touch OpenShift Virtualization with a 10 foot pole.
My reasons are primarily: 
 - Complexity on another level.
 - Backup options.
 - Performance
 - Way too expensive compared to RHV.
 - Lack of CPU and memory hot-plug.
 - No support for hot-plugging of virtio disks.
 - No hot-unplug.
 - No storage migration.
 - Don't get me started on networking...
 - Almost everything about it is more complex than RHV or VMware, and yet it's lacking important features that we've had in RHV and VMware for over a decade.

I'm afraid that Red Hat's decision to be hip and cool instead of giving their customers what they wanted and needed, will bite them in the end.
RHV would have sold by the boatloads because of the VMware/Broadcom fiasco, of that I'm sure.

It doesn't matter how much marketing material you throw at our face. Openshift Virtualization should not, in my opinion, be advertised as a replacement for RHV.
No one mentions the shortcomings.
If I was was running a "container first" company and also needed to run a dozen legacy VMs, then fine, let's utilize the existing Openshift installation and run them with Openshift Virtualization. But to replace RHV and manage hundreds or thousands of VMs on Openshift Virtualization.. Hard pass!

I'm not saying that Openshift Virtualization will never be usable. It may very well be, in the future. But closing the curtains on RHV when they did, was way premature.

[ipalagin]

A few more missing features:

- Too many tasks can't be completed from the WUI and require YAML editing.  Creating PVC?  Click on YAML and add a few details.  Labelling a VM to expose its web service? Nah, that label is for different purpose, open YAML editor and add your label under metadata.

- Misguiding raw WUI. Disk has "Edit" option?  Ok, go ahead and edit.  Saved?  No changes? Hah, it's because you can't edit this resource!  You need to delete and re-create it.

- No cloning for running VMs, no sealing VMs from the WUI.

- No central task/events log with events ID, usable out of the box.  For example, I'm creating a VM snapshot, and it fails.  To find this state on OpenShift, I would have to remember to check the VM snapshot state from WUI or CLI, and if I stop the VM, the launch pod will be gone along with its log.  In VMware/RHV, the VM icon would have an alert + record in the event log, which is easy to find. Again, I'm sure it's possible to setup Prometheus with all required monitoring in place, but what effort will it be?

It appears that RedHat invested 99% of the OpenShift Virtualization budget into marketing in 1% into technology.  Result: convincing sales pitch for executives and a nightmare for sysadmins.  Sign of successful integration between IBM and RedHat -  there was a joke "if you want to torture your Java developers, buy them WebSphere".