gevent-multiprocess

gevent-multiprocess.py

import sys
from gevent import server
from gevent.baseserver import _tcp_listener
from gevent.monkey import patch_all; patch_all()
from multiprocessing import Process, current_process, cpu_count


def note(format, *args):
    sys.stderr.write('[%s]\t%s\n' % (current_process().name, format%args))

def echo(socket, address):
    print 'New connection from %s:%s' % address
    fileobj = socket.makefile()
    fileobj.write('Welcome to the echo server! Type quit to exit.\r\n')
    fileobj.write('In %s\r\n' % current_process().name)
    fileobj.flush()
    while True:
        line = fileobj.readline()
        if not line:
            print "client disconnected"
            break
        if line.strip().lower() == 'quit':
            print "client quit"
            break
        fileobj.write(current_process().name + '\t' + line)
        fileobj.flush()
        print "echoed", repr(line)

listener = _tcp_listener(('127.0.0.1', 8001)) 

def serve_forever(listener):
    note('starting server')
    server.StreamServer(listener, echo).serve_forever()
    
number_of_processes = 5
print 'Starting %s processes' % number_of_processes
for i in range(number_of_processes):
    Process(target=serve_forever, args=(listener,)).start()

serve_forever(listener)

Do you have an example of something similar for pywsgi? It'd be great to use gevent without gunicorn and take advantage of multiple processes!

It says it's 5 processes but eventually it will be 6 because you start the main in the end...
What's the difference between just forking with gevent.hubs?

I'm using gevent 1.0 on Mac Maverick, there's no _tcp_listener for gevent.baseserver, do you know why?

It has been moved to gevent.server in v1.0, so:

-from gevent.baseserver import _tcp_listener
+from gevent.server import _tcp_listener