Question

我的清漆缓存工作正常，除了每隔几个小时就会出现烦人的问题，后端探测器会报告我的4个后端中的3个。因此网站无法打开，显然会报告503。

奇怪的是每次发生这种情况时，我都会去Domain Dossier（http://centralops.net/co/DomainDossier.aspx）并进行＆＃34;服务扫描＆＃34;。扫描将显示503错误，但显示所有标题，只有X-Varnish有一个数字，显然缓存不起作用....

但是在我完成＆＃34;服务扫描＆＃34;之后，我去做了一个＆＃34; varnishadm backend.list＆＃34;，我将100％的时间看到数字计数开始从＆＃34;生病0/10＆＃34;最后＆＃34;健康10/10＆＃34;。就像我说的那样，我通常会有3个病区，所以我做了＃34;服务扫描＆＃34;再次为另外两个生病的域名，他们每个人都成功探测ping并重新开始工作....

任何人都有同样的问题吗？

Apache：监听domain.ip1：8080监听domain.ip2：8080监听domain.ip4：8080（domain2使用与domain3相同的ip）

这是default.vcl：

backend domain1 {
  .host = "domain.ip1";
  .port = "8080";
  .connect_timeout = 600s;
  .saintmode_threshold = 0;
  .first_byte_timeout = 600s;
  .between_bytes_timeout = 600s;
  .probe = {
      .url = "/favicon.ico";
   .interval = 10s;
   .timeout = 3s;
   .window = 10;
   .threshold = 3;
  # .initial = 3;
   .expected_response = 301;
}
}

backend domain2 {
  .host = "domain.ip2";
  .port = "8080";
  .connect_timeout = 600s;
  .first_byte_timeout = 600s;
  .between_bytes_timeout = 600s;
  .probe = {
   .url = "/favicon.ico";
   .interval = 10s;
   .timeout = 3s;
   .window = 10;
   .threshold = 3;
 #  .initial = 3;
   .expected_response = 200;
}
}

backend domain3 {
  .host = "domain.ip3";
  .port = "8080";
  .connect_timeout = 600s;
  .first_byte_timeout = 600s;
  .between_bytes_timeout = 600s;
  .probe = {
      .url = "/favicon.ico";
   .interval = 10s;
   .timeout = 3s;
   .window = 10;
   .threshold = 3;
  # .initial = 3;
   .expected_response = 200;
}   
}
backend domain4 {
  .host = "domain.ip4";
  .port = "8080";
  .connect_timeout = 600s;
  .first_byte_timeout = 600s;
  .between_bytes_timeout = 600s;
  .probe = {
     .url = "/favicon.ico";
   .interval = 10s;
   .timeout = 3s;
   .window = 10;
   .threshold = 3;
  # .initial = 3;
   .expected_response = 200;
}
}


sub vcl_recv {
      if  (req.http.host ~ "^www.domain1.com") {
           set req.http.host = "www.domain1.com";
           set req.backend = domain1;
       return (lookup);
  }
  if  (req.http.host ~ "^www.domain2.com") { 
       set req.http.host = "www.domain2.com";
       set req.backend = domain2;
       return (lookup);
 }
  if  (req.http.host ~ "^www.domain3.com") { 
       set req.http.host = "www.domain3.com";
       set req.backend = domain3;
       return (lookup);
 }
 if  (req.http.host ~ "^www.domain4.com") { 
       set req.http.host = "www.domain4.com";
       set req.backend  = domain4;
       return (lookup);
 }

   }

当后端健康时，探测器10/10，标题如下：

HTTP/1.1 301 Moved Permanently
Server: Apache
Location: http://www.domain1.com/
Cache-Control: max-age=900
Expires: Tue, 24 Jun 2014 07:30:25 GMT
Content-Type: text/html; charset=iso-8859-1
Content-Length: 304
Accept-Ranges: bytes
Date: Tue, 24 Jun 2014 07:15:25 GMT
X-Varnish: 1997501007
Age: 0
Via: 1.1 varnish
Connection: close

OR

HTTP/1.1 301 Moved Permanently
Server: Apache
Location: http://www.domain1.com/
Cache-Control: max-age=900
Expires: Tue, 24 Jun 2014 07:30:25 GMT
Content-Type: text/html; charset=iso-8859-1
Content-Length: 304
Accept-Ranges: bytes
Date: Tue, 24 Jun 2014 07:18:53 GMT
X-Varnish: 1997501047 1997501007
Age: 208
Via: 1.1 varnish
Connection: close

当后端生病时，我使用Domain Dossier获取标题，它将显示如下：

HTTP/1.1 503
Server: Apache
Location: http://www.domain1.com/
Cache-Control: max-age=900
Expires: Tue, 24 Jun 2014 07:30:25 GMT
Content-Type: text/html; charset=iso-8859-1
Content-Length: 304
Accept-Ranges: bytes
Date: Tue, 24 Jun 2014 07:15:25 GMT
X-Varnish: 1997501007
Age: 0
Via: 1.1 varnish
Connection: close

/ etc / sysconfig / varnish文件：

DAEMON_OPTS="-a :80\
             -T localhost:6082 \
             -f /etc/varnish/default.vcl \
             -S /etc/varnish/secret \
             -s malloc,512m \
             -p ping_interval=10 \
             -p thread_pools=2 \
             -p thread_pool_max=8000 \
             -p thread_pool_min=100 \
             -p thread_pool_add_delay=2 \
             -p sess_timeout=60 \
             -p cli_timeout=120  \
"

Varnish Probe再次运行在Domain Dossier的“服务扫描”之后

0 个答案: