are there any known artifical limitations of
WinNT Workstation as opposed to e.g. WinNT
Server regarding the scheduling scheme of
the threads?

For e.g. the medusa thread receives a http
request and "passes" it to the pool of ZORB

We have a small Zope product which remains
sleeping in a loop until a file contains a
matching line. We trigger the execution of
this product by requesting the proper URL
and also are able to monitor its doings.

With WinNT Server and W2K Server Zope is
still able to respond towards further http
requests, even though the one ZORB worker
thread is actually waiting in the loop.

With WinNT Workstation we receive no errors
and the medusa thread still receives requests
which are queued. But there is no real progress
along the ZORB threads because only one worker
seems to receive timeslots and that is the one
waiting and sleeping within the loop. The rest
of the thread pool is "active" but seems to
receive no time slots. So effectivly Zope comes
to a quite a halt but will of course continue
nicely, if the waiting ZORB thread is allowed
to exit the loop. The queued requests are then
handled as if nothing has happend.

Is this a known behaviour of WinNT Workstation
as opposed to e.g. WinNT Server or W2K Server?


Zope-Dev maillist  -  [EMAIL PROTECTED]
**  No cross posts or HTML encoding!  **
(Related lists - 
 http://lists.zope.org/mailman/listinfo/zope )

Reply via email to