Varnish系列之四：负载均衡的架构与实现

vcl中定义调度器说明：

    1. 轮询方式

backend appServer1 {

    .host = "172.16.100.52";

    .port = "80";

    .probe = {

        .url = "/test.html";

        .expected_response = 200;

    }

}

backend appServer2 {

    .host = "172.16.100.53";

    .port = "80";

}

import directors;

sub vcl_init {

    new rr = directors.round_robin();

    rr.add_backend(appServer1);

    rr.add_backend(appServer2);

}

sub vcl_recv {

    set req.backend_hint = rr.backend();

    ......

}

    2. hash方式

backend appServer1 {

    .host = "172.16.100.52";

    .port = "80";

}

backend appServer2 {

    .host = "172.16.100.53";

    .port = "80";

}

import directors;

sub vcl_init {

    new hash = directors.hash();

    hash.add_backend(appServer1, 1.0);

    hash.add_backend(appServer2, 1.0);

}

sub vcl_recv {

    set req.backend_hint = hash.backend(req.http.X-Forwarded-For);

    ......

}

说明：当varnish前端用nginx做反代时，在nginx中需要做如下配置：

            proxy_pass http://varnish_server_ip:port;

            proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;