nginx服务器access日志中大量400 bad request错误的解决方法

在access.log中有大量400错误,并以每天几百m的速度增加,占用大量空间.

tail -f /opt/nginx/logs/access.log

    116.236.228.180 - - [15/dec/2010:11:00:15 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:15 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:15 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:15 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:15 +0800] "-" 400 0 "-" "-"
    119.97.196.7 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    119.97.196.7 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    219.243.95.197 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"
    116.236.228.180 - - [15/dec/2010:11:00:16 +0800] "-" 400 0 "-" "-"

网上大把的文章说是http头/cookie过大引起的,可以修改nginx.conf中两参数来修正.

    client_header_buffer_size 16k;
          large_client_header_buffers 4 32k;

修改后

    client_header_buffer_size 64k;
         large_client_header_buffers 4 64k;

没有效果,就算我把nginx0.7.62升到最新的0.8.54也没能解决.
在官方论坛中nginx作者提到空主机头不会返回自定义的状态码,是返回400错误.

最后修正如下
改为原先的值

    client_header_buffer_size 16k;
         large_client_header_buffers 4 32k;

关闭默认主机的日志记录就可以解决问题

    server {
    listen *:80 default;
    server_name _;
    return 444;
    access_log   off;
    }