vCenter vpxd-svcs service health check fails and restarts due to timeout threshold

Products

• VMware vCenter Server

Issue/Introduction

A vCenter Server alert frequently triggers, stating the vpxd-svcs service is in a red status. The /var/log/vmware/vmon/vmon.log contains the following error sequence indicating a health check retrieval failure and subsequent automated service restart: Wa(03)  <vpxd-svcs> Health of service failed. Health data: {"localizable_msgs": [{"id": "com.vmware.vmon.svc_health_fail", "default_message": "Failed to retrieve service health.", "args": []}], "_service_name": "vpxd-svcs", "_trigger_threaddump_on_failure": 0} In(05)  <vpxd-svcs> Restarting service.

Environment

vCenter Server 8.0

Cause

The default health check timeout threshold of 30 seconds is insufficient for the vpxd-svcs service to respond under certain operating conditions or loads. This delay causes the vmon service to register a health retrieval failure and forcefully terminate and restart the vpxd-svcs process.

Resolution

• Establish an SSH session to the affected vCenter Server Appliance and log in as the root user.

• Backup the configuration file: cp vi /etc/vmware/vmware-vmon/svcCfgfiles/vpxd-svcs.json /root/vpxd-svcs.json 

• Open the vpxd-svcs service configuration file using a text editor (e.g., vi): vi /etc/vmware/vmware-vmon/svcCfgfiles/vpxd-svcs.json

• Locate the timeout parameter string and modify the value from -t 30 to -t 90.

• Save the changes and close the configuration file.

• Restart the vmon service to apply the updated configuration threshold.