Fix site start/stop status by removing stale socket file

[bcotrim]

Related issues

• Fixes STU-1278 STU-1255

Proposed Changes

• Remove stale Unix socket file (~/.studio/pm2/events.sock) before binding in the _events command
• Only applies to Unix-based systems (macOS/Linux) since Windows named pipes are automatically cleaned up by the OS

Root cause: When Studio crashes or is force-killed, the socket file is left behind. On restart, the _events command cannot bind to this socket (EADDRINUSE), preventing site-event updates from flowing to the UI.

Testing Instructions

1. Start Studio in dev mode (npm start)
2. Verify normal start/stop works - click Start/Stop on a site, status should update
3. Force-kill Studio without proper cleanup:

   pkill -9 -f "Electron.app"

4. Verify the stale socket file exists:

   ls -la ~/.studio/pm2/events.sock

5. Start Studio again (npm start)
6. Click Start/Stop on a site - status should update correctly (the fix removes the stale socket automatically)

Pre-merge Checklist

• Have you checked for TypeScript, React or other console errors?

[wpmobilebot]

📊 Performance Test Results

Comparing 611152d vs trunk

site-editor

Metric	trunk	611152d	Diff	Change
load	2921.00 ms	2914.00 ms	-7.00 ms	🟢 -0.2%

site-startup

Metric	trunk	611152d	Diff	Change
siteCreation	7091.00 ms	7112.00 ms	+21.00 ms	🔴 0.3%
siteStartup	3929.00 ms	3921.00 ms	-8.00 ms	🟢 -0.2%

---

Results are median values from multiple test runs.

Legend: 🟢 Improvement (faster) | 🔴 Regression (slower) | ⚪ No change

[sejas]

Nice! it solved my issue! 👌
Now I can see the sites starting and stopping as expected.

[fredrikekelund]

pm2-axon is supposed to automatically handle EADDRINUSE errors in Socket::bind. See source.

The gist of the logic is that:

• If net.createServer emits an EADDRINUSE error.
  • Create a net.Socket and try to connect to the socket.
  • If this fails, it means that no server is listening on the socket, so pm2-axon removes the socket file and connects again.

If this fixes the underlying problem, that's great 👍 Still, I'm curious how the pm2-axon logic fails and why…