Red Hat OpenShift Virtualization (based on Kube-Virt) enables to run VMs on Kubernetes cluster. Kubernetes is becoming the platform of choice for many customers to run their containerized workloads. Kubernetes cluster can have 1000s of nodes. Scheduling the workloads across the nodes and restart of failed workloads are handled as you expect with the vSphere HA cluster and DRS features. Tags are utilized for placement affinity rules. Like vMotion, VMs can be migrated from one host to another using Live Migration.
Virtual machines can be created using the OpenShift Console or define it on YAML files and apply it similar to container workloads. You can also clone from VM template. The templates are stored on Image registry. Similar to Content Library on VMware environments, the image registry can be used to distribute the templates to multiple locations.
As VAAI (vSphere API for Storage Array Integration) which offloads storage operations to storage array, Kubernetes uses Container Storage Interface (CSI) to handle the storage operations specific to array features. For ONTAP, we provide Trident CSI to handle storage provisioning, cloning, snapshots, etc.
Trident supports both file (NFS v3,v4.x, SMB) and block (iSCSI, NVMe/TCP) protocols. Similar to vSphere datastores, you have the option to store multiple VM data disks on same ONTAP volume with trident economy drivers or use dedicated ones like vVols with nas, nas-flexgroup or san drivers.
VM data disks are like First Class Disks (FCD) which are handled with Persistent Volume Claims (PVC) to request size with specific attributes. VM data disks (PVC) can be protected with Astra Control for backup and restore operations, changing the storage class (like storage vMotion), etc.
For high availability requirements, Metro cluster solution can be consumed with Trident to provide similar solution with vSphere Metro Storage Cluster. On AWS, Multi-AZ FSxN along with trident can be consumed to tolerate on AZ failures.
OpenShift Virtualization supports bridge & overlay networking like distributed vswitch and NSX-T environment. OpenShift Virtualization uses Multus to handle multiple networks with CNI (Container Network Interface) to separate the data networks from the management network. VMs can also be placed on specific VLAN networks.
GPU support for VMs is also available. It can be consumed as passthrough PCI device, vGPU profiles or using SR-IOV.
OpenShift Virtualization can be deployed anywhere from Edge to Core to Cloud. OpenShift Pipelines or other Kubernetes tools can be used for CI/CD. Ansible Automation Platform is often used for automation like Aria Automation to integrate with Servicenow for catalog requests of Virtual machines or applications. For centralized monitoring, various Kubernetes monitoring tool including Cloud Insights are available to handle operations similar to Aria Operations.
Red Hat Advanced Cluster Management for Kubernetes handles centralized policy management across the multiple instances of clusters. Astra Control handles ONTAP Kubernetes data operations across the clusters.
To explore the features on your own, install Virtualization Operator from OperatorHub into an existing OpenShift Cluster and have baremetal nodepool into the cluster to host the VM pods. If you prefer to consume as service, you can explore with Red Hat OpenShift Service on AWS (ROSA) with FSxN or similar services from other cloud providers.
To migrate existing VMs from vSphere to OpenShift Virtualization, migration toolkit operator is available and use that toolkit to perform bulk or single VM migration to try. For additional details, refer https://docs.netapp.com/us-en/netapp-solutions/containers/rh-os-n_use_case_openshift_virtualization_workflow_vm_migration_using_mtv.html
Our team will be available in KubeCon as well as in Red Hat Summit. Feel free to stop by booth for further interactions.
