fix: an answer whose client went away ends quietly, and a shell whose client left before it opened is stopped.

ADR.md

@@ -1342,3 +1342,28 @@ method.
 Consequences: A change to one family of paths touches one module and the routing in `handler`.
 State shared between requests, such as the open shells, lives in the module that owns it; other
 modules change it and never rebind it.
+
+## ADR-48: Exceptions are handled; no traceback reaches the logs
+
+Status: Accepted
+
+Context: The services log to the journal, which administrators read through the Logs page and
+`journalctl`. Python prints a traceback for an exception nothing handles: `socketserver` for a
+failed request, `threading` for a failed thread and the interpreter for the process. A client
+that closes its connection before the answer is written, as a page does when it is left or
+reloaded, raises `BrokenPipeError` or `ConnectionResetError` in the handler. A traceback is many
+lines of internals for one event and hides the lines around it.
+
+Decision: The code handles the exceptions it can expect where they occur, and answers or logs
+them with their reason in one line. An exception left unhandled is a defect to be fixed. As the
+last line of defence, every server replaces the default reports:
+
+- `vpnrouter_serve.serve`, which the web service and the metrics service run, sets the servers'
+  `handle_error`, `threading.excepthook` and `sys.excepthook`.
+- The bootstrap server sets its server's `handle_error` and `sys.excepthook`.
+
+A `ConnectionError` of a client that went away is not logged; any other exception is logged in
+one line with its type and message.
+
+Consequences: The logs hold one line per failure, without its stack. A failure that reaches the
+last line of defence is found by its message and fixed where it occurs, with its own handling.

README.md

@@ -1160,6 +1160,7 @@ vre /etc/vpn-router/vpn-router.json ~/notes.txt
 | `docker` | `docker.exposed_ports` and the rules of the `VPN-ROUTER-DOCKER` chain |
 | `postgresql` | `clusters`: each PostgreSQL cluster of `pg_lsclusters` with its version, name, port, status and data directory, and for one that is online its `settings`: `server_version`, `listen_addresses`, `port`, `max_connections`, `shared_buffers`, `effective_cache_size`, `work_mem`, `password_encryption`, `ssl` and `timezone`; `null` when `pg_lsclusters` fails |
 | `caddy` | The installed Caddy: `version`, `source`, `upstream` once Caddy's repository is configured and `ubuntu` otherwise, and `service`, the state of `caddy.service`; `null` when Caddy is not installed |
+| `oidc` | The OpenID Provider: `enabled`, and `issuer`, its issuer URL, `null` while it is disabled |
 
 ### The /postgresql page
 

debian-package/debian/changelog

@@ -1,3 +1,34 @@
+vpn-router (0.37.3-1) unstable; urgency=medium
+
+  fix: an answer whose client went away ends quietly, and a shell whose client left before it opened is stopped.
+
+  * The web service, the metrics service and the bootstrap server stop
+    writing an answer when its client closes the connection, and a site
+    file whose download is left no longer gets a second answer.
+  * A shell whose client closed the connection before it opened is
+    stopped and its end is logged, instead of running on.
+
+ -- Slawomir Koszewski <slawek@koszewscy.waw.pl>  Fri, 30 Oct 2026 02:03:26 +0100
+
+vpn-router (0.37.2-1) unstable; urgency=medium
+
+  fix: the web service, the metrics service and the bootstrap server log a failure in one line instead of a traceback, and nothing when a client goes away.
+
+  * A request whose client closed the connection is no longer logged.
+  * Any other failure of a request, a thread or the service is logged in
+    one line with the exception's type and message.
+
+ -- Slawomir Koszewski <slawek@koszewscy.waw.pl>  Fri, 30 Oct 2026 01:59:19 +0100
+
+vpn-router (0.37.1-1) unstable; urgency=medium
+
+  fix: /config answers the OpenID Provider's state and issuer.
+
+  * /config has oidc: enabled, and issuer, the issuer URL, null while the
+    provider is disabled.
+
+ -- Slawomir Koszewski <slawek@koszewscy.waw.pl>  Fri, 30 Oct 2026 01:53:55 +0100
+
 vpn-router (0.37.0-1) unstable; urgency=medium
 
   feature: the OpenID Provider has the client credentials and device authorization grants, public clients, and service accounts as resources with roles and scopes in JWT access tokens.

debian-package/src/usr/lib/vpn-router/bootstrap

@@ -8,6 +8,7 @@ the logs.
 """
 
 import subprocess
+import sys
 from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
 
