Exploring compute options in Red Hat OpenShift as a VMware admin
Monitoring, migration, resource balancing, there are several compute options that go into any tool for managing your infrastructure. In this section, we will explore common features in Red Hat® OpenShift®, and what Red Hat® OpenShift® Virtualization compute menu options mean for VMware vSphere admins.
What will you learn?
- How OpenShift menu and features map to ones you may know from VMware
What do you need before starting?
- VMware vSphere 8.0 or higher (as referenced in this path)
- Red Hat OpenShift 4.18 (as referenced in this path)
Compute feature mapping chart
This section addresses commonly used compute features in VMware and how they map to deploying, managing, and maintaining virtual machines in OpenShift.
| Feature | VMware | OpenShift |
| Resource balancing | Dynamic resource scheduling (DRS) | Descheduler pod eviction policy |
| Host / VM metrics | vCenter, Aria Operations | OpenShift Metrics and Monitoring, Red Hat Advanced Cluster Management |
| Compute Live migration | vMotion | Live Migration |
| CPU overcommitment | Yes | Yes |
| Memory Overcommitment Methods | Ballooning and Transparent Page Sharing (TPS), Swap | KSM and free page reporting |
| (Co)Scheduling constraints | Affinity rules for VMs to VMs and VMs to Hosts | (anti)affinity rules for VMs, Pods and hosts |
| Node failure | Node failure detection and VM rescheduling happens within 15-30 seconds, even without vCenter | Supported through fencing agents' environments. See KCS for HA configuration. |
| Dynamic reconfiguration | CPU, Memory, Disk, network and some additional hardware is supported for hot add/remove and reconfiguration | Hot add supports storage, CPU, SR-IOV and Bridge Network, and memory |
| Compute acceleration | PCI passthrough for CPUs, VGPU supported | GPU, generic PCI passthrough, and CPU passthrough supported, (NVIDIA) vGPU supported |
| Template management | Template VMs, OVA/OVF deployment, and content libraries offer the ability to provision VMs using a simplified process | Catalog to manage boot sources, VM templates and instance types. VMs can be created from VM templates, instance types or declarative raw VM definitions when associated with a specific boot source. |
| VM export/import | Import and export VMs using the OVF and OVA formats | Import and export OVA/VMs using migration toolkit for virtualization (export example). |
Menu mapping chart
This section addresses what compute menu selection items in OpenShift Virtualization mean from a VMware vSphere perspective. As noted in previous resources, OpenShift is conceptually different in select areas from VMware to achieve the same or similar items. These instances will be marked as “N/A”.
| OpenShift Virtualization menu item | VMware comparable | Explanation |
| Nodes | Host and Cluster > Host | List and troubleshoot current nodes in the cluster and their configuration and runtime details |
| NodeHealthChecks | Config > vSphere DRS | Configure remediation actions for different node states. There is a range of remediation operators that go from soft rebooting a node that is in Not Ready or Unknown state, hard rebooting it from the BMC APIs or forcing a reinstallation of the node after unsuccessful reboot remediations |
| Machines (Automation) | N/A | List and troubleshoot the metadata of nodes installed and automated through the OpenShift Machine API operator using manual or automatic scaling |
| MachineSet (Automation) | N/A | List and troubleshoot the metadata of nodes installed and automated through the OpenShift Machine API operator using manual or automatic scaling |
| MachineSet (Automation) | Host Profiles | Configure automation profiles to manually or automatically scale infrastructure. A scaled MachineSet will create a Machine object to allocate an available bare metal host and trigger the installation and configuration defined in the MachineSet |
| MachineAutoscalers (Automation) | N/A | Define autoscaling rules to scale a MachineSet |
| MachineHealthChecks | N/A | Alternative mechanism to NodeHealthChecks when the desired remediation is to delete a Machine, which triggers a fresh installation of the node. In cloud providers it causes the deletion of the node and the creation of a new bare metal node |
| Bare Metal Host (Automation) | N/A | List and operate bare metal servers added to the cluster. Bare metal hosts include manually installed servers and servers made available for and installed by MachineSet scaling. Integrated with BMC APIs it enables full server lifecycle |
| MachineConfigs (Automation) | Host Profiles | Manage declarative host configurations to ensure consistent groups of configurations per machine pools and avoid configuration drifts |
| MachineConfigPools (Automation) | Host Profiles | Manage different host configuration pools allocating MachineConfig objects to nodes based on roles. MachineConfigPools also allow to segment nodes and define different HA policies when rolling platform updates/upgrades |
| Hardware Devices | N/A | N/A (specialty hardware enablement) |
Now that you have explored the common compute options available in OpenShift as to how they pertain to VMware vSphere, it's time to see comparisons between observability offerings.
