Request from 10.x.x.x a few seconds after original request

Question

When debugging my Rails application I noticed often when I do a request, about 4 seconds later the same page is requested from a 10.x.x.x IP address. I googled for the IP addresses and if I understand correctly it is a router. I find this very strange. Can anyone explain what this is?

Excerpt from my Rails production.log:

Started GET "/blogs/71" for (My IP address) at Sun Oct 09 12:14:12 +0000 2011
  Processing by BlogsController#show as HTML
  Parameters: {"id"=>"71"}
Rendered shared/_google_maps.html.erb (0.2ms)
Rendered blogs/_comments.html.erb (22.0ms)
Rendered blogs/_comment_form.html.erb (2.9ms)
Rendered shared/_header.html.erb (23.8ms)
Rendered shared/_menu.html.erb (0.1ms)
Rendered shared/_footer.html.erb (0.3ms)
Rendered blogs/show.html.erb within layouts/application (155.2ms)
Completed 200 OK in 157ms (Views: 155.0ms | ActiveRecord: 1.1ms)


Started GET "/blogs/71" for 10.36.219.118 at Sun Oct 09 12:14:16 +0000 2011
  Processing by BlogsController#show as 
  Parameters: {"id"=>"71"}
Rendered shared/_google_maps.html.erb (0.1ms)
Rendered blogs/_comments.html.erb (16.1ms)
Rendered blogs/_comment_form.html.erb (12.0ms)
Rendered shared/_header.html.erb (0.1ms)
Rendered shared/_menu.html.erb (0.1ms)
Rendered shared/_footer.html.erb (0.2ms)
Rendered blogs/show.html.erb within layouts/application (123.2ms)
Completed 200 OK in 152ms (Views: 132.5ms | ActiveRecord: 1.1ms)

Don't know if it's relevant, but the server is Ubuntu, web server is nginx with passenger and REE.

I have a recaptcha on the pages where this occurs, could that be causing this?

Not sure if this is off-topic here. I apologize in advance if it is.

Answer 1

10.X.X.X is a local network address, so without being inside your building, we can't tell what it is from here

Look in the nginx access logs -- is nginx logging multiple requests happening? If so, looking at the user-agent string should tell you what software is making them

Answer 2

I'd start by tracing the request-path through the entire stack. It could look something like this:

1. The request is received at TCP/80 on Websrv.
2. Nginx is bound to TCP/80 and processes the request
3. The request is handled by these specific directives in the config file
4. The request is passed off to Passenger, where it enters the Ruby environment

If I had to guess where the problem may sit (and I do have to guess) I'd put money on #3 up there. Something in the nginx config is causing the request to duplicate and come around again, with the IP address changing part way through.

Then examine your config to see where the heck that 10 address comes from; it could be assigned to a different NIC on your server, a proxy engine somewhere, or on a completely different device on your network. Locating that IP address will help isolate the double-request issue.