More fcgi problems on apache


#1

I’m having tons of problems with fcgi on apache2. The app didn’t
crash over the weekend when it wasn’t used. I ran it this morning and
it was fine. It’s an inhouse app so it wasn’t used for 3 days over the
weekend. It did however crash today. Does this sound like a bad fcgi
setup or just a very buggy fcgi/Apache combo?

Charlie B.
http://www.recentrambles.com


#2

Give up now and go with mod_proxy, mod_proxy_balancer,
and Mongrel instead.

– -- Tom M.


#3

The decision isn’t mine. I’m just the programmer. This is the first
app my company has ever ran that isn’t plain vanilla cgi and the first
non perl app. I think the company is going to try out lighty ran behind
apache. They aren’t keen on using mongrel quite yet. It’s still a
little too new for them. Does the above setup sound good?

On Tue, 2006-05-30 at 12:35 -0700, Tom M. wrote:

3 days over the weekend. It did however crash today. Does this
sound like a bad fcgi setup or just a very buggy fcgi/Apache combo?


Rails mailing list
removed_email_address@domain.invalid
http://lists.rubyonrails.org/mailman/listinfo/rails

Charlie B.
http://www.recentrambles.com


#4

I had an app that was continually crashing after a few days, and it
seemed
like a usage thing, happening at 4am randomly every other night, often
2-3
days it would run fine.

I figured out it was my logrotate cron job - restarting apache (and
lighttpd
especially). That, and don’t put your fcgi sockets in /tmp, especially
if
you have a cron job to clean that out.

I took a moment to learn how to use the external spinners/reapers
scripts to
launch external fastcgi instances as well, and now lighttpd and apache
restarts don’t mess with my fastcgi instances.

hope that helps.

Charles Brian Q.
www.seebq.com