The symptoms are as follows: 1. Go to a given URL, which doesn't respond (browser sits and spins) 2. Doing a 'top' on the server shows python nowhere to be seen (so it's not infinite-loop-processor-hogging) 3. ps shows the python threads are all still there, but none of them will respond. 3. doing a ./stop sucessfully stops Zope (the broswer returns a 'reset by peer') 4. doing a ./start yields the following the the stupid log file:
Can anyone suggest how I might go about tracking this down? It's really beginning to piss me off. That said, I've seen quite a few posts now complaining about Zope hanging and crashing, so I'm actually beginning to worry about using Zope from a stability point of view.
These symptoms are indeed very similar to what we had on the demo.iuveno-net.de site. We had an average "up-time" of around an hour or so, and the system also didn't really crash but just "hang" ... Right now we don't face these problems any more. As far as I know, the only thing Stephan changed on the server was exchanging the CoreSession tracking by the newest version (0.8 I think), which had to be patched to run for us AFAIK. Now the server is reliable again. I was really scared by this issue, too, as we had the same problems on a demo server for a customer ... Joachim.