In Files

Parent

Mongrel::HttpServer

This is the main driver of Mongrel, while the Mongrel::HttpParser and Mongrel::URIClassifier make up the majority of how the server functions. It’s a very simple class that just has a thread accepting connections and a simple HttpServer.process_client function to do the heavy lifting with the IO and Ruby.

You use it by doing the following:

  server = HttpServer.new("0.0.0.0", 3000)
  server.register("/stuff", MyNiftyHandler.new)
  server.run.join

The last line can be just server.run if you don’t want to join the thread used. If you don’t though Ruby will mysteriously just exit on you.

Ruby’s thread implementation is “interesting” to say the least. Experiments with many different types of IO processing simply cannot make a dent in it. Future releases of Mongrel will find other creative ways to make threads faster, but don’t hold your breath until Ruby 1.9 is actually finally useful.

Attributes

acceptor[R]
workers[R]
classifier[R]
host[R]
port[R]
throttle[R]
timeout[R]
num_processors[R]

Public Class Methods

new(host, port, num_processors=950, throttle=0, timeout=60) click to toggle source

Creates a working server on host:port (strange things happen if port isn’t a Number). Use HttpServer::run to start the server and HttpServer.acceptor.join to join the thread that’s processing incoming requests on the socket.

The num_processors optional argument is the maximum number of concurrent processors to accept, anything over this is closed immediately to maintain server processing performance. This may seem mean but it is the most efficient way to deal with overload. Other schemes involve still parsing the client’s request which defeats the point of an overload handling system.

The throttle parameter is a sleep timeout (in hundredths of a second) that is placed between socket.accept calls in order to give the server a cheap throttle time. It defaults to 0 and actually if it is 0 then the sleep is not done at all.

     # File lib/mongrel.rb, line 96
 96:     def initialize(host, port, num_processors=950, throttle=0, timeout=60)
 97:       
 98:       tries = 0
 99:       @socket = TCPServer.new(host, port) 
100:       
101:       @classifier = URIClassifier.new
102:       @host = host
103:       @port = port
104:       @workers = ThreadGroup.new
105:       @throttle = throttle / 100.0
106:       @num_processors = num_processors
107:       @timeout = timeout
108:     end

Public Instance Methods

configure_socket_options() click to toggle source
     # File lib/mongrel.rb, line 246
246:     def configure_socket_options
247:       case RUBY_PLATFORM
248:       when /linux/
249:         # 9 is currently TCP_DEFER_ACCEPT
250:         $tcp_defer_accept_opts = [Socket::SOL_TCP, 9, 1]
251:         $tcp_cork_opts = [Socket::SOL_TCP, 3, 1]
252:       when /freebsd(([1-4]\..{1,2})|5\.[0-4])/
253:         # Do nothing, just closing a bug when freebsd <= 5.4
254:       when /freebsd/
255:         # Use the HTTP accept filter if available.
256:         # The struct made by pack() is defined in /usr/include/sys/socket.h as accept_filter_arg
257:         unless `/sbin/sysctl -nq net.inet.accf.http`.empty?
258:           $tcp_defer_accept_opts = [Socket::SOL_SOCKET, Socket::SO_ACCEPTFILTER, ['httpready', nil].pack('a16a240')]
259:         end
260:       end
261:     end
graceful_shutdown() click to toggle source

Performs a wait on all the currently running threads and kills any that take too long. It waits by @timeout seconds, which can be set in .initialize or via mongrel_rails. The @throttle setting does extend this waiting period by that much longer.

     # File lib/mongrel.rb, line 239
239:     def graceful_shutdown
240:       while reap_dead_workers("shutdown") > 0
241:         STDERR.puts "Waiting for #{@workers.list.length} requests to finish, could take #{@timeout + @throttle} seconds."
242:         sleep @timeout / 10
243:       end
244:     end
process_client(client) click to toggle source

Does the majority of the IO processing. It has been written in Ruby using about 7 different IO processing strategies and no matter how it’s done the performance just does not improve. It is currently carefully constructed to make sure that it gets the best possible performance, but anyone who thinks they can make it faster is more than welcome to take a crack at it.

