8.1 Virtual Machine Placement
When configured, vSphere HA and DRS can automatically perform virtual machine placement for initial power on or when the cluster is under resource contention. When running DAGs across virtual machines in a vSphere HA or DRS cluster, to avoid a single point of failure, WSFC-configured virtual machines must be kept apart, on different physical hosts. The following practices can help achieve this requirement:
• Create a DRS anti-affinity rule to keep virtual machines on different hosts.
• Enable strict enforcement of affinity rules.
• Set the DRS automation level for WSFC virtual machines to partially automatic.
In addition, when the ForceAffinePoweron option for vSphere DRS is set to 1, strict enforcement of anti-affinity rules is enforced. When possible, VMware also recommends using multiple server racks for the ESXi hosts, and distributing the vSphere clusters across the server cabinets to minimize the impact of a single component failure.
Refer to
Setup for Failover Clustering and Microsoft Cluster Service (
https://pubs.vmware.com/vsphere-60/topic/com.vmware.ICbase/PDF/vsphere-esxi-vcenter-server-60-setup-mscs.pdf) for detailed setup procedures. To set up anti-affinity rules to keep virtual machines on separate hosts, follow the instructions at
http://kb.vmware.com/kb/2003128.