Back in 2022, VMware unveiled vSphere 8 during the famous VMware Explore event in Las Vegas (formerly known as VMworld). That major version brought numerous new features and improvements, among which we find VSAN ESA, high performance workloads improvements, Green metrics, support for Data Processing units (DPU), the end of vSphere Update Manager in sight, vSphere configuration profiles and more.
Along with all this also came a new release cycle which now includes Initial Availability (IA) before General Availability (GA). As a result, vSphere 8 Update 1 will go through this cycle. The drive for this change was mostly due to all the issues customers encountered with vSphere 7 which brought several production environments down and drove VMware to pull vSphere 7 Update 3 from downloads as it was buggy.
There isn’t much data out there about the current adoption rate of vSphere 8, however, VMware hopes to entice customers to keep their environments up to date with vSphere 8 Update 1 which brings with it, a number of interesting improvements and changes. In this blog, we will go through these one by one.
vSphere Configuration Profiles
vSphere Configuration Profiles were announced with vSphere 8 as a technology preview to improve on Host profiles. The legacy host config tracking feature is quite tricky and often deviates from the baseline for whatever reason which drives admins to just skip a bunch of checks, essentially crippling the capabilities of the feature.
vSphere Configuration profiles aims to offer a better and less clunky alternative (or should I say replacement). In vSphere 8 Update 1, it is a fully supported feature that allows Administrators to manage the host configuration at a cluster level. Note that it currently doesn’t support environments using VMware NSX.
It will allow you to
- Set desired configuration at the cluster level in JSON format
- Check that hosts are compliant with desired configuration
- If non-compliant, remediate hosts to bring them into compliance
- Support for vSphere Distributed Switch configuration
Note that if the cluster is still using baseline-based life-cycle management, it must be converted to use image-based life-cycle management.
vSphere Lifecycle Manager for Standalone Hosts
vSphere 8 improved host Lifecycle Manager for standalone ESXi hosts using vSphere APIs. In vSphere 8 Update 1, the vSphere Client now supports standalone ESXi hosts to compose a desired image, remediate, check compliance… Pretty much everything you expect to be able to do in vSphere Lifecycle Manager on a vSphere cluster will be available for standalone hosts, including staging and ESXi Quick Boot.
Now, I would be interested to know what drove that decision, if it is edge compute cases or simply to reduce the risk of wildly outdated standalone ESXi hosts in the wild.
A nice feature that comes with it, which applies to edge cases is the ability to define custom image depots, which should address edge environments where the depot sits in the same location as the hosts that has poor connectivity. In this case, the custom depot will act as a staging location to avoid running a task for hours and risking problems due to slow or high latency connections.
Host different GPU workloads on a single GPU
In previous versions of vSphere, all NVIDIA vGPU workloads on an ESXi host must use the same vGPU profile type and GPU memory size.
vSphere 8 Update 1 allows you to apply different vGPU profile types (A, B, C, Q) to your virtual machines to share resources more effectively across workloads. Note that the memory size on each profile must be the same though.
| Series | Optimal Workload |
|---|---|
| Q-series | Virtual workstations for creative and technical professionals who require the performance and features of Quadro technology |
| C-series | Compute-intensive server workloads, such as artificial intelligence (AI), deep learning, or high-performance computing (HPC) |
| B-series | Virtual desktops for business professionals and knowledge workers |
| A-series | App streaming or session-based solutions for virtual applications users |
Integration of VMware Skyline Health Diagnostics with vCenter
Skyline Health Diagnostics is a self-service diagnostic platform that helps vSphere administrators acquire insights inside their environment. It provides issues diagnosis, failure resolution in an automatic fashion, scans for VMSA and identifying impact of future upgrades.
VI admins use it to troubleshoot issues before contacting support and in vSphere 8 Update 1, Skyline Health Diagnostics is integrated with vCenter.
VM Service Bring Your Own Image
Back in vSphere 7 Update 2a, the VM service was introduced which lets you initiate the creation of virtual machines from within a Kubernetes cluster, in a way comparable with the likes of Cross-plane or Cluster API.
In vSphere 8 Update 1, support for custom images from the customer has been added. Images must support CloudInit or vAppConfig to configure the VM on first boot. These images are added as VM templates to a Content library.
