在ec2上使用多php-fpm服务器实例运行nginx时连接错误

Question

我试图在AmazonEC2上运行php5，使用多个分离的php fpm服务器，通过nginx上的上游块进行负载平衡。我正在用两个t1进行测试。微实例，但当我尝试加载php文件时，浏览器上出现502错误网关。（静态html文件工作正常，但无法使php文件正常工作。）

这是我的nginx错误日志。

2012/07/11 12:28:21[错误]18626#0:*1 recv（）在从上游读取响应头时失败（104：由对等方重置连接），客户端：xxx。xxx。xxx。xxx，服务器：www.example。com，请求：“GET/HTTP/1.1”，上游：fastcgi://10.xxx.xxx.xxx:9000主持人：“www.example。com“

有时我也会这样。

2012/07/11 13:25:51[错误]1157#0:*4从上游读取响应标头时，上游连接过早关闭，客户端：xxx。xxx。xxx。xxx，服务器：www.example。com，请求：“GET/HTTP/1.1”，上游：fastcgi://10.xxx.xxx.xxx:9000主持人：“www.example。com“

我花了一些时间从ec2 sequritygroup/iptables打开9000端口，并在nginx和php-fpm上声明本地ip地址，所以我认为这不是问题。（我曾经有连接拒绝错误日志）

有人能帮我吗？？下面是我的服务器设置和首选项。

[实例1]

• t1.microCentOS 6.2.2
• nginx/1.2.2

[实例2]

• t1。micro CentOS 6.2。2
• PHP5.3。14（fpm fcgi）Zend发动机v2。3.0，带加速装置v0。9.6

[nginx.conf]

user  nginx nginx;
worker_processes  1;
worker_rlimit_nofile 1024;
worker_priority -5;

error_log  /var/log/nginx/error.log warn;
pid        /var/run/nginx.pid;

events {
    multi_accept on;
    worker_connections  1024;
}


http {
    include       /etc/nginx/mime.types;
    default_type  application/octet-stream;
    server_tokens   off;

    log_format  main  '$remote_addr - $remote_user [$time_local] "$request" '
                      '$status $body_bytes_sent "$http_referer" '
                      '"$http_user_agent" "$http_x_forwarded_for"';

    access_log  /var/log/nginx/access.log  main;

    sendfile        on;
    #tcp_nopush     on;

    keepalive_timeout  0;

    gzip  on;

    upstream apserver {
        ip_hash;
        server ip-10-xxx-xxx-xxx.ap-northeast-1.compute.internal:9000;
    }

    include /etc/nginx/conf.d/*.conf;
}

[example.conf]

server {
    listen       80;
    server_name  www.example.com;

    charset utf-8;
    access_log  /var/log/nginx/www.example.com.access.log  main;
    error_log   /var/log/nginx/www.example.com.error.log  debug;
    root    /var/www;

    location / {
        index   index.php index.html index.html;
        if (-f $request_filename) {
            expires max;
            break;
        }

        if (!-e $request_filename) {
            rewrite ^(.+)/index\.php/(.*)$ $1/index.php?q=$2 last;
        }
    }

        location ~ \.php$ {
        fastcgi_send_timeout  10m;
            fastcgi_read_timeout 10m;
        fastcgi_connect_timeout 10m;
        fastcgi_pass    apserver;
        fastcgi_index   index.php;
        fastcgi_param   SCRIPT_FILENAME     $document_root$fastcgi_script_name;
        fastcgi_param   PATH_INFO       $fastcgi_script_name;
        include     /etc/nginx/fastcgi_params;
    }
}

[php fpm.d/www.conf]

[www]
listen = ip-10-xxx-xxx-xxx.ap-northeast-1.compute.internal:9000
listen.backlog = -1
listen.allowed_clients = ip-10-yyy-yyy-yyy.ap-northeast-1.compute.internal

; Tried testing with below and got the same error
;listen = 9000
;listen.allowed_clients = any

listen.owner = prod
listen.group = prod
listen.mode = 0666
user = prod
group = prod

pm = dynamic
pm.max_children = 10
pm.start_servers = 5
pm.min_spare_servers = 3
pm.max_spare_servers = 8
pm.max_requests = 500

request_terminate_timeout = 30
request_slowlog_timeout = 2
slowlog = /var/log/php-fpm/www-slow.log

php_admin_value[error_log] = /var/log/php-fpm/www-error.log
php_admin_flag[log_errors] = on
php_admin_flag[expose_php] = off

Answer 1

我终于明白了！！

我仍然不知道为什么，但通过在我的每个实例上分配弹性ip，并使用私有ip而不是私有DNS地址，解决了我的问题。

所以我的conf文件现在看起来像这样以防万一。

[nginx.conf]

upstream apserver {
    ip_hash;
    server 10.xxx.xxx.xxx:9000;
}

[php fpm.d/www.conf]

[www]
listen = 10.xxx.xxx.xxx:9000
listen.backlog = -1
listen.allowed_clients = 10.yyy.yyy.yyy

谢谢！