 PORT = 17081
@@ -42,6 +43,7 @@ def report():
 class Handler(BaseHTTPRequestHandler):
     def do_GET(self):
         body = report().encode()
+        try:
             self.send_response(200)
             self.send_header('Content-Type', 'text/plain; charset=utf-8')
             self.send_header('Content-Length', str(len(body)))
@@ -49,7 +51,20 @@ class Handler(BaseHTTPRequestHandler):
             self.send_header('Refresh', '10')
             self.end_headers()
             self.wfile.write(body)
+        except ConnectionError:
+            # The client went away before the answer was written.
+            self.close_connection = True
+
+
+def failed(request, client_address):
+    """Log a failed request in one line, and nothing when the client went away."""
+    failure = sys.exc_info()[1]
+    if not isinstance(failure, ConnectionError):
+        print(f'a request from {client_address[0]} failed: {type(failure).__name__}: {failure}', file=sys.stderr, flush=True)
 
 
 if __name__ == '__main__':
-    ThreadingHTTPServer(('', PORT), Handler).serve_forever()
+    sys.excepthook = lambda kind, value, _: print(f'failed: {kind.__name__}: {value}', file=sys.stderr, flush=True)
+    server = ThreadingHTTPServer(('', PORT), Handler)
+    server.handle_error = failed
+    server.serve_forever()

debian-package/src/usr/lib/vpn-router/metrics

@@ -405,6 +405,7 @@ class Handler(BaseHTTPRequestHandler):
                     registry = CollectorRegistry()
                     registry.register(Sample(current))
                     body, kind, code = generate_latest(registry), CONTENT_TYPE_LATEST, 200
+        try:
             self.send_response(code)
             self.send_header('Content-Type', kind)
             self.send_header('Content-Length', str(len(body)))
@@ -414,6 +415,9 @@ class Handler(BaseHTTPRequestHandler):
             self.end_headers()
             if send_body:
                 self.wfile.write(body)
+        except ConnectionError:
+            # The client went away before the answer was written.
+            self.close_connection = True
 
     def do_GET(self):
         self.answer(True)

debian-package/src/usr/lib/vpn-router/vpnrouter_serve.py

@@ -4,6 +4,7 @@ Only the Python standard library is used, so the bootstrap server, which must ru
 is broken, can use it.
 """
 
+import sys
 import threading
 import time
 from http.server import ThreadingHTTPServer
@@ -14,9 +15,21 @@ def serve(wanted, handler, log, interval):
 
     wanted answers ([(address, port), ...], '') or ([], the reason not to listen). A server whose
     place is no longer wanted is stopped, and a new one listens at each new place; a place that
-    cannot be bound is tried again the next round. Every new message is logged with log. Does
+    cannot be bound is tried again the next round. Every new message is logged with log. A request
-    not return.
+    whose client went away is not logged; any other exception a request, a thread or the process
+    does not handle is logged with log in one line, without a traceback. Does not return.
     """
+    def failed(request, client_address):
+        failure = sys.exc_info()[1]
+        if not isinstance(failure, ConnectionError):
+            log(f'a request from {client_address[0]} failed: {type(failure).__name__}: {failure}')
+
+    def thread_failed(hook):
+        if not issubclass(hook.exc_type, (SystemExit, ConnectionError)):
+            log(f'{hook.thread.name if hook.thread else "a thread"} failed: {hook.exc_type.__name__}: {hook.exc_value}')
+
+    threading.excepthook = thread_failed
+    sys.excepthook = lambda kind, value, _: log(f'failed: {kind.__name__}: {value}')
     servers, reported = {}, None
     while True:
         targets, reason = wanted()
@@ -34,6 +47,7 @@ def serve(wanted, handler, log, interval):
             except OSError as exc:
                 messages.append(f'cannot listen on {target[0]}:{target[1]}: {exc}')
                 continue
+            server.handle_error = failed
             thread = threading.Thread(target=server.serve_forever, name=f'server {target[0]}', daemon=True)
             thread.start()
             servers[target] = (server, thread)

debian-package/src/usr/lib/vpn-router/vpnrouter_web/handler.py

@@ -188,6 +188,7 @@ class Handler(BaseHTTPRequestHandler):
             code = 200
         else:
             body, kind, code = b'Not found\n', 'text/plain; charset=utf-8', 404
+        try:
             self.send_response(code)
             self.send_header('Content-Type', kind)
             self.send_header('Content-Length', str(len(body)))
@@ -197,6 +198,9 @@ class Handler(BaseHTTPRequestHandler):
             self.end_headers()
             if send_body:
                 self.wfile.write(body)
+        except ConnectionError:
+            # The client went away before the answer was written.
+            self.close_connection = True
 
     def raw_body(self):
         """The request body's bytes, b'' without one or over BODY_LIMIT."""

