Hi OpenClaw team,

I am seeing a reproducible mismatch between openclaw status and the actual systemd user-unit state on a Linux VPS after updating to OpenClaw 2026.5.7.

openclaw status reports the gateway service as stopped/failed, but the gateway is reachable, Discord is OK, and systemctl --user show reports the unit as active/running with a successful result.

This looks related to older systemd status reporting issues such as #40275 and #27267, and possibly the fix in #39325, but it still reproduces for me on 2026.5.7.

Environment

• OpenClaw: 2026.5.7 (eeef486)
• Install type: global npm/pnpm install, managed via OpenClaw systemd user service
• OS: Ubuntu 24.04.4 LTS
• Kernel: 6.8.0-106-generic
• Node.js: 22.22.2
• Gateway mode: local loopback
• Gateway URL: ws://127.0.0.1:18789
• systemd unit type: user service
• User linger: enabled (Linger=yes)

What openclaw status reports

Gateway              local · ws://127.0.0.1:18789 (local loopback) · reachable ~65ms · auth token
Gateway service      systemd user installed · enabled · stopped (state failed)
Plugin compatibility none
Discord              ON · OK · token config (...) · accounts 1/1

So the status output says the gateway is reachable and Discord is OK, but also says the systemd service is stopped/failed.

What systemd reports

$ XDG_RUNTIME_DIR=/run/user/0 systemctl --user show openclaw-gateway \
  -p ActiveState -p SubState -p Result -p ExecMainStatus -p UnitFileState

Result=success
ExecMainStatus=0
ActiveState=active
SubState=running
UnitFileState=enabled

systemctl --user status openclaw-gateway --no-pager -l also shows the service active and running:

● openclaw-gateway.service - OpenClaw Gateway
     Loaded: loaded (...; enabled; preset: enabled)
     Active: active (running)
   Main PID: <node pid> (node)
      CGroup: .../openclaw-gateway.service
              ├─<pid> /usr/bin/node /usr/lib/node_modules/openclaw/dist/index.js gateway --port 18789
              └─<pid> node .../flybridge/dist/server.js

No failed units are reported:

$ systemctl --failed --no-pager
0 loaded units listed.

$ XDG_RUNTIME_DIR=/run/user/0 systemctl --user --failed --no-pager
0 loaded units listed.

Gateway ports are listening:

127.0.0.1:18789  node .../openclaw/dist/index.js gateway --port 18789
*:9998           node .../flybridge/dist/server.js

Other health checks

$ openclaw config validate
Config valid: ~/.openclaw/openclaw.json

$ openclaw plugins doctor
No plugin issues detected.

openclaw status reports Discord as OK:

Discord  ON  OK  token config (...) · accounts 1/1

Expected behavior

If the systemd user unit is active/running and the gateway is reachable, openclaw status should report the gateway service as active/running instead of stopped/failed.

Actual behavior

openclaw status reports:

Gateway service      systemd user installed · enabled · stopped (state failed)

while systemd reports:

ActiveState=active
SubState=running
Result=success
ExecMainStatus=0

Notes

This happened after openclaw update --yes from 2026.5.6 to 2026.5.7. The update completed the core update successfully, but the automatic gateway restart timed out after 60 seconds waiting for health checks. Shortly afterward the gateway was actually running and reachable on 2026.5.7.

The service also had a recent unit refresh from openclaw doctor, followed by:

XDG_RUNTIME_DIR=/run/user/0 systemctl --user daemon-reload

Resetting the failed state did not change the openclaw status display:

XDG_RUNTIME_DIR=/run/user/0 systemctl --user reset-failed openclaw-gateway

Systemd still reports the unit as active/running and with no failed units, while openclaw status continues to show stopped (state failed).

Thank you for looking into this.