SocketServer.Fo rkingMixIn and zombie processes

[nisimura]

Hi,

I noticed that when I used SocketServer.Fo rkingMixIn
(ForkingTCPServ er), there were always zombie processes around. I
searched for where waitpid() is called in ForkingMixIn and found it in
SocketServer.py :

def process_request (self, request, client_address) :
"""Fork a new subprocess to process the request."""
self.collect_ch ildren()
pid = os.fork()
if pid:
# Parent process
if self.active_chi ldren is None:
self.active_chi ldren = []
self.active_chi ldren.append(pi d)
self.close_requ est(request)
return
else:
# Child process.
# This must never return, hence os._exit()!
try:
self.finish_req uest(request, client_address)
os._exit(0)
except:
try:
self.handle_err or(request, client_address)
finally:
os._exit(1)

So waitpid() is called just before fork(). This means that if next
request does not come in, zombie processes are not reclaimed forever!
Is this an intentional behavior? Is there any workaround?

--
Kazu Nisimura