Re: [Zope] Finding crash information
Not on my box. Completely silent. -Jon Michel Pelletier <michel@digicool.com> writes:
Do these sudden disapearances leave core files?
-Michel
Jon Prettyman wrote:
Not on my box. Completely silent.
-Jon
Michel Pelletier <michel@digicool.com> writes:
Do these sudden disapearances leave core files?
-Michel
If it's the same error we were seeing, then it isn't that Zope crashes, it hangs, becomes non-responsive. It's not spinning, it's just sitting there not answering. The reason why (in our case) is because of an uncaptured exception in the medusa code. Whenever medusa encounters an uncaptured exception in a thread, it closes it down. Well, it was encountering an uncaptured exception in the main listening thread, so it shut it down... now Zope isn't listening anymore... so it just sits there. If you want debug output without using debug mode, add this code to your start script (reverse the slashes for Win32, of course): STUPID_LOG_FILE=$PYTHONHOME/var/event.log export STUPID_LOG_FILE Debug information will get stored in event.log. -- Nick Garcia | ngarcia@codeit.com CodeIt Computing | http://codeit.com
Thanks a lot Nick I haven't tried out your suggestions yet but I *hope* they point to the problem. You write nevertheless that in your case Zope hangs, whereas in my case the process just dies silently and doesn't even trigger the manager to restart it. I will try my own supervisor on it and see how it behaves. Pavlos
participants (3)
-
Jon Prettyman -
Nick Garcia -
Pavlos Christoforou