mirror of
https://github.com/masterking32/MasterHttpRelayVPN.git
synced 2026-05-17 21:24:37 +03:00
Abolfazl
157 lines
4.9 KiB
Markdown
157 lines
4.9 KiB
Markdown
# Troubleshooting
|
|
|
|
Start with the symptom you see. Most problems are configuration, certificate trust, or an outdated Apps Script deployment.
|
|
|
|
## Certificate Errors
|
|
|
|
Symptoms:
|
|
|
|
- Browser says the connection is not private.
|
|
- Telegram or another app works, but normal HTTPS sites do not.
|
|
- Chrome or Edge still fails after you installed the certificate.
|
|
|
|
Fix:
|
|
|
|
1. Make sure the proxy has run at least once so `ca/ca.crt` exists.
|
|
2. Install `ca/ca.crt` as a trusted root certificate.
|
|
3. Fully close and reopen the browser. On Windows, check Task Manager for background Chrome or Edge processes.
|
|
4. For Firefox, import the CA separately from **Settings** -> **Privacy & Security** -> **Certificates** -> **View Certificates** -> **Authorities**.
|
|
|
|
You can also run:
|
|
|
|
```bash
|
|
python main.py --install-cert
|
|
```
|
|
|
|
## `unauthorized`
|
|
|
|
The shared secret does not match.
|
|
|
|
Fix:
|
|
|
|
1. Open [apps_script/Code.gs](../apps_script/Code.gs).
|
|
2. Find `const AUTH_KEY = "...";`.
|
|
3. Make sure it exactly matches `auth_key` in `config.json`.
|
|
4. Deploy a new Apps Script deployment after changing [apps_script/Code.gs](../apps_script/Code.gs).
|
|
|
|
## `Config not found`
|
|
|
|
Run the wizard:
|
|
|
|
```bash
|
|
python setup.py
|
|
```
|
|
|
|
Or copy [config.example.json](../config.example.json) to `config.json` and fill `script_id` plus `auth_key`.
|
|
|
|
## `502 Bad JSON`
|
|
|
|
Google returned HTML or another unexpected response instead of relay JSON.
|
|
|
|
Common causes:
|
|
|
|
- Wrong Deployment ID.
|
|
- Apps Script daily quota is exhausted.
|
|
- You edited [apps_script/Code.gs](../apps_script/Code.gs) but did not create a new deployment.
|
|
- The deployment access setting is not **Anyone**.
|
|
|
|
Fix:
|
|
|
|
1. Create a new Apps Script deployment.
|
|
2. Copy the new Deployment ID into `config.json`.
|
|
3. Confirm the deployment is a Web App with **Execute as: Me** and **Who has access: Anyone**.
|
|
4. If quota is exhausted, wait for the quota reset or add more deployments with `script_ids`.
|
|
|
|
## Page Looks Like Random Characters
|
|
|
|
Symptoms:
|
|
|
|
- A website opens as unreadable text like `�` and random symbols.
|
|
- The issue appears only for some users or only on some websites.
|
|
- HTML, JavaScript, or JSON is shown as binary-looking output instead of a normal page.
|
|
|
|
Most likely cause:
|
|
|
|
The target website sent a compressed response, but the browser received it without a usable `Content-Encoding` header. This usually happens when an old Apps Script deployment or exit node still forwards `Accept-Encoding` to the target website.
|
|
|
|
Fix:
|
|
|
|
1. Update this project and install dependencies again with `pip install -r requirements.txt`.
|
|
2. Redeploy [apps_script/Code.gs](../apps_script/Code.gs) as a new Apps Script deployment.
|
|
3. Copy the new Deployment ID into `config.json` if it changed.
|
|
4. If you use a Deno exit node, redeploy [apps_script/deno_deploy.ts](../apps_script/deno_deploy.ts).
|
|
5. Restart the proxy and fully reopen the browser.
|
|
|
|
## Connection Timeout
|
|
|
|
The current `google_ip` may be blocked or slow on your network.
|
|
|
|
Run:
|
|
|
|
```bash
|
|
python main.py --scan
|
|
```
|
|
|
|
Then set the recommended IP in `config.json`:
|
|
|
|
```json
|
|
"google_ip": "RECOMMENDED_IP"
|
|
```
|
|
|
|
Restart the proxy.
|
|
|
|
## Slow Browsing
|
|
|
|
Try these in order:
|
|
|
|
1. Install all dependencies from [requirements.txt](../requirements.txt), especially `h2`.
|
|
2. Run `python main.py --scan` and update `google_ip`.
|
|
3. Deploy multiple Apps Script projects and use `script_ids`.
|
|
4. Keep `log_level` at `INFO` unless debugging.
|
|
5. Use an exit node only when needed, because it adds another hop.
|
|
|
|
## Browser Proxy Is Set But Sites Do Not Load
|
|
|
|
Check:
|
|
|
|
1. The terminal says HTTP proxy is listening on `127.0.0.1:8085`.
|
|
2. Browser proxy type is **HTTP**, not HTTPS.
|
|
3. HTTPS traffic is also configured to use the same HTTP proxy.
|
|
4. The CA is installed and the browser was fully restarted.
|
|
|
|
## SOCKS5 Works Differently Than HTTP Proxy
|
|
|
|
Some SOCKS5 clients resolve domains locally and only send raw IPs to the proxy. That can break routes that depend on hostnames.
|
|
|
|
Use HTTP proxy `127.0.0.1:8085` for browsers and apps that support it. This is especially important for Telegram-style cases where hostname handling matters.
|
|
|
|
## YouTube Opens But Video Playback Fails
|
|
|
|
Try:
|
|
|
|
```json
|
|
"youtube_via_relay": true
|
|
```
|
|
|
|
Then restart the proxy. This routes YouTube through the Apps Script relay instead of the direct Google SNI-rewrite path. It may use more Apps Script executions.
|
|
|
|
If `exit_node.mode` is `full`, YouTube may also be routed through the relay so the exit node can handle all traffic.
|
|
|
|
## Docker Certificate Confusion
|
|
|
|
The container generates `ca/ca.crt` into the host `./ca` volume, but certificate installation inside the container does not trust it on your host OS.
|
|
|
|
Install `ca/ca.crt` manually on the host browser or OS. See [Docker Guide](DOCKER.md).
|
|
|
|
## Reset The Local Certificate
|
|
|
|
To remove the certificate from trust stores:
|
|
|
|
```bash
|
|
python main.py --uninstall-cert
|
|
```
|
|
|
|
To generate a fresh CA, stop the proxy and delete the `ca/` folder. The next run will create a new one.
|
|
|
|
Never share the `ca/` folder.
|