清漆,是什么导致了 hit-for-pass?

Varnish, What is causing hit-for-pass?

这是清漆日志(已编辑 - 已扩展以提供更多信息):

*   << Request  >> 172651    
-   Begin          req 172650 rxreq
-   Timestamp      Start: 1455791132.581465 0.000000 0.000000
-   Timestamp      Req: 1455791132.581465 0.000000 0.000000
-   ReqStart       127.0.0.1 47991
-   ReqMethod      GET
-   ReqURL         /
-   ReqProtocol    HTTP/1.1
-   ReqHeader      X-Real-IP: x.x.x.x
-   ReqHeader      X-Forwarded-For: x.x.x.1, x.x.x.2
-   ReqHeader      X-Forwarded-Proto: https
-   ReqHeader      X-Forwarded-Port: 443
-   ReqHeader      Host: mydomain.com
-   ReqHeader      User-Agent: Go-http-client/1.1  (//benchmark app, but regular browser client happens the same
-   ReqHeader      Cf-Connecting-Ip: x.x.x.x
-   ReqHeader      Cf-Ipcountry: AND
-   ReqHeader      Cf-Origin-Https: off
-   ReqHeader      Cf-Ray: 27690752ab232e21-NRT
-   ReqHeader      Cf-Visitor: {"scheme":"https"}
-   ReqUnset       X-Forwarded-For: x.x.x.1, x.x.x.2
-   ReqHeader      X-Forwarded-For: x.x.x.ext, x.x.x.1, x.x.x.2
-   VCL_call       RECV
-   VCL_return     hash
-   VCL_call       HASH
-   VCL_return     lookup
-   Debug          "XXXX HIT-FOR-PASS"
-   HitPass        393349
-   VCL_call       PASS
-   VCL_return     fetch
-   Link           bereq 172652 pass
-   Timestamp      Fetch: 1455791132.581733 0.000268 0.000268
-   RespProtocol   HTTP/1.1
-   RespStatus     200
-   RespReason     OK
-   RespHeader     Server: nginx
-   RespHeader     Date: Thu, 18 Feb 2016 10:25:32 GMT
-   RespHeader     Content-Type: text/html; charset=utf-8
-   RespHeader     Vary: Accept-Encoding, Accept-Encoding
-   RespHeader     Etag: "1455788472-0"
-   RespHeader     Content-Language: ja
-   RespHeader     Cache-Control: no-cache
-   RespHeader     Last-Modified: Thu, 18 Feb 2016 09:41:12 GMT
-   RespHeader     X-Content-Options: nosniff
-   RespHeader     X-Micro-Cache: HIT
-   RespHeader     X-Varnish: 172651
-   RespHeader     Age: 0
-   RespHeader     Via: 1.1 varnish-v4
-   VCL_call       DELIVER
-   VCL_return     deliver
-   Timestamp      Process: 1455791132.581744 0.000279 0.000011
-   RespHeader     Accept-Ranges: bytes
-   RespHeader     Transfer-Encoding: chunked
-   Debug          "RES_MODE 8"
-   RespHeader     Connection: keep-alive
-   Timestamp      Resp: 1455791132.581794 0.000329 0.000049
-   ReqAcct        336 0 336 436 57069 57505
-   End            

我读到这是在设置 set-cookie header 或 cache-control: max-age=0 时创建的。

好吧,我在我的子 vcl_backend_response 中有这个试图避免传球命中。这很粗鲁,但我不知道后端是什么导致了这次通过:

sub vcl_backend_response {
  unset beresp.http.set-cookie;
  unset beresp.http.expires;
  set beresp.http.Cache-Control = "max-age=6000";
}

我正在摆脱 Expire,因为它被设置为时间戳 0。所以我认为可能是这样。然后我试着玩 Cache-control 但还是一样。我即将以这种速度取消所有 headers...

毕竟是304响应,为什么不缓存呢?

更多信息:后端是一个 Drupal 应用程序。

编辑 1:关于 vcl_recv

是的,我确实有一些东西,我正在返回某些路径的通行证(不是我正在检查的路径),我正在取消设置文件的 cookie,我正在取消所有非 drupal 的 cookie session对于其余的。它或多或少是 copy/paste drupal 默认子 vcl_recv。如果你想要我可以在这里复制。

顺便说一句,我更改了其他内容,现在我在响应 200 中也通过了测试(但是当我使用基准测试工具时)。有人删除了说可能是因为响应 304 的答案。我正在检查那个,但答案不见了:(

好的,这是 vcl_recv:

sub vcl_recv {
 if (req.url ~ "^/status\.php$" ||
    req.url ~ "^/admin/.*$" ||
     [...]
     req.url ~ "^.*\.js.*$" ) 
    {
       return (pass);
    }

  if (req.url ~ "^/admin/content/backup_migrate/export") {
    return (pipe);
  }

  ## COOKIES ##
  if (req.url ~ "(?i)\.(css|js|jpg|jpeg|gif|png|ico|svg|map)(\?.*)?$") {
    unset req.http.Cookie;
  }
  if (req.http.Cookie) {
    set req.http.Cookie = ";" + req.http.Cookie;
    set req.http.Cookie = regsuball(req.http.Cookie, "; +", ";"); 
    set req.http.Cookie = regsuball(req.http.Cookie, ";(SESS[a-z0-9]+|SSESS[a-z0-9]+|NO_CACHE)=", "; =");
    set req.http.Cookie = regsuball(req.http.Cookie, ";[^ ][^;]*", "");
    set req.http.Cookie = regsuball(req.http.Cookie, "^[; ]+|[; ]+$", "");
   if (req.http.Cookie == "") {
        unset req.http.Cookie;
    }
    # if any of our cookies of interest are still there, we disable caching and pass the request
    else {
        return (pass);
    }
 }
}

