Any progress on the restart bug front? I am having trouble with regular shutdowns too, not only the previously reported restart weirdness.
When I start the linux binary install with the 'start' script on RedHat 6.1 (with the latest upgrades), I get strange behavior with 'shutdown' in the Control Panel. When I do shutdown, and go back to the management screen, Zope is still running, until I do *another* shutdown. Sometimes I even have to do three shutdowns to get Zope down. This happens with clean Zope 2.1.2 installs.
This is really annoying when you're developing products. The restart bug was difficult, but not being able to shut down reliably makes even my workaround unworkable.
So is there any progress on this? Presumably the restart and shutdown problems are related. Last I saw it was discarded as a RedHat 5.2 specific bug, but it definitely isn't. At least not here.
I have exactly the same problems on RH6.1 and Zope 2.1.2 compiled from source (I also did a cvs checkout yesterday, but it had the same problems). I first thought it was me doing something wrong, since it was my first attempt to write a product in python, and that it was "broken" almost all the time. I also get a lot of 2000-01-07T18:09:02 ERROR(200) ZODB Couldn't load state for '\000\000\000\000\000\000\002\320' is that related, or is that because of the broken product? It seems to disappear if I kill zope and all python processes and restart. I ended up starting zope with ./start and killing it with ctrl+c yesterday. I had enough problems trying to understand zclasses etc than to bother about if things got reloaded all the time... /Magnus Heino