     # File lib/mongrel.rb, line 115
115:     def process_client(client)
116:       begin
117:         parser = HttpParser.new
118:         params = HttpParams.new
119:         request = nil
120:         data = client.readpartial(Const::CHUNK_SIZE)
121:         nparsed = 0
122: 
123:         # Assumption: nparsed will always be less since data will get filled with more
124:         # after each parsing.  If it doesn't get more then there was a problem
125:         # with the read operation on the client socket.  Effect is to stop processing when the
126:         # socket can't fill the buffer for further parsing.
127:         while nparsed < data.length
128:           nparsed = parser.execute(params, data, nparsed)
129: 
130:           if parser.finished?
131:             if not params[Const::REQUEST_PATH]
132:               # it might be a dumbass full host request header
133:               uri = URI.parse(params[Const::REQUEST_URI])
134:               params[Const::REQUEST_PATH] = uri.path
135:             end
136: 
137:             raise "No REQUEST PATH" if not params[Const::REQUEST_PATH]
138: 
139:             script_name, path_info, handlers = @classifier.resolve(params[Const::REQUEST_PATH])
140: 
141:             if handlers
142:               params[Const::PATH_INFO] = path_info
143:               params[Const::SCRIPT_NAME] = script_name
144: 
145:               # From http://www.ietf.org/rfc/rfc3875 :
146:               # "Script authors should be aware that the REMOTE_ADDR and REMOTE_HOST
147:               #  meta-variables (see sections 4.1.8 and 4.1.9) may not identify the
148:               #  ultimate source of the request.  They identify the client for the
149:               #  immediate request to the server; that client may be a proxy, gateway,
150:               #  or other intermediary acting on behalf of the actual source client."
151:               params[Const::REMOTE_ADDR] = client.peeraddr.last
152: 
153:               # select handlers that want more detailed request notification
154:               notifiers = handlers.select { |h| h.request_notify }
155:               request = HttpRequest.new(params, client, notifiers)
156: 
157:               # in the case of large file uploads the user could close the socket, so skip those requests
158:               break if request.body == nil  # nil signals from HttpRequest::initialize that the request was aborted
159: 
160:               # request is good so far, continue processing the response
161:               response = HttpResponse.new(client)
162: 
163:               # Process each handler in registered order until we run out or one finalizes the response.
164:               handlers.each do |handler|
165:                 handler.process(request, response)
166:                 break if response.done or client.closed?
167:               end
168: 
169:               # And finally, if nobody closed the response off, we finalize it.
170:               unless response.done or client.closed? 
171:                 response.finished
172:               end
173:             else
174:               # Didn't find it, return a stock 404 response.
175:               client.write(Const::ERROR_404_RESPONSE)
176:             end
177: 
178:             break #done
179:           else
180:             # Parser is not done, queue up more data to read and continue parsing
181:             chunk = client.readpartial(Const::CHUNK_SIZE)
182:             break if !chunk or chunk.length == 0  # read failed, stop processing
183: 
184:             data << chunk
185:             if data.length >= Const::MAX_HEADER
186:               raise HttpParserError.new("HEADER is longer than allowed, aborting client early.")
187:             end
188:           end
189:         end
190:       rescue EOFError,Errno::ECONNRESET,Errno::EPIPE,Errno::EINVAL,Errno::EBADF
191:         client.close rescue nil
192:       rescue HttpParserError => e
193:         STDERR.puts "#{Time.now}: HTTP parse error, malformed request (#{params[Const::HTTP_X_FORWARDED_FOR] || client.peeraddr.last}): #{e.inspect}"
194:         STDERR.puts "#{Time.now}: REQUEST DATA: #{data.inspect}\n---\nPARAMS: #{params.inspect}\n---\n"
195:       rescue Errno::EMFILE
196:         reap_dead_workers('too many files')
197:       rescue Object => e
198:         STDERR.puts "#{Time.now}: Read error: #{e.inspect}"
199:         STDERR.puts e.backtrace.join("\n")
200:       ensure
201:         begin
202:           client.close
203:         rescue IOError
204:           # Already closed
205:         rescue Object => e
206:           STDERR.puts "#{Time.now}: Client error: #{e.inspect}"
207:           STDERR.puts e.backtrace.join("\n")
208:         end
209:         request.body.close! if request and request.body.class == Tempfile
210:       end
211:     end
reap_dead_workers(reason='unknown') click to toggle source

Used internally to kill off any worker threads that have taken too long to complete processing. Only called if there are too many processors currently servicing. It returns the count of workers still active after the reap is done. It only runs if there are workers to reap.

