In vSphere 8.0 Update 3 we introduce Embedded vSphere Cluster Services (Embedded vCLS). Based on vSphere Pod technology and managed at a system level, we’ve focused on reducing impact to day-to-day operations while maintaining full compatibility.
Embedded vCLS is automatically activated on vSphere clusters that contain at least one ESXi 8 Update 3 host and is managed by vCenter 8 Update 3. During updates from older versions, the cluster automatically transitions to Embedded vCLS once at least one host is updated to ESXi 8 Update 3.
Note: You must always update vCenter to 8 Update 3 before updating hosts. See the VMware Product Interoperability Matrix for more information.
Why do we call this Embedded vCLS and what has changed?
Embedded vCLS uses the CRX runtime technology first developed for the vSphere IaaS control plane (formerly vSphere with Tanzu).
Embedded vCLS VMs have no storage footprint and run entirely in host memory. The ESXi host spins up the Embedded vCLS VM(s) directly. There is no OVF deployment pushed from vCenter and EAM (ESX Agent Manager) is no longer involved.
vSphere vMotion and Storage vMotion are not supported when using Embedded vCLS. These functions are no longer needed due to the changes to a vSphere Pod based VM and lack of a storage footprint. For example, when placing a host currently running an Embedded vCLS VM into maintenance mode, the host will automatically spin-down the Embedded vCLS VM and another host in the cluster will automatically spin-up a new Embedded vCLS VM. Embedded vCLS VMs do not consume any datastore capacity, therefore, Storage vMotion of the VMs is never required.
The number of vCLS VMs per cluster has also been reduced from up to three VMs to two VMs when using Embedded vCLS. A single node cluster will use a single Embedded vCLS VM and clusters of two or more hosts will use two Embedded vCLS VMs.
You can easily identify the Cluster Service type from the summary tab of the vSphere cluster.
A cluster using the new Embedded vCLS reports as such. A vSphere 8 U2 cluster, for example, reports a Cluster Service type “vCLS”
What remains the same?
- Embedded vCLS VMs are system managed and typical VM operations are disallowed. Only power-off and hard stop operations are permitted for troubleshooting purposes.
- Embedded vCLS is required for vSphere DRS and vSphere HA to operate optimally.
- Anti-affinity with vCLS VMs, if needed, is achieved using a Compute Policy.
- Retreat Mode is still available for troubleshooting purposes.
What happens with a mixed version cluster?
Embedded vCLS is automatically activated on vSphere clusters that contain at least one ESXi 8 Update 3 host and is managed by vCenter 8 Update 3. During updates from older versions, the cluster automatically transitions to Embedded vCLS once at least one host is updated to ESXi 8 Update 3.
What happens if you update just one host in a cluster to ESXi 8 Update 3 and then subsequently remove that host from the cluster or place that host into maintenance mode?
In that scenario, where the only ESXi 8 Update 3 host is no longer available, vCLS will automatically revert to the original format where vCLS VMs are deployed by vCenter from OVF (i.e. not Embedded vCLS).
If the ESXi 8 Update 3 host is returned to the cluster or exits maintenance mode, then vCLS will again transition to using the Embedded vCLS format.
Note: Embedded vCLS will only revert to vCLS if there are no ESXi 8 Update 3 or later hosts available to run virtual machines in the cluster
