Virtual Machines Do Not Automatically Failback to Original Host After vSphere HA Recovery

Products

• VMware vCenter Server

Issue/Introduction

After a vSphere High Availability (HA) event (such as a host network isolation or failure test), virtual machines restart on surviving hosts successfully. Upon restoring the affected host to a healthy state, the virtual machines do not automatically migrate back to their original host.

Environment

vSphere 8.x

Cause

vSphere HA is a reactive recovery mechanism designed solely to restart virtual machines on healthy hosts during a failure. HA does not possess built-in failback logic to track and return workloads to their original host. Proactive workload balancing and failback are handled exclusively by the vSphere Distributed Resource Scheduler (DRS).

Resolution

Process to verify the DRS automation level of the cluster.  Once the ESXi host has been restored, DRS can proceed to automate workload balancing determined from its calculated logic.

1. Log in to the vSphere Client.

2. Select the affected vSphere cluster in the inventory.

3. Navigate to Configure > Services > vSphere DRS and click Edit.

4. Verify that vSphere DRS is turned ON.

5. Expand DRS Automation and set the Automation Level to Fully Automated.

Note: If the Automation Level is configured as Manual or Partially Automated, administrators must manually review and apply the generated DRS Recommendations within the vSphere Client to migrate virtual machines back to the recovered host.

Additional Information

vSphere Resource Management