Restarting Management Agents in ESXi

Products

• VMware vSphere ESXi

Issue/Introduction

Connectivity issues with vCenter or ESXi hosts often arise from unresponsive management services (hostd, vpxa). This article provides the standard procedure to restart these services to restore management plane communication.

Warning:

Do not use the services.sh restart command or Restart Management agents in ESXi Using Direct Console User Interface (DCUI) On ESXi hosts:

• If you are running vSAN

esxcli vsan cluster get

• If Link Aggregation Control Protocol (LACP) is in use

esxcli network vswitch dvs vmware lacp status get

• If NSX is configured in the environment

esxcli software vib list | grep -E 'nsx|vsipfwlib'

• If you are using shared graphics in a View environment (VGPU, vDGA, vSGA)

Only restart single services.

NOTE: Restarting ESXi management agents may:

• cause temporary disruptions
• can lead to system instability requiring a host reboot
• impact currently running tasks, including impacting performance on the guest virtual machines of that host.

Environment

VMware vSphere ESXi 7.x
VMware vSphere ESXi 8.x
VMware vSphere ESXi 9.x

VMware Cloud Foundation 4.x
VMware Cloud Foundation 5.x
VMware Cloud Foundation 9.x

Cause

Management agents (hostd and vpxa) may enter a hung state due to high resource contention, stuck tasks, or integration with third-party software.

Resolution

Restart management agents in ESXi using the Host Client:

In the ESXi Host Client, select Host > Manage > Services and select the service (vpxa or hostd) to restart

Restart management agents in ESXi Using ESXi Shell or Secure Shell (SSH):

1. Access the Host: Connect to the affected ESXi host via SSH (ensure SSH is enabled in the DCUI).
2. Restart hostd Service: Execute the following command: /etc/init.d/hostd restart
3. Restart vpxa Service: Execute the following command: /etc/init.d/vpxa restart
4. Verify Status: Allow 1–2 minutes for the agents to re-initialize and check the vSphere Client to see if the host status returns to "Connected."

Note: If services fail to restart, inspect /var/log/hostd.log for underlying errors (e.g., out of memory, disk pressure).

Additional Information

For VMware Cloud Foundation (VCF) environments, ensure you check the SDDC Manager inventory status after services are restored, as inventory sync issues may persist if the host was disconnected for an extended period.