Opened 10 years ago

Closed 10 years ago

#687 closed defect (wontfix)

Keep-alive with fastcgi not working with HEAD

Reported by: Michael Bonfils Owned by:
Priority: minor Milestone:
Component: nginx-core Version: 1.6.x
Keywords: fastcgi keepalive head Cc: patrick.dos.santos@…
uname -a: Linux ubuntu14-vm 3.13.0-32-generic #57-Ubuntu SMP Tue Jul 15 03:51:08 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
nginx -V: nginx version: nginx/1.7.8
built by gcc 4.8.2 (Ubuntu 4.8.2-19ubuntu1)
TLS SNI support enabled
configure arguments: --prefix=/etc/nginx --sbin-path=/usr/sbin/nginx --conf-path=/etc/nginx/nginx.conf --error-log-path=/var/log/nginx/error.log --http-log-path=/var/log/nginx/access.log --pid-path=/var/run/nginx.pid --lock-path=/var/run/nginx.lock --http-client-body-temp-path=/var/cache/nginx/client_temp --http-proxy-temp-path=/var/cache/nginx/proxy_temp --http-fastcgi-temp-path=/var/cache/nginx/fastcgi_temp --http-uwsgi-temp-path=/var/cache/nginx/uwsgi_temp --http-scgi-temp-path=/var/cache/nginx/scgi_temp --user=nginx --group=nginx --with-http_ssl_module --with-http_realip_module --with-http_addition_module --with-http_sub_module --with-http_dav_module --with-http_flv_module --with-http_mp4_module --with-http_gunzip_module --with-http_gzip_static_module --with-http_random_index_module --with-http_secure_link_module --with-http_stub_status_module --with-http_auth_request_module --with-mail --with-mail_ssl_module --with-file-aio --with-http_spdy_module --with-cc-opt='-g -O2 -fstack-protector --param=ssp-buffer-size=4 -Wformat -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2' --with-ld-opt='-Wl,-Bsymbolic-functions -Wl,-z,relro -Wl,--as-needed' --with-ipv6

Description

During our test, we found that while issue HEAD requests to nginx on fastcgi backend, we found that it doesn't reuse TCP connection to fastcgi backend for HEAD requests.

How to reproduce :

$ netstat -tupan | grep 1039
tcp        0      0 0.0.0.0:1039            0.0.0.0:*               LISTEN      23795/dewpoint.fcgi

$ curl -I http://localhost http://localhost/README http://localhost/README http://localhost/README http://localhost/README http://localhost/README http://localhost/README http://localhost http://localhost http://localhost/README http://localhost/README http://localhost

$ netstat -tupan | grep 1039
tcp        0      0 0.0.0.0:1039            0.0.0.0:*               LISTEN      23795/dewpoint.fcgi
tcp        0      0 127.0.0.1:39797         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39802         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39801         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39790         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39798         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39805         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39796         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39791         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39794         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39800         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39795         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39799         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39804         127.0.0.1:1039          TIME_WAIT   -               
tcp        0      0 127.0.0.1:39803         127.0.0.1:1039          TIME_WAIT   -               

In error.log in debug mode ::

2014/12/23 15:33:50 [info] 27766#0: *1 client 127.0.0.1 closed keepalive connection

With GET requests, keep alive between nginx and fastcgi backend is ok

$ netstat -tupan | grep 1039
tcp        0      0 0.0.0.0:1039            0.0.0.0:*               LISTEN      23795/dewpoint.fcgi

$ curl -v http://localhost http://localhost/README http://localhost/README http://localhost/README http://localhost/README http://localhost/README http://localhost/README http://localhost http://localhost http://localhost/README http://localhost/README http://localhost

$ netstat -tupan | grep 1039
tcp        0      0 0.0.0.0:1039            0.0.0.0:*               LISTEN      23795/dewpoint.fcgi
tcp        0   1024 127.0.0.1:1039          127.0.0.1:39808         ESTABLISHED 23795/dewpoint.fcgi
tcp        0      0 127.0.0.1:39808         127.0.0.1:1039          ESTABLISHED 27695/nginx: worker

Issue was reproduced in 1.6.2 and 1.7.8

Attachments (2)

dump_head.pcap (19.1 KB ) - added by Michael Bonfils 10 years ago.
Dump of network between nginx and fastcgi backend
nginx.conf (1.5 KB ) - added by Michael Bonfils 10 years ago.
Nginx configuration

Download all attachments as: .zip

Change History (3)

by Michael Bonfils, 10 years ago

Attachment: dump_head.pcap added

Dump of network between nginx and fastcgi backend

by Michael Bonfils, 10 years ago

Attachment: nginx.conf added

Nginx configuration

comment:1 by Maxim Dounin, 10 years ago

Resolution: wontfix
Status: newclosed

Yes, this is expected behaviour. In case of a HEAD request (and in other cases when response body isn't returned) we do not read the response body from a backend (if any), and close the connection instead.

Note: See TracTickets for help on using tickets.