DRS evacuates VMs after ESXi reboot due to InsufficientAgentVmsDeployment and missing Guest Introspection SVMs

Products

• VMware vSphere ESXi

Issue/Introduction

After an ESXi host reboot, workloads fail to run on specific cluster nodes.

Manual attempts to vMotion or migrate a Virtual Machine (VM) to the affected hosts succeed temporarily, but the Distributed Resource Scheduler (DRS) immediately intervenes and evacuates the VM back to other hosts in the cluster.

Initial symptoms mimic a standard networking availability or uplink failure.   vCenter Server flags the affected hosts with an InsufficientAgentVmsDeployment error.

Environment

VMware NSX 4.x

VMware vCenter Server 8.x

Cause

Following an ESXi host reboot, the ESX Agent Manager (EAM) fails to successfully provision and initialize the Guest Introspection Service Virtual Machines (SVMs).

By design, if a cluster enforces strict security or introspection policies, DRS treats the missing security agent as a critical compliance failure and automatically evacuates the node to protect workload integrity.

Resolution

Execute the following steps sequentially to resolve the EAM agency errors and restore normal DRS workload distribution:

Step 1: Verify Agent & Service VM Status

1. Log in to the vSphere Client.
   
   
2. Navigate to Administration > ESX Agent Manager (or check the NSX Manager dashboard depending on your architecture version).
   
   
3. Review the deployment status of the Guest Introspection Service VMs (SVMs).
   
   
4. Filter or look specifically for errors mapped to hosts.  (Note down the exact state (e.g., Missing, Power Off, Installation Failed).

Step 2: Log Collection & Inspection:   If the visual status does not identify the root cause (such as a missing storage path or VLAN mismatch for the SVMs), inspect the relevant logs:

1. ESX Agent Manager Logs: Review eam.log on the vCenter Server to determine why EAM failed to provision or communicate with the SVMs on the impacted hosts.
   
   
2. NSX Manager Logs: Verify communication timeouts between the NSX Manager and the host transport nodes during the boot sequence.

Step 3: Trigger EAM Remediation: EAM frequently requires a manual trigger to retry deployment following a reboot timeout.

1. Automated Resolve:
   
   
• Navigate to vSphere Client > Home > Administration > ESX Agent Manager.
  
  
• Select the failing Guest Introspection Agency.
  
  
• Click Resolve or Resolve All to force vCenter Server to redeploy/reconfigure the missing agents.
  
  
2. Manual Redeployment:
   
   
• If the Automated Resolve fails, navigate to the NSX Service Deployment tab.
  
  
• Select the affected cluster/hosts and initiate a Redeploy of the Guest Introspection service for the impacted nodes.

Step 4: Verification

1. Confirm the Guest Introspection SVMs display a Healthy/Green status on the hosts.
   
   
2. Verify the InsufficientAgentVmsDeployment alert clears automatically.
   
   
3. Manually migrate a non-critical VM to one of the affected hosts.
   
   
4. Verify that DRS permits the VM to reside on the host without triggering an immediate evacuation.