A CloudAdmin user can establish policies and profiles in the SDDC that govern the placement of workload VMs.
Creating and Managing Compute Policies
Compute policies provide a way to specify how the vSphere Distributed Resource Scheduler (DRS) should place VMs on hosts in a resource pool. Use the vSphere client Compute Policies editor to create and delete compute policies.
You can create or delete, but not modify, a compute policy. If you delete a category tag used in the definition of the policy, the policy is also deleted. The system does not check for policy conflicts. If, for example, multiple VMs subject to the same VM-Host affinity policy are also subject to a VM-VM anti-affinity policy, DRS will be unable to place the VMs in a way that complies with both policies.
Affinity policies in your VMware Cloud on AWS SDDC are not the same as the vSphere DRS affinity rules you can create on premises. The policies can be used in similar ways, but have significant operational differences. A compute policy applies to all hosts in an SDDC, and cannot typically be enforced in the same way that a DRS "must" policy is enforced. The policy create/delete pages have more information about operational details for each policy type.
Impact of Compute Policies During Host Replacement and Maintenance
If you implement a compute policy based on host tags, it's important to review the potential impact when a tagged host is temporarily taken out of service for maintenance or fails and is replaced. Either event can reduce available resources while the replacement or maintenance activity is completed. We recommend allocating a sufficient number of tagged hosts to support the workloads being placed by compute policy when a single tagged host is temporarily unavailable.
Impact of Compute Policies on Cluster Conversion
When planning for a cluster conversion, special consideration is required when host tags are being used with compute policy. Cluster conversion does not replace host tags on new (converted) hosts in the cluster, which can cause resource starvation where tagged hosts are required for workload placement. We recommend temporarily disabling the compute policy, or temporarily relocating affected workloads to another cluster prior to the cluster conversion. See Converting Host Types in Clusters.
Monitoring Compliance
Open the VM Summary page in the vSphere client to view the compute policies that apply to a VM and its compliance status with each policy.
