nsx-opsagent-appliance service fails to start after restarting nsx-edge-api-server service

Products

• VMware NSX

Issue/Introduction

In VMware NSX, when the nsx-edge-api-server service is stopped, the nsx-opsagent-appliance service is simultaneously stopped . However, when the nsx-edge-api-server service is subsequently restarted, the nsx-opsagent-appliance service does not automatically restart and remains offline.

Following logs will indicate nsx-edge-api-server and nsx-opsagent going down

syslog.log

2026-04-14T05:23:32.014Z ######### nsx-edge-api-server 263885 - -   * Stopping nsx-edge-api-server
2026-04-14T05:23:32.867Z #########  nsx-edge-api-server 263885 - -     ...done.
2026-04-14T05:23:32.038Z#########  systemd 1 - -  nsx-edge-api-server.service: Deactivated successfully.

2026-04-14T05:23:26.680Z nsx-opsagent-appliance 263804 Shutting down NSX-OpsAgent watchdog
2026-04-14T05:23:27.664Z nsx-opsagent-appliance 263804 Shutting down NSX-OpsAgent service
2026-04-14T05:23:32.161Z systemd 1 nsx-opsagent-appliance.service: Deactivated successfully.
2026-04-14T05:23:32.540Z systemd 1 Stopped NSX OpsAgent.

During a configuration push operation (potentially an NSX Manager push for password policy changes), only the NAPI Service was restarted, but OpsAgent was not, causing OpsAgent to remain in a stopped state for an extended period. 

api_roothelper.log

2026-04-14 05:53:05,758  napi.root.node.services.telegraf_conf  WARNING  Signal to Telegraf failed: ... returned non-zero exit status 1.
2026-04-14 05:53:06,372  napi.root.node.services.utils  INFO  Service Action: stop ntp
2026-04-14 05:53:15,463  napi.root.node.services.utils  INFO  Service Action: start ntp

Environment

VMware Cloud Foundation (VCF) / VMware NSX

Cause

The nsx-edge-api-server.service systemd configuration utilizes a RequiredBy dependency on nsx-opsagent-appliance.service. While this dependency ensures the NSX API (NAPI) is running before the opsagent starts during bootup, the strict RequiredBy binding forces the opsagent to halt when NAPI is stopped without triggering a corresponding restart.

Resolution

This issue is resolved in VMware Cloud Foundation (VCF) 9.1.

To work around this issue, manually start the affected services on the NSX Edge node.

• Log in to the affected NSX Edge node via SSH as the root:

manually restart the nsx-opsagent-appliance service by running the following command: /etc/init.d/nsx-opsagent-appliance restart

• If the issue still persists, restart nsx-proxy service by running the command : /etc/init.d/nsx-proxy restart