cppcms fixes (#8349)

* Update nginx config

* another try?

* try again

* again!

* revert most unnecessary changes

* fix cppcms postgres

frameworks/C++/cppcms/config-nginx-mysql.json

@@ -4,7 +4,8 @@
     },
     "service": {
         "api": "fastcgi",
-        "socket": "/var/tmp/cppcms.sock"
+        "socket": "/var/tmp/cppcms.sock",
+        "backlog": 65535
     },
     "http": {
         "script": "/"

frameworks/C++/cppcms/config-nginx-postgresql.json

@@ -4,7 +4,8 @@
     },
     "service": {
         "api": "fastcgi",
-        "socket": "/var/tmp/cppcms.sock"
+        "socket": "/var/tmp/cppcms.sock",
+        "backlog": 65535
     },
     "http": {
         "script": "/"

frameworks/C++/cppcms/nginx.conf

@@ -7,32 +7,23 @@ error_log stderr error;
 events {
     # This needed to be increased because the nginx error log said so.
     # http://nginx.org/en/docs/ngx_core_module.html#worker_connections
-    worker_connections  65535;
+    worker_connections 65535;
     multi_accept on;
 }
 
 http {
     default_type  application/octet-stream;
     client_body_temp_path      /tmp;
-
-    # turn off request logging for performance
     access_log off;
+    server_tokens off;
 
-    # I think these only options affect static file serving
-    sendfile        on;
-    tcp_nopush      on;
+    sendfile on;
+    tcp_nopush on;
 
-    # Allow many HTTP Keep-Alive requests in a single TCP connection before
-    # closing it (the default is 100). This will minimize the total number
-    # of TCP connections opened/closed. The problem is that this may cause
-    # some worker processes to be handling too connections relative to the
-    # other workers based on an initial imbalance, so this is disabled for
-    # now.
+    keepalive_timeout 65;
+    keepalive_disable none;
     keepalive_requests 1000;
 
-    #keepalive_timeout  0;
-    keepalive_timeout  65;
-
     server {
         # For information on deferred, see:
         # http://nginx.org/en/docs/http/ngx_http_core_module.html#listen