ESXi Host Fails to Exit Maintenance Mode Automatically After NSX Upgrade Due to "HostInstallStatusCheck" Failure

Products

• VCF Operations
• VMware NSX

Issue/Introduction

• Following an NSX host upgrade, one or more ESXi hosts fail to exit Maintenance Mode automatically.

• The issue may affect a single host or multiple hosts within the cluster.

• Manual Maintenance mode exit of the host works via vCenter.

• In the vSphere Update Manager (VUM) logs, the HostInstallStatusCheck times out and returns a NOT_OK status:

info vmware-vum-server[18884] ... [com.vmware.nsxt.HostInstallStatusCheck] returned status: WAIT
info vmware-vum-server[21370] ... [com.vmware.nsxt.HostInstallStatusCheck] returned status: NOT_OK
info vmware-vum-server[16042] ... CheckHostHealth - check - issues size = 1 - status = 3 - check = com.vmware.nsxt.HostInstallStatusCheck name = Host installation status check - description = Results in failure if the deployment status of host is not successful or if NSX services on the host are not at target version

• In the NSX Manager (`nsxapi.log`), the management connection for the affected transport node shows as `UNKNOWN` with missing heartbeats:
    
    WARN ... getClientHeartbeatStatus: client <UUID>, no record for heartbeat found.
  INFO ... Management connection status for node <UUID> is UNKNOWN
  INFO ... Not invoking issue finder ... since MPA is not up

• Checking the affected ESXi host locally reveals that critical NSX management services are offline or failing:

  • nsx-cfgagent is not running.

  • nsx_exporter is not running.

  • nsx-proxy is failing with AssertionError: client failed to connect

 

Environment

VMware NSX

Cause

The failure to exit maintenance mode occurs because the HostInstallStatusCheck condition was not met. This check fails due to a lack of Management Plane Agent (MPA) connectivity. At the infrastructure layer, the core NSX services—specifically cfg-agent, nsx-proxy, and nsx-exporter—are down and unresponsive on the host.

Because these local management services are offline, the host cannot communicate its updated health, heartbeat, or version status back to the NSX Manager. Consequently, the vSphere automated upgrade workflow cannot verify that the NSX services have successfully reached the target version, and it safely halts the workflow by leaving the host in Maintenance Mode.

Resolution

Immediate Remediation (Post-Upgrade)
===============================
If the host is already upgraded but stuck in Exit Maintenance Mode:

1.Log in to the affected ESXi host via SSH.

2.Verify the status of the NSX services.

3.Restart the core NSX management agents on the host to restore MPA connectivity to the NSX Manager

/etc/init.d/nsx-proxy restart
/etc/init.d/nsx-cfgagent restart
/etc/init.d/nsx-opsagent restart

4.Once the NSX Manager UI reports the Transport Node status as **Up/Healthy**, manually exit the host from Maintenance Mode in vCenter Server.

### **Preventative Action (Pre-Upgrade)**

To prevent this issue during future cluster upgrades, ensure all host transport nodes are healthy and communicating with the NSX Manager before initiating the upgrade workflow.

1. Check the NSX UI under **System > Fabric > Nodes > Host Transport Nodes**.
2. Verify that the **Manager Connectivity** and **Controller Connectivity** are both listed as **Up**.
3. Resolve any disconnected nodes prior to starting the VUM/NSX lifecycle upgrade process.