http request has no response, the configure with 'master_process off and listen reuseport'

[bullerdu@…]

When the following configuration is used, no response is received when sending an http request.

user  root;

master_process              off;
worker_processes            4;
worker_cpu_affinity         auto;

error_log  logs/error.log  debug;

pid        logs/nginx.pid;

#load_module modules/ngx_http_image_filter_module.so;

events {
    use                 epoll;
    accept_mutex        off;
    worker_connections  102400;
}

http {
    include       mime.types;
    default_type  application/octet-stream;

    #log_format  main  '$remote_addr - $remote_user [$time_local] "$request" '
    #                  '$status $body_bytes_sent "$http_referer" '
    #                  '"$http_user_agent" "$http_x_forwarded_for"';

    #access_log  logs/access.log  main;

    sendfile        on;
    #tcp_nopush     on;

    #keepalive_timeout  0;
    keepalive_timeout  65;

    #gzip  on;

    server {
        listen 80 reuseport;

        location / {
            return 200 "ok";
        }
    }
}

the result with no response:

$curl -sv 'localhost:80/'
* About to connect() to localhost port 80 (#0)
*   Trying 127.0.0.1...
* Connected to localhost (127.0.0.1) port 80 (#0)
> GET / HTTP/1.1
> User-Agent: curl/7.29.0
> Host: localhost
> Accept: */*
>
^C

The fault occurs because worker_processes enable multiple processes with 'listen 80 reuseport;' in the master off mode.
My question is, since master_process off conflicts with worker_processes, why does nginx not disable worker_processes to configure multiple processes in master off mode

A patch that solves this configuration problem is listed below.

diff --git a/src/core/ngx_connection.c b/src/core/ngx_connection.c
index fe729a78..ab210dd8 100644
--- a/src/core/ngx_connection.c
+++ b/src/core/ngx_connection.c
@@ -112,6 +112,10 @@ ngx_clone_listening(ngx_cycle_t *cycle, ngx_listening_t *ls)

     ccf = (ngx_core_conf_t *) ngx_get_conf(cycle->conf_ctx, ngx_core_module);

+    if (ccf->master == 0) {
+        return NGX_OK;
+    }
+
     for (n = 1; n < ccf->worker_processes; n++) {

         /* create a socket for each worker process */

[Maxim Dounin]

My question is, since master_process off conflicts with worker_processes, why does nginx not disable worker_processes to configure multiple processes in master off mode

Forcing worker_processes to be 1 in case of master_process off; might solve some of the minor issues like this one, though generally is not enough to work without master process.

Note well that the master_process off; mode is for nginx development, and implies understanding of various limitations, see ​docs. It is expected that some features simply won't work or will break things, up to segmentation faults in some cases (see ticket #945 for an example).

A patch that solves this configuration problem is listed below.

Thanks for the patch. I've submitted a ​slightly different one for review.

[Maxim Dounin <mdounin@…>]

In 8105:09463dd9c504/nginx:

Disabled cloning of sockets without master process (ticket #2403).

Cloning of listening sockets for each worker process does not make sense
when working without master process, and causes some of the connections
not to be accepted if worker_processes is set