debian-package/src/usr/lib/vpn-router/vpnrouter_web/shell.py

@@ -367,11 +367,14 @@ def shell(handler, name):
     def refuse(code, message):
         log(f'shell of {name} from {handler.client()} refused: {message}')
         body = json.dumps({'error': message}).encode() + b'\n'
+        try:
             handler.send_response(code)
             handler.send_header('Content-Type', 'application/json')
             handler.send_header('Content-Length', str(len(body)))
             handler.end_headers()
             handler.wfile.write(body)
+        except ConnectionError:
+            handler.close_connection = True
 
     key = handler.headers.get('Sec-WebSocket-Key', '').strip()
     if handler.headers.get('Upgrade', '').lower() != 'websocket' or 'upgrade' not in handler.headers.get('Connection', '').lower() \
@@ -408,12 +411,21 @@ def shell(handler, name):
     finally:
         os.close(terminal)
     accept = base64.b64encode(hashlib.sha1((key + WEBSOCKET_GUID).encode()).digest()).decode()
+    handler.close_connection = True
+    try:
         handler.send_response(101, 'Switching Protocols')
         handler.send_header('Upgrade', 'websocket')
         handler.send_header('Connection', 'Upgrade')
         handler.send_header('Sec-WebSocket-Accept', accept)
         handler.end_headers()
-    handler.close_connection = True
+    except ConnectionError:
+        # The client went away before the shell opened; the shell started for it is stopped.
+        run('systemctl', 'stop', unit)
+        os.close(master)
+        process.kill()
+        process.wait()
+        log(f'shell of {name} from {handler.client()} ended before it opened: the client closed the connection')
+        return
     client, started, idle = handler.client(), time.monotonic(), options['shell_idle_minutes'] * 60
     log(f'shell of {name} from {client} opened as {account.pw_name} in {unit}')
 

debian-package/src/usr/lib/vpn-router/vpnrouter_web/sites.py

@@ -112,6 +112,7 @@ def site_files(handler, name, relative, query, payload, send_body):
                 return 404, {'error': f'{relative} does not exist'}
             with open(target, 'rb') as handle:
                 size = os.fstat(handle.fileno()).st_size
+                try:
                     handler.send_response(200)
                     handler.send_header('Content-Type', mimetypes.guess_type(target.name)[0] or 'application/octet-stream')
                     handler.send_header('Content-Length', str(size))
@@ -121,6 +122,9 @@ def site_files(handler, name, relative, query, payload, send_body):
                     handler.end_headers()
                     if send_body:
                         shutil.copyfileobj(handle, handler.wfile, vpnrouter_sites.CHUNK)
+                except ConnectionError:
+                    # The client went away before the file was sent.
+                    handler.close_connection = True
             return None
         if handler.command == 'PUT':
             length = handler.headers.get('Content-Length')

debian-package/src/usr/lib/vpn-router/vpnrouter_web/status.py

@@ -211,9 +211,11 @@ def postgresql_databases():
 
 def details(state):
     """What /config shows: the configuration without its secrets, and what the router resolved and runs; caddy is the
-    installed Caddy, {version, source, service}, service the state of caddy.service, null when it is not installed."""
+    installed Caddy, {version, source, service}, service the state of caddy.service, null when it is not installed; oidc
+    is {enabled, issuer} of the OpenID Provider, issuer null while it is disabled."""
     configuration = state.get('configuration', {})
     caddy = package_version('caddy')
+    oidc = configuration.get('oidc', {}).get('enabled') is True
     return {
         'configuration': configuration,
         'network': dict(state.get('network', {}), routes=routes()),
@@ -230,6 +232,7 @@ def details(state):
         # Caddy is upstream's once its repository is configured (vpn-router-caddy).
         'caddy': {'version': caddy, 'source': 'upstream' if vpnrouter.CADDY_APT_SOURCE.is_file() else 'ubuntu',
                   'service': unit_state('caddy.service')} if caddy else None,
+        'oidc': {'enabled': oidc, 'issuer': vpnrouter.oidc_issuer(configuration) if oidc else None},
     }