vCenter High Availability Basic Configuration stuck/fails during Witness/Passive node clone operation

Products

• VMware vCenter Server

Issue/Introduction

Symptoms:

• VCHA Basic Configuration stuck/fails at 96%/33% when vCenter Server is stored on NFSv3 Datastore.
• vCenter Server service may be restarted by vmware-vmon service during the VCHA configuration.
• In the vpxd.log file, you see entries similar to:

xxxx-xx-xxTxx:xx:xx.xxxZ error vpxd[7F725BE7C700] [Originator@6876 sub=VmProv opID=FlowBasedWizard-apply-638-ngc-35-f-01] Get exception while executing action vpx.vmprov.RemoveSnapshot: N5Vmomi5Fault17HostCommunication9ExceptionE(vmodl.fault.HostCommunication)

 

Environment

VMware vCenter Server 7.0.x
VMware vCenter Server Appliance 6.7.x
VMware vCenter Server Appliance 6.5.x

Cause

This issue occurs due to limitation from NFSv3 locking mechanism.

Resolution

While deploying VCHA, ensure target and source ESXi hosts used in clone operations is the same ESXi host.

To avoid anti-affinity rule violation error during the VCHA configuration when same ESXi host is used for active,passive and witness node deployment:

•  Set "config.vpxd.vcha.drsAntiAffinity" to False in vCenter Advance settings

Note: After the configuration nodes can be moved to different ESXI hosts and the above Advance settings can be revert back to True