vSphere Replication appliance shows high CPU utilization when many ESXi hosts are in Maintenance Mode

Products

• VMware Live Recovery

Issue/Introduction

In vCenter vSphere client, the vSphere Replication (VR) appliance shows high CPU utilization, often approaching 100%, even though no large number of active replication tasks are running.

The issue may persist for an extended period and can impact the overall performance of the vSphere Replication appliance.

Environment

 

• vCenter managing the environment

• vSphere Replication appliance 9.0.x deployed

• Many ESXi hosts in Maintenance Mode

• The VR appliance CPU usage is observed from vCenter vSphere client

 

Cause

vSphere Replication continuously scans all ESXi hosts in the vCenter inventory to maintain replication state and host awareness. This behavior includes hosts that are in Maintenance Mode.

When ESXi hosts enter Maintenance Mode, certain services and ports required by vSphere Replication (for example, hbrsrvuw) are no longer available. As a result, the vSphere Replication appliance repeatedly attempts to communicate with these hosts.

These repeated scan and retry operations can generate a large number of internal work items and events, which in turn increases the workload on vSphere Replication services and causes high CPU utilization on the VR appliance.

Resolution

If the ESXi hosts can be exited from Maintenance Mode, this is the preferred resolution.

If the hosts cannot be taken out of Maintenance Mode in the short term, vSphere Replication scanning can be prevented by applying a specific tag in vCenter.

Workaround: Follow this KB https://knowledge.broadcom.com/external/article/405139/vcenter-task-flooded-with-cause-a-genera.html to exclude hosts from vSphere Replication scanning by assigning the tag com.vmware.vr.disallowedHost to the ESXi hosts that are currently in maintenance mode

After the tag is applied, vSphere Replication will no longer scan these hosts, which significantly reduces unnecessary work items and helps lower the CPU utilization of the vSphere Replication appliance.

Additional Information

 

• This behavior is observed when vSphere Replication repeatedly attempts to communicate with ESXi hosts that are not able to respond due to Maintenance Mode.

• This workaround is intended for temporary situations where hosts must remain in Maintenance Mode.

• Remember to remove the tag after Maintenance Mode is cleared to ensure normal vSphere Replication behavior.