Author: Helwig, Till Helge
Thanks for your ideas.
Sadly, I can't use the built-in server monitor because our version doesn't have
one. And finding out which template causes the problems is quite hard because the
applications is built upon Fusebox and Mach-II. Yes...both of them. So the
template context is usually insanely huge and most of the time the reported error
location is index.cfm, which is only the Mach-II controller. Annoying.
I found an entry in the logs telling me that some request exceeded the time limit
because of tag cfloop. Some digging unearthed some hilariously bad piece of code,
which I deep responsible. A loop over a query contains a set of new cfquery tags.
I assume that at some point either the processing in CF or the database gets
bored of this and breaks down. ;)
I only wonder why the whole server doesn't work anymore afterwards. From the
error message it looks like the overly long runtime is recognized and the request
is being aborted. So why doesn't the server recover? I can't have an application
that is able to break the server. Any ideas what I could do about that?
Refactoring the code is the obvious solution, but right now I don't have the time
(and the nerves).
Von: Russ Michaels [mailto:email@example.com]
Gesendet: Montag, 11. Juni 2012 17:30
Betreff: Re: CF breaks down with error 500 (GC)
check the runtime logs for some hints.
Also I would suggest installing FusionReactor to do some further debugging and
Have you made use of the built in server monitor as this can tell you info about
memory usage and what is using memory where.
----- Excess quoted text cut - see Original Post for more -----