Linode Forum

[Mark_Stenks]

Hi @ all,

i don't know how to tuning the performance about my web servers...
i have 2 linode 1024 balanced with LinodeBalancer, serving a website with a high traffic.

configured like this :

NodeBalancers :
Algorithm = leastConnection
Session Stickiness = Table

two web servers :

Code: Ubuntu 10.04 32Bit (Linode Kernel 3.2.1-linode40)
nginx version: nginx/1.0.11
PHP 5.3.5-1ubuntu7.2ppa1~lucid with Suhosin-Patch
php-apc module enabled.
php5-fpm

all this packets was installed from the ppa :

Code: ppa:nginx/stable
ppa:nginx/php5

nginx is configured like this :

Code: user www-data;
  worker_processes  4;
  worker_rlimit_nofile 32768;
  error_log  /var/log/nginx/error.log;
  pid        /var/run/nginx.pid;
  events {
      worker_connections  4096;
      multi_accept on;
      accept_mutex_delay 50ms;
  }
http {
    tcp_nodelay        on;
    expires -1;
    server_tokens off;
    sendfile on;
    tcp_nopush on;
    keepalive_timeout 60 60;
    types_hash_max_size 2048;
    include /etc/nginx/mime.types;
    default_type application/octet-stream;
    access_log /var/log/nginx/access.log;
    error_log /var/log/nginx/error.log;
    map $scheme $server_https {
        default off;
        https on;
    }
    include /etc/nginx/conf.d/*.conf;
    include /etc/nginx/sites-enabled/*;
}

php-fpm :

Code: [www]
listen = /etc/php5/php5-fpm.sock
user = www-data
group = www-data
pm = dynamic
pm.max_children = 8
pm.start_servers = 2
pm.min_spare_servers = 1
pm.max_spare_servers = 3
pm.max_requests = 1000
request_terminate_timeout = 30s
#request_slowlog_timeout = 1s
#slowlog = /var/log/php5-fpm/$pool.log.slow
rlimit_files = 32768
rlimit_core = unlimited
chdir = /

and the website is configured like this :

Code: server {
    server_name webdomain.org;
    listen   80; ## listen for ipv4; this line is default and implied
    access_log  /var/log/nginx/web_access.log combined  buffer=32k;
    error_log   /var/log/nginx/web_error.log info;
    root /home/web/;

    index index.php index.html index.htm;

    try_files $uri $uri/ @rewrite;

    location @rewrite {
        rewrite ^/(.*)$ /index.php/$1;
    }
    location ~ \.php {
        # try_files $uri =404;
        #auth_basic            "Restricted";
        #auth_basic_user_file  /etc/nginx/htpasswd;

        fastcgi_index index.php;
        fastcgi_pass   backend;

        include fastcgi_params;
        fastcgi_split_path_info ^(.+\.php)(/.+)$;
        fastcgi_param PATH_INFO $fastcgi_path_info;
        fastcgi_param PATH_TRANSLATED $document_root$fastcgi_path_info;
        fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
    }

    location ~ /\.ht {
        deny all;
    }
}
upstream backend {
        server unix:/etc/php5/php5-fpm.sock;
}


this configuration is the same in the web server no.2

i'm trying to test the website with ab and this is the results :

ab -n 50000 -c 1000 http://localhost/

Code: Server Software:        nginx
Server Hostname:        localhost
Server Port:            80

Document Path:          /en
Document Length:        188 bytes

Concurrency Level:      1000
Time taken for tests:   7.112 seconds
Complete requests:      50000
Failed requests:        0
Write errors:           0
Non-2xx responses:      50296
Total transferred:      19011888 bytes
HTML transferred:       9455648 bytes
Requests per second:    7030.27 [#/sec] (mean)
Time per request:       142.242 [ms] (mean)
Time per request:       0.142 [ms] (mean, across all concurrent requests)
Transfer rate:          2610.52 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:       28   60  11.1     60     108
Processing:    36   80  16.4     80     155
Waiting:       25   61  16.5     60     121
Total:         90  141  14.2    140     203

Percentage of the requests served within a certain time (ms)
  50%    140
  66%    146
  75%    150
  80%    152
  90%    159
  95%    164
  98%    173
  99%    182
 100%    203 (longest request)

from this output is not bad, (i think) but if i test the website with loadimpact i have a bad bad bad result.

500 concurrent connection in 10 minutes :



i don't know where to begin.
any help is very very appreciated!

Thanks

[Guspaz]

Your benchmark is showing you can serve over 7000 requests per second... It does't look like you need any tuning to me, unless the page you were benchmarking (your index file) isn't representative of actual user load.