# \[Workaround\] White screen / enableClientEncryption error with reverse proxy + HTTP Basic Auth

**URL:** <https://community.silverbullet.md/t/workaround-white-screen-enableclientencryption-error-with-reverse-proxy-http-basic-auth/4002>\
**Category:** Troubleshooting\
**Created:** [March 30, 2026, 4:27pm UTC](https://community.silverbullet.md/t/workaround-white-screen-enableclientencryption-error-with-reverse-proxy-http-basic-auth/4002 "2026-03-30T16:27:11Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![ZTKenan](https://community.silverbullet.md/user_avatar/community.silverbullet.md/ztkenan/32/3365_2.png) [@ZTKenan](https://community.silverbullet.md/u/ZTKenan)\
**Post date:** [March 30, 2026, 4:27pm UTC](https://community.silverbullet.md/t/workaround-white-screen-enableclientencryption-error-with-reverse-proxy-http-basic-auth/4002/1 "2026-03-30T16:27:11Z")

</div>

## White screen / "Cannot set properties of undefined (setting 'enableClientEncryption')" - Root cause and solution when using reverse proxy + HTTP Basic Auth

_This article is summarized by LLM based on my conversations with LLM during the debugging process._

### Symptom

- SilverBullet shows a white screen with only HTML skeleton loaded.
- Browser console error: `TypeError: Cannot set properties of undefined (setting 'enableClientEncryption')`
- `service_worker.js` fails to register with HTTP 401 or MIME type `text/html`.
- The issue is intermittent – sometimes the page loads after multiple refreshes, especially on faster devices/desktops.

**My setup:**

- SilverBullet v2.5.2 on Windows
- Nginx reverse proxy with HTTP Basic Auth
- CDN (EdgeOne) providing HTTPS edge, HTTP origin pull
- NSSM for Windows service
- Access: public internet → CDN → Nginx → SilverBullet

**Key observation:**

- Local access ([http://localhost:3000](http://localhost:3000)) always works fine.
- Public access (https://_my\_domain_) results in white screen / Service Worker registration failure.
- The issue is intermittent - more frequent on mobile devices, less frequent on desktop. Multiple refreshes sometimes succeed.

### Root Cause

**A race condition between Service Worker installation and the first data requests.**

When SilverBullet loads:

1. It tries to install the Service Worker.
2. **It does NOT wait** for the Service Worker to be ready before sending critical requests (`.config`, `.fs/CONFIG.md`, etc.).

If the Service Worker installs _slowly_ (e.g., on mobile devices, or first visit), these data requests bypass the Service Worker and go directly to the reverse proxy (Nginx/Caddy). If the proxy requires HTTP Basic Auth, these requests return **401** because the browser hasn't provided credentials yet. The frontend code then fails to initialize, resulting in a white screen.

The "cache problem" is real – but the deeper issue is that the frontend logic doesn't await Service Worker readiness before making API calls.

### Why it's intermittent

- **Desktop / "Desktop mode" on mobile** → faster Service Worker installation → wins the race → page loads.
- **Mobile / slow network** → slower installation → loses the race → white screen.
- **Multiple refreshes** → Service Worker may have installed in the background → subsequent loads succeed.

### Solution

The fix makes Nginx automatically attach Basic Auth credentials when proxying to SilverBullet. This ensures that even if a request bypasses the Service Worker, the backend still receives valid credentials.

First, encode your username:password in Base64:

```bash
echo -n "username:password" | base64

```

Feel free to adjust the Base64 command syntax for your platform (`base64` on Linux/macOS, `certutil` or PowerShell on Windows).

Then add to your Nginx `location /` block:

```nginx
location / {
    auth_basic "Restricted";
    auth_basic_user_file .htpasswd;
    
    # Forward credentials to backend automatically
    proxy_set_header Authorization "Basic dXNlcm5hbWU6cGFzc3dvcmQ=";
    
    proxy_pass http://127.0.0.1:3000;
    proxy_http_version 1.1;
    proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection "upgrade";
    proxy_set_header Host $host;
    # ... other proxy settings
}

```

#### 3. Clear all caches and reset client

After updating Nginx, manually clear:

- Site data (Application → Storage → Clear site data)
- Cache storage

_LLM also mentioned to do the following actions, but I'm not sure whether they would help:_

```auto
https://your-domain.com/?resetClient=1

```

Manually clear Service Workers (`chrome://serviceworker-internals/`)

### Why this works (LLM explanation)

- The Service Worker can be installed without authentication.
- The `Authorization` header ensures all proxied requests (whether via Service Worker or direct) are authenticated.
- No more race condition – the backend always sees valid credentials.

### Result

As for me, although I can still see `Error: Failed to register a ServiceWorker for scope ('https://*my_domain*/') with script ('https://*my_domain*/service_worker.js'): The script has an unsupported MIME type ('text/html').` error, I finally managed to edit my documents from public internet with my mobile phone.

---

<div class="post-metadata">

**Author:** ![ZTKenan](https://community.silverbullet.md/user_avatar/community.silverbullet.md/ztkenan/32/3365_2.png) [@ZTKenan](https://community.silverbullet.md/u/ZTKenan)\
**Post date:** [March 30, 2026, 4:29pm UTC](https://community.silverbullet.md/t/workaround-white-screen-enableclientencryption-error-with-reverse-proxy-http-basic-auth/4002/2 "2026-03-30T16:29:44Z")

</div>

Refrence: [White screen only after upgrade to latest version 2.5.1 - #5 by LumenYoung](https://community.silverbullet.md/t/white-screen-only-after-upgrade-to-latest-version-2-5-1/3945/5)

---

<div class="post-metadata">

**Author:** ![LumenYoung](https://community.silverbullet.md/user_avatar/community.silverbullet.md/lumenyoung/32/368_2.png) [@LumenYoung](https://community.silverbullet.md/u/LumenYoung)\
**Post date:** [April 28, 2026, 11:20am UTC](https://community.silverbullet.md/t/workaround-white-screen-enableclientencryption-error-with-reverse-proxy-http-basic-auth/4002/3 "2026-04-28T11:20:43Z")

</div>

I think I locate a more important fix to be extending the boot timeout.

I created a PR for this [Increase boot fetch timeout for authenticated reverse proxies by LumenYoung · Pull Request #1953 · silverbulletmd/silverbullet · GitHub](https://github.com/silverbulletmd/silverbullet/pull/1953)