     # File lib/mongrel.rb, line 217
217:     def reap_dead_workers(reason='unknown')
218:       if @workers.list.length > 0
219:         STDERR.puts "#{Time.now}: Reaping #{@workers.list.length} threads for slow workers because of '#{reason}'"
220:         error_msg = "Mongrel timed out this thread: #{reason}"
221:         mark = Time.now
222:         @workers.list.each do |worker|
223:           worker[:started_on] = Time.now if not worker[:started_on]
224: 
225:           if mark - worker[:started_on] > @timeout + @throttle
226:             STDERR.puts "Thread #{worker.inspect} is too old, killing."
227:             worker.raise(TimeoutError.new(error_msg))
228:           end
229:         end
230:       end
231: 
232:       return @workers.list.length
233:     end
register(uri, handler, in_front=false) click to toggle source

Simply registers a handler with the internal URIClassifier. When the URI is found in the prefix of a request then your handler’s HttpHandler::process method is called. See Mongrel::URIClassifier#register for more information.

If you set in_front=true then the passed in handler will be put in the front of the list for that particular URI. Otherwise it’s placed at the end of the list.

     # File lib/mongrel.rb, line 326
326:     def register(uri, handler, in_front=false)
327:       begin
328:         @classifier.register(uri, [handler])
329:       rescue URIClassifier::RegistrationError => e
330:         handlers = @classifier.resolve(uri)[2]
331:         if handlers
332:           # Already registered
333:           method_name = in_front ? 'unshift' : 'push'
334:           handlers.send(method_name, handler)
335:         else
336:           raise
337:         end
338:       end
339:       handler.listener = self
340:     end
run() click to toggle source

Runs the thing. It returns the thread used so you can “join” it. You can also access the HttpServer::acceptor attribute to get the thread later.

     # File lib/mongrel.rb, line 265
265:     def run
266:       BasicSocket.do_not_reverse_lookup=true
267: 
268:       configure_socket_options
269: 
270:       if defined?($tcp_defer_accept_opts) and $tcp_defer_accept_opts
271:         @socket.setsockopt(*$tcp_defer_accept_opts) rescue nil
272:       end
273: 
274:       @acceptor = Thread.new do
275:         begin
276:           while true
277:             begin
278:               client = @socket.accept
279:   
280:               if defined?($tcp_cork_opts) and $tcp_cork_opts
281:                 client.setsockopt(*$tcp_cork_opts) rescue nil
282:               end
283:   
284:               worker_list = @workers.list
285:   
286:               if worker_list.length >= @num_processors
287:                 STDERR.puts "Server overloaded with #{worker_list.length} processors (#@num_processors max). Dropping connection."
288:                 client.close rescue nil
289:                 reap_dead_workers("max processors")
290:               else
291:                 thread = Thread.new(client) {|c| process_client(c) }
292:                 thread[:started_on] = Time.now
293:                 @workers.add(thread)
294:   
295:                 sleep @throttle if @throttle > 0
296:               end
297:             rescue StopServer
298:               break
299:             rescue Errno::EMFILE
300:               reap_dead_workers("too many open files")
301:               sleep 0.5
302:             rescue Errno::ECONNABORTED
303:               # client closed the socket even before accept
304:               client.close rescue nil
305:             rescue Object => e
306:               STDERR.puts "#{Time.now}: Unhandled listen loop exception #{e.inspect}."
307:               STDERR.puts e.backtrace.join("\n")
308:             end
309:           end
310:           graceful_shutdown
311:         ensure
312:           @socket.close
313:           # STDERR.puts "#{Time.now}: Closed socket."
314:         end
315:       end
316: 
317:       return @acceptor
318:     end
stop(synchronous=false) click to toggle source

Stops the acceptor thread and then causes the worker threads to finish off the request queue before finally exiting.

     # File lib/mongrel.rb, line 351
351:     def stop(synchronous=false)
352:       @acceptor.raise(StopServer.new)
353: 
354:       if synchronous
355:         sleep(0.5) while @acceptor.alive?
356:       end
357:     end
unregister(uri) click to toggle source

Removes any handlers registered at the given URI. See Mongrel::URIClassifier#unregister for more information. Remember this removes them all so the entire processing chain goes away.

     # File lib/mongrel.rb, line 345
345:     def unregister(uri)
346:       @classifier.unregister(uri)
347:     end

Disabled; run with --debug to generate this.

[Validate]

Generated with the Darkfish Rdoc Generator 1.1.6.