Merge pull request #61 from vbanos/remote-server-timeout

Make remote server connection timeout configurable

tests/test_warcprox.py

@@ -249,6 +249,14 @@ class _TestHttpRequestHandler(http_server.BaseHTTPRequestHandler):
         elif self.path == '/empty-response':
             headers = b''
             payload = b''
+        elif self.path == '/slow-response':
+            time.sleep(6)
+            headers = (b'HTTP/1.1 200 OK\r\n'
+                    +  b'Content-Type: text/plain\r\n'
+                    +  b'\r\n')
+            payload = b'Test.'
+            actual_headers = (b'Content-Type: text/plain\r\n'
+                           + b'Content-Length: ' + str(len(payload)).encode('ascii') + b'\r\n')
         else:
             payload = b'404 Not Found\n'
             headers = (b'HTTP/1.1 404 Not Found\r\n'
@@ -356,7 +364,8 @@ def warcprox_(request):
             '--port=0',
             '--playback-port=0',
             '--onion-tor-socks-proxy=localhost:9050',
-            '--crawl-log-dir=crawl-logs']
+            '--crawl-log-dir=crawl-logs',
+            '--socket-timeout=4']
     if request.config.getoption('--rethinkdb-dedup-url'):
         argv.append('--rethinkdb-dedup-url=%s' % request.config.getoption('--rethinkdb-dedup-url'))
         # test these here only
@@ -1711,6 +1720,16 @@ def test_long_warcprox_meta(
         with pytest.raises(StopIteration):
             next(rec_iter)
 
+def test_socket_timeout_response(
+        warcprox_, http_daemon, https_daemon, archiving_proxies,
+        playback_proxies):
+    """Response will timeout because we use --socket-timeout=4 whereas the
+    target URL will return after 6 sec.
+    """
+    url = 'http://localhost:%s/slow-response' % http_daemon.server_port
+    response = requests.get(url, proxies=archiving_proxies, verify=False)
+    assert response.status_code == 502
+
 def test_empty_response(
         warcprox_, http_daemon, https_daemon, archiving_proxies,
         playback_proxies):

warcprox/main.py

@@ -162,6 +162,10 @@ def _build_arg_parser(prog='warcprox'):
             default=None, help=(
                 'host:port of tor socks proxy, used only to connect to '
                 '.onion sites'))
+    # Configurable connection socket timeout, default is 60 sec.
+    arg_parser.add_argument(
+            '--socket-timeout', dest='socket_timeout', type=float,
+            default=None, help=argparse.SUPPRESS)
     arg_parser.add_argument(
             '--crawl-log-dir', dest='crawl_log_dir', default=None, help=(
                 'if specified, write crawl log files in the specified '

warcprox/mitmproxy.py

@@ -205,12 +205,13 @@ class MitmProxyHandler(http_server.BaseHTTPRequestHandler):
     and records the bytes in transit as it proxies them.
     '''
     logger = logging.getLogger("warcprox.mitmproxy.MitmProxyHandler")
+    _socket_timeout = 60
 
     def __init__(self, request, client_address, server):
         threading.current_thread().name = 'MitmProxyHandler(tid={},started={},client={}:{})'.format(warcprox.gettid(), datetime.datetime.utcnow().isoformat(), client_address[0], client_address[1])
         self.is_connect = False
         self._headers_buffer = []
-        request.settimeout(60)  # XXX what value should this have?
+        request.settimeout(self._socket_timeout)
         http_server.BaseHTTPRequestHandler.__init__(self, request, client_address, server)
 
     def _determine_host_port(self):
@@ -247,8 +248,7 @@ class MitmProxyHandler(http_server.BaseHTTPRequestHandler):
             self._remote_server_sock = socket.socket()
             self._remote_server_sock.setsockopt(socket.IPPROTO_TCP, socket.TCP_NODELAY, 1)
 
-        # XXX what value should this timeout have?
+        self._remote_server_sock.settimeout(self._socket_timeout)
-        self._remote_server_sock.settimeout(60)
         self._remote_server_sock.connect((self.hostname, int(self.port)))
 
         # Wrap socket if SSL is required

warcprox/warcproxy.py

@@ -397,6 +397,9 @@ class SingleThreadedWarcProxy(http_server.HTTPServer, object):
                 WarcProxyHandler.onion_tor_socks_proxy_host = options.onion_tor_socks_proxy
                 WarcProxyHandler.onion_tor_socks_proxy_port = None
 
+        if options.socket_timeout:
+            WarcProxyHandler._socket_timeout = options.socket_timeout
+
         http_server.HTTPServer.__init__(
                 self, server_address, WarcProxyHandler, bind_and_activate=True)