Hi All,
I posted a while back about how my app gets really slow after a while
and dies. When I kill Mongrel, I get a trace that looks like it came
out of the ruby profiler. (see my earlier post if interested:
http://groups.google.com/group/rubyonrails-talk/browse_frm/thread/da1523ccbc1547ce
)
What dawned on me today was that the page that seems to set off this
slowness happens to be the user profile page. Maybe that’s a
coincidence… but maybe not.
I can load the user profile page thousands of times in a functional test
with no issues, but on a real server it crawls (7-10 seconds, eventually
killing the server). I’m wondering if that’s because in the real
environment it’s invoking the routes, and the fact that my route has the
word “profile” in it could somehow be invoking the profiler. My route
looks like this:
map.connect ‘/person/:id/section/profile’, :controller =>
‘mycontroller’, :action => ‘user_profile’
Is it possible that this is somehow invoking the profiler? It seems
impossible, but admittedly I don’t know the ins and outs of routes.
I’m having difficulties troubleshooting this because it’s not
reproducible in the test environment. Any ideas?
Many thanks,
Tom