Bug

openclaw gateway restart always times out after 60 seconds on any Linux system without lsof installed, even though the gateway is running and healthy.

Severity: HIGH — affects every restart on Arch Linux, Alpine, minimal Debian/Ubuntu, containers, and any distro that doesn't ship lsof by default.

Reproduction

1. Uninstall lsof (or use a system that doesn't have it)
2. Run openclaw gateway restart
3. Observe: timeout after 60s with error:

Timed out after 60s waiting for gateway port 18789 to become healthy.
Port 18789 is already in use.
- Port is in use but process details are unavailable (install lsof or run as an admin user).
Port diagnostics errors: Error: spawn lsof ENOENT
Gateway restart timed out after 60s waiting for health checks.

4. The gateway is actually running fine — curl http://127.0.0.1:18789/health returns 200.

Root Cause

In src/cli/daemon-cli/restart-health.ts, the ownsPort logic:

const ownsPort =
  runtimePid != null
    ? portUsage.listeners.some((listener) => listenerOwnedByRuntimePid({ listener, runtimePid }))
    : gatewayListeners.length > 0 ||
      (portUsage.status === "busy" && portUsage.listeners.length === 0);

When lsof is missing, inspectPortUsage() returns { status: "busy", listeners: [] }. The runtimePid != null branch calls .some() on an empty array → always false → ownsPort = false → healthy = false → loops for 120 attempts × 500ms = 60s timeout.

The runtimePid == null branch already has the correct fallback (portUsage.status === "busy" && portUsage.listeners.length === 0), but the runtimePid != null branch is missing it.

Fix

PR: #32613