我正在尝试将 Nginx 配置为缓存反向代理。如果有必要的话,原始服务器是 Apache,它托管一个 WordPress 实例。
反向代理功能按预期工作,但缓存似乎不起作用。如果我连续两次获取相同的静态资源,我就会得到x-proxy-cache: MISS
两次。
assodigitale.it 是域名,138.201.87.123 是原始服务器 IP 地址,138.201.87.124 是 Nginx 代理 IP 地址。
原始服务器似乎回复允许代理缓存资源:
$ curl --connect-to ::138.201.87.123:443 --http2 -I https://assodigitale.it/wp-content/uploads/2018/03/aereo.jpg
HTTP/2 200
date: Sun, 11 Mar 2018 20:59:39 GMT
server: Apache/2.4.25 (Debian)
content-length: 32989
strict-transport-security: max-age=31536000; includeSubdomains; preload
last-modified: Wed, 07 Mar 2018 09:34:41 GMT
etag: "80dd-566cf44ca2952"
accept-ranges: bytes
vary: Accept-Encoding
cache-control: max-age=1209600, public
x-content-type-options: nosniff
content-type: image/jpeg
正如预期的那样,对代理服务器的第一个请求导致了 MISS:
$ curl --connect-to ::138.201.87.124:443 --http2 -I https://assodigitale.it/wp-content/uploads/2018/03/aereo.jpg
HTTP/2 200
server: nginx/1.13.9
date: Sun, 11 Mar 2018 21:04:00 GMT
content-type: image/jpeg
content-length: 32989
strict-transport-security: max-age=31536000; includeSubdomains; preload
last-modified: Wed, 07 Mar 2018 09:34:41 GMT
etag: "80dd-566cf44ca2952"
vary: Accept-Encoding
cache-control: max-age=1209600, public
x-content-type-options: nosniff
x-proxy-cache: MISS
strict-transport-security: max-age=4838400; includeSubDomains; preload
accept-ranges: bytes
对 Nginx 代理的第二个请求应该导致 HIT,但却导致了另一个 MISS:
$ curl --connect-to ::138.201.87.124:443 --http2 -I https://assodigitale.it/wp-content/uploads/2018/03/aereo.jpg
HTTP/2 200
server: nginx/1.13.9
date: Sun, 11 Mar 2018 21:05:52 GMT
content-type: image/jpeg
content-length: 32989
strict-transport-security: max-age=31536000; includeSubdomains; preload
last-modified: Wed, 07 Mar 2018 09:34:41 GMT
etag: "80dd-566cf44ca2952"
vary: Accept-Encoding
cache-control: max-age=1209600, public
x-content-type-options: nosniff
x-proxy-cache: MISS
strict-transport-security: max-age=4838400; includeSubDomains; preload
accept-ranges: bytes
以下是我的 nginx 配置的相关部分:
proxy_cache_path /srv/cache/nginx levels=1:2 keys_zone=revproxy:2000m inactive=2880m use_temp_path=off;
proxy_cache_key "$scheme$request_method$host$request_uri";
proxy_cache_methods GET HEAD;
proxy_cache_valid any 1m;
proxy_cache_valid 200 1440m;
server {
listen 443 ssl http2;
ssl on;
server_name assodigitale.it;
ssl_certificate /etc/letsencrypt/live/assodigitale.it/fullchain.pem;
ssl_certificate_key /etc/letsencrypt/live/assodigitale.it/privkey.pem;
ssl_session_timeout 1d;
ssl_session_cache shared:SSL:50m;
ssl_session_tickets off;
ssl_protocols TLSv1.2;
ssl_ciphers 'ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-RSA-CHACHA20-POLY1305:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA256';
ssl_prefer_server_ciphers on;
location / {
proxy_cache revproxy;
add_header X-Proxy-Cache $upstream_cache_status;
add_header Strict-Transport-Security "max-age=4838400; includeSubDomains; preload";
proxy_pass https://138.201.87.123;
proxy_cache_use_stale updating error timeout http_500 http_502 http_503 http_504;
proxy_cache_bypass $http_x_forceflushcacheurl;
proxy_cache_lock on;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_hide_header Upgrade;
proxy_buffering off;
proxy_connect_timeout 600;
proxy_send_timeout 600;
proxy_read_timeout 600;
send_timeout 600;
proxy_ignore_headers Set-Cookie;
http2_push_preload on;
client_max_body_size 64M;
}
}
该/srv/cache/nginx
目录具有 755 个权限和www-data
所有者,并且 Nginx 以 的身份运行www-data
。事实上,Nginx 确实在其中写入了它的文件夹,即0 1 2 3 4 5 6 7 8 9 a b c d e f
,但对于一个相当大且流量远超普通流量的网站来说,现在占用的总空间为 344Kb。
尝试curl
上述相同的命令,但使用页面而不是图像,会产生相同的结果,总是 MISS。
为什么Nginx拒绝缓存资源?
答案1
您应该设置 proxy_buffering on
,否则 nginx 将不会缓存响应!
官方文档说:
当禁用缓冲时,响应会在收到后立即同步传递给客户端。nginxwill not try to read the whole response from the proxied server.
一次可以从服务器接收的数据的最大大小由 proxy_buffer_size 指令设置。
答案2
我从我运行的另一个类似的 Nginx 代理复制了配置,并将其调整到适合网站,现在可以正常工作。
这是我当前正在使用的配置:
proxy_cache_path /srv/cache/nginx levels=1:2 keys_zone=revproxy:2000m inactive=2880m use_temp_path=off;
proxy_cache_key "$scheme$request_method$host$request_uri";
proxy_cache_methods GET HEAD;
proxy_cache_valid any 1m;
proxy_cache_valid 200 1440m;
server {
listen 443 ssl http2;
ssl on;
server_name assodigitale.it;
ssl_certificate /etc/letsencrypt/live/assodigitale.it/fullchain.pem;
ssl_certificate_key /etc/letsencrypt/live/assodigitale.it/privkey.pem;
ssl_protocols SSLv3 TLSv1 TLSv1.1 TLSv1.2;
ssl_ciphers RC4:HIGH:!aNULL:!MD5;
ssl_prefer_server_ciphers on;
keepalive_timeout 60;
ssl_session_timeout 10m;
location / {
proxy_cache revproxy;
add_header X-Proxy-Cache $upstream_cache_status;
proxy_pass https://138.201.87.123;
proxy_next_upstream error timeout invalid_header http_500 http_502 http_503 http_504;
proxy_cache_use_stale updating error timeout http_500 http_502 http_503 http_504;
proxy_cache_lock on;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_hide_header Upgrade;
proxy_buffering on;
proxy_connect_timeout 600;
proxy_send_timeout 600;
proxy_read_timeout 600;
send_timeout 600;
client_max_body_size 64M;
}
}
虽然这可以作为一个答案,因为它解决了问题,但我不明白它为什么有效(或者更准确地说,为什么以前的配置不起作用),所以我不会接受我自己的答案。
也许有人可以发现使缓存起作用的两种配置中的具体差异:这将是一个可以接受的答案。
答案3
我们遇到了类似的问题。我们使用 nginx 作为 s3 bucket 的代理缓存(因此我们可以支持 IP 白名单)。我们注意到,尽管我们的代理在 https 上提供服务,但我们并没有从 https 端点获取数据。似乎 nginx 不喜欢这样。一旦我们从 切换http
到https
,就没有问题了。