Too many 400 bad request

Server NGINX
1

Hi everyone,

I’ve deployed a server for rsync public mirror,and I use nginx as web
server to display some html for visitors.

Now the problem is, too many 400 errors.

I paste “only a small amount of them” below:

=================================
216.216.33.242 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
174.128.254.98 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
216.216.33.242 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
207.99.82.229 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
173.255.246.181 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
206.53.187.66 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
173.255.246.181 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
216.40.205.154 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
67.222.104.104 - - [21/Aug/2012:03:02:05 -0700] “GET
/centos/5.8/extras/x86_64/repodata/repomd.xml HTTP/1.1” 200 2146 “-”
“urlgrabber/3.1.0 yum/3.2.22” “-”
75.125.106.50 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
75.125.106.50 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
64.28.85.10 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
38.107.176.34 - - [21/Aug/2012:03:02:05 -0700] “GET
/centos/5.8/os/i386/repodata/repomd.xml HTTP/1.1” 200 1140 “-”
“urlgrabber/3.1.0 yum/3.2.22” “-”
174.36.195.24 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
132.239.242.24 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
64.156.195.66 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
199.180.132.136 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
173.228.120.18 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
209.119.87.126 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
174.137.174.201 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
208.70.208.200 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”

Well,this really make me confused.And I’ve asked other mirror admins who
use nginx,they all have this problem.

Is it caused by nginx?
Do anyone know the solution?
Thanks.

Yours sincerely,
Christopher M.

Ambassador/Contributor of Fedora Project and many others.

2

Hello!

On Wed, Aug 22, 2012 at 12:38:23PM +0800, Christopher M. wrote:

216.216.33.242 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
174.128.254.98 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”
216.216.33.242 - - [21/Aug/2012:03:02:05 -0700] “-” 400 0 “-” “-” “-”

Try looking e.g. here for a possible cause:
http://mailman.nginx.org/pipermail/nginx/2011-September/029045.html

At least Chrome is known to cause such log entries if there aren’t
many resources on a page requested due to additional connections
being open and then closed without sending any single request in
them.

[…]

Well,this really make me confused.And I’ve asked other mirror admins who
use nginx,they all have this problem.

Is it caused by nginx?
Do anyone know the solution?
Thanks.

It’s just a way how nginx logs connections which are just opened
and then closed. Confusing part is that some other web servers
doesn’t seem to log anything in this case.

Maxim D.

3

I think I’ve met empty request…I will do a test tonight,thanks.