[Zope] ZoperServer hang up ...

[Dieter Maurer]

Volkert Barr writes:
 > our ZoperServer (Zope-2.4.1-linux2-x86.tar, running on Suse 7.1) hangs
 > up with=20
 > the follwing message while accessing it via a Netscape Browser. Any
 > hints ?
 >=20
 > With reagrds,
 > Volkert
 >=20
 >=20
 > 2001-11-02T09:57:48 ERROR(200) ZServer uncaptured python exception,
 > closing channel <select-trigger (pipe) at 81e7ad4>
 > (exceptions.OSError:[Errno 11] Die Ressource ist zur Zeit nicht
 > verf=FCgbar [/home/project/ZopeServer/lib/python2.1/asyncore.py|poll|9=
5]
 > [/home/project/ZopeServer/lib/python2.1/asyncore.py|handle_read_event|=
383]
 > [/home/project/ZopeServer/ZServer/medusa/thread/select_trigger.py|hand=
le_read|77]
 > [/home/project/ZopeServer/lib/python2.1/asyncore.py|recv|335]
 > [/home/project/ZopeServer/lib/python2.1/asyncore.py|recv|517])
 > Unhandled exception in thread:
That's a normal (and usually harmless) message when the client
closes the connection before Zope could return the response.

 > Traceback (most recent call last):
 >   File "/home/project/ZopeServer/ZServer/PubCore/ZServerPublisher.py",
 > line 97, in __init__
 >     response._finish()
 >   File "/home/project/ZopeServer/ZServer/HTTPResponse.py", line 277, i=
n
 > _finish
 >     stdout.close()
 >   File "/home/project/ZopeServer/ZServer/HTTPResponse.py", line 334, i=
n
 > close
 >     Wakeup()
 >   File
 > "/home/project/ZopeServer/ZServer/medusa/thread/select_trigger.py", li=
ne
 > 74, in pull_trigger
 >     os.write (self.trigger, 'x')
 > OSError: [Errno 32] Daten=FCbergabe unterbrochen (broken pipe)
 > Unhandled exception in thread:
 > Traceback (most recent call last):
 >   File "/home/project/ZopeServer/ZServer/PubCore/ZServerPublisher.py",
 > line 97, in __init__
 >     response._finish()
 >   File "/home/project/ZopeServer/ZServer/HTTPResponse.py", line 277, i=
n
 > _finish
 >     stdout.close()
 >   File "/home/project/ZopeServer/ZServer/HTTPResponse.py", line 334, i=
n
 > close
 >     Wakeup()
 >   File
 > "/home/project/ZopeServer/ZServer/medusa/thread/select_trigger.py", li=
ne
 > 74, in pull_trigger
 >     os.write (self.trigger, 'x')
 > OSError: [Errno 32] Daten=FCbergabe unterbrochen (broken pipe)
Maybe, the ZServer thread died by this exception.
Then, Zope could no longer respond to HTTP requests.
Do you lose one thread after the hand (look at the number of
Python processes before and after the hangup).

If so, this would be a serious bug in ZServer - newly introduced
not to far ago...



Dieter