它几乎总是在所有 http. 304 个响应和 200 个响应。我已经编辑了上面的 varnishlog 以显示完整示例以提供更多信息。

编辑: 所以这是在显示命中通过之前的请求日志。肯定是传球命中率的罪魁祸首:

-   ReqHeader      X-Forwarded-For: 127.0.0.1
-   VCL_return     hash
-   VCL_call       HASH
-   VCL_return     lookup
-   VCL_call       MISS
-   VCL_return     fetch
-   Link           bereq 3 fetch
-   Timestamp      Fetch: 1456731597.992472 0.099948 0.099948
-   RespProtocol   HTTP/1.1
-   RespStatus     200
-   RespReason     OK
-   RespHeader     Date: Mon, 29 Feb 2016 07:39:57 GMT
-   RespHeader     Content-Type: text/html; charset=utf-8
-   RespHeader     Content-Encoding: gzip
-   RespHeader     Vary: Accept-Encoding
-   RespHeader     Etag: "1456730796-0"
-   RespHeader     Content-Language: ja
-   RespHeader     Cache-Control: no-cache
-   RespHeader     Last-Modified: Mon, 29 Feb 2016 07:26:36 GMT
-   RespHeader     Expires: Thu, 01 Jan 1970 00:00:01 GMT
-   RespHeader     X-Content-Options: nosniff
-   RespHeader     X-Micro-Cache: EXPIRED
-   RespHeader     X-Varnish: 2
-   RespHeader     Age: 0
-   RespHeader     Via: 1.1 varnish-v4
-   RespUnset      Etag: "1456730796-0"
-   RespHeader     ETag: W/"1456730796-0"
-   VCL_call       DELIVER
-   RespHeader     X-Cache: MISS
-   RespHeader     X-Cache-Hits: 0
-   RespUnset      X-Varnish: 2
-   RespUnset      Via: 1.1 varnish-v4
-   VCL_return     deliver
-   Timestamp      Process: 1456731597.992517 0.099993 0.000044
-   RespUnset      Content-Encoding: gzip
-   RespHeader     Accept-Ranges: bytes
-   RespHeader     Content-Length: 53980
-   Debug          "RES_MODE 42"
-   RespHeader     Connection: keep-alive
-   Gzip           U D - 12752 53980 80 101936 101946
-   Timestamp      Resp: 1456731597.992817 0.100293 0.000300
-   ReqAcct        746 0 746 437 53980 54417

您的问题可能恰恰是 304 响应:

清漆 3 书说:

Only the following status codes will be cached by default:

200: OK
203: Non-Authoritative Information
300: Multiple Choices
301: Moved Permanently
302: Moved Temporarily
307: Temporary Redirect
410: Gone
404: Not Found

You can still cache other status codes, but you will have to set the beresp.ttl to a positive value in vcl_fetch yourself.

http://book.varnish-software.com/3.0/VCL_Basics.html

令我惊讶的是,Varnish 4 书中还提到“304:未修改”是可缓存的。 http://book.varnish-software.com/4.0/chapters/VCL_Basics.html

这对我来说没有意义,因为如果 Varnish 正在缓存条件请求的 304 响应,那么对同一资源的每个非条件请求也应该得到没有响应主体的 304 响应,这似乎毫无用处。

这个问题太长了,恐怕对其他人也没有用。

感谢其他答案给出的 link 我看到了这个(vcl_backend_response 的默认实现):

sub vcl_backend_response {
if (beresp.ttl <= 0s ||
    beresp.http.Set-Cookie ||
    beresp.http.Surrogate-control ~ "no-store" ||
    (!beresp.http.Surrogate-Control &&
    beresp.http.Cache-Control ~ "no-cache|no-store|private") ||
    beresp.http.Vary == "*") {
    /*
     * Mark as "Hit-For-Pass" for the next 2 minutes
     */
    set beresp.ttl = 120s;
    set beresp.uncacheable = true;
}
return (deliver);
}

这就是所有命中传球的条件。多亏了它,我可以看到导致通过命中的原因正是这个 beresp.ttl <=0。

那么哪个值有 beres.ttl?在同一个 link 上,我们读到: beresp.ttl 被初始化为它在

中找到的第一个值
The s-maxage variable in the Cache-Control response header field
The max-age variable in the Cache-Control response header field
The Expires response header field
The default_ttl parameter.

修改vcl_backend_response中的header没关系,因为ttl已经设置好了!

很好,看看我的 headers 我可以从后端看到:

 -  RespHeader Expires: Thu, 01 Jan 1970 00:00:01 GMT

这是导致传球成功的原因。这就是我的问题的答案。现在,为什么我有这个 header 我不知道,那将是一个不同的问题。 (我用过期 30d 修改了 nginx;到所有位置都没有结果)