Here is an example of a manifest to create a virtual machine from Kubernetes. In the example below, the contents of the CloudInit to configure the machine must be in a configmap named tensor-configmap.
apiVersion: vmoperator.vmware.com/v1alpha1
kind: VirtualMachine
metadata:
labels:
vm-selector: vmware-tanzu-jumpbox3
name: vmware-tanzu-tensor
namespace: tkg
spec:
imageName: ubuntu-jammy-vm2
className: best-effort-xsmall
powerState: poweredOn
storageClass: k8s-shared
advancedOptions:
defaultVolumeProvisioningOptions:
thinProvisioned: true
volumes:
– name: tensor
persistentVolumeClaim:
claimName: tensor-pvc3
networkInterfaces:
– networkType: vsphere-distributed
networkName: k8s-workload
readinessProbe:
tcpSocket:
port: 22
vmMetadata:
configMapName: tensor-configmap
transport: CloudInit
Supervisor Services now support vSphere Distributed Switch
Supervisor Services are vSphere certified Kubernetes operators to offer DevOps engineers access to the service APIs so they can create their own instances on Supervisor Clusters in their own namespaces. They can then be used either in vSphere Pods or Tanzu Kubernetes clusters.
The supervisor service catalog includes
- Backup and Recovery service backed by Velero
- Certificate management service
- Registry service backed by Harbor obviously
- Ingress backed by Contour
- And more…
In vSphere 8 Update 1, in addition to the VM Service, Supervisor Services are now available when using vSphere Distributed Switches.
Nvidia NVSwitch Support
If you haven’t heard of these, NVIDIA NVSwitch is a high-performance networking fabric designed for interconnecting multiple GPUs and servers in a datacenter environment. vSphere 8 Update 1 brings support for these NVSwitches for high performance and connects up to 8 GPUs to the same VM.
Nvidia NVSwitches have a lot of use cases in high-performance computing (HPC) and AI (Artificial Intelligence) where applications require more than one GPU. Because some applications hit a bottleneck at the bus level, the NVSwitch offers greater bandwidth. For instance, PCIe Gen5 x16 can transmit up to 64 GB/s while NVSwitch backed architecture can transmit 450 GB/s both ways with a pair of GPUs.
Improved vSphere Green Metrics
vSphere 8 introduced a new set of metrics to vSphere with the green metrics. vSphere Metrics work for hosts and VMs and track energy consumption to establish the efficiency of your infrastructure.
vSphere green metrics include
- Host energy consumption not used by VMs (overhead)
- Host is idle consumption
- Host’s consumption used by VMs
- VM-level power consumption taking VM size into account (vSphere 8 Update 1)
Usage metrics are calculated based on VM resource usage compared to the hosts power usage which is collected through IPMI integration.
There is another interesting metric which is static power. This represents what the power draw would be for that particular VM if it were a bare metal physical host.
Authentication with Okta federation
vSphere Identity management in vSphere supported Active Directory and OpenLDAP up until now. It is great to see that vSphere 8 Update 1 brings support for Okta for both vCenter and NSX Manager as a 3rd party identity providers.
This should greatly simplify identity federation and SSO, especially with the rise of Cloud Native adoption and how popular providers like Okta are when put behind something like Dex.
VMFS and vSANDirect Disk Provisioning Storage Policy
Cloud Native Storage is a capability of vSphere that allows Kubernetes to provision disks to populate PV (Persistent Volumes) with a native vSphere disk (VMFS, VSAN). This works through the CSI (Container Storage Interface) and lets administrators visualize disks and space utilisation of Kubernetes volumes in the vSphere client.
Here is an example of a manifest to create Kubernetes storage class that can be used to provision disks directly on VSAN:
kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
name: vsan-default-policy
annotations:
storageclass.kubernetes.io/is-default-class: “true”
provisioner: csi.vsphere.vmware.com
parameters:
storagepolicyname: “vSAN Default Storage Policy”
In vSphere 8 Update 1, you can choose EZT, LZT, or Thin provisioning via the SPBM policy for cloud native storage. It will facilitate compliance checks in SPBM for volume allocation rules in a storage policy such as the operations supported on disks and FCDs (First Class Disks).
If you are interested in the topic, we discussed cloud native disks in VMware Cloud Director (VCD) in the blog already: https://www.bdrsuite.com/blog/vmware-cloud-director-named-disks/
Other improvements
A number of other enhancements are included in vSphere 8 Update 1 such as
- ESXi Quick Boot for Secure Systems
- vSphere Fault Tolerance with vTPM
- VM DirectPath I/O Hot-Plug for NVMe
- New vVols VASA 5.0 Spec
- NVMe-TCP support for vVols
- Maximum WSFC clusters per ESXi host from 3 to 16
- Enhanced XCOPY to Datastores Across Different Storage Arrays
- NFSv3 vmkPortBinding
- vVol VM Swap improvements for faster Power on/off and vMotion performance
- Lower minimum reclamation rate for UNMAP to 10MBPS to reduce disruption on low performing storage backends
- Improved NFS resiliency with retry NFS mounts on failure and NFS mount validation
Wrap up
In conclusion, vSphere 8 Update 1 represents a significant step forward with a range of new features and enhancements. This update delivers improved performance, security, and scalability, making it an essential upgrade for any enterprise looking to stay ahead of the curve in today’s fast-paced IT landscape.
Follow our Twitter and Facebook feeds for new releases, updates, insightful posts and more.
