I notice our server is rapidly responding 500 errors for random pages.
Our setup is pretty basic. We are running Nginx 0.7.6 with Unicorn and a
Rails app on a Ubuntu 10.4 server.
What ever the problem, it doesn’t seem to happen at the application
level since Unicorn logs don’t contain any errors. Next we looked at
Nginx’s error logs and there is nothing suspicious but even though the
site seemingly works when we monitor the Nginx’s access log there are a
lot of 500 error responses
for random pages of our site.
This is really strange since our server isn’t under a lot of load. It
uses a 4 core Linode instance and we have 4 Nginx worker processes.
Could someone people point us to the right direction?
Thanks
Hello!
On Wed, Mar 06, 2013 at 01:23:28AM +0100, Arman M. wrote:
This is really strange since our server isn’t under a lot of load. It
uses a 4 core Linode instance and we have 4 Nginx worker processes.
Could someone people point us to the right direction?
If it indeed happens in nginx, I would suppose it’s some
configuration problem, something like rewrite loop, or proxy loop,
or something similar. It should have reasons logged to error log
though. To make sure it’s indeed happens in nginx, and not
returned by your application, you may try logging $upstream_status
variable, see docs here:
http://nginx.org/en/docs/http/ngx_http_upstream_module.html
You may also use debug log to check what goes wrong in details,
see here:
http://nginx.org/en/docs/debugging_log.html
But in any case 0.7.6 is a 5 years old development version, and
you may want to upgrade.
–
Maxim D.
http://nginx.org/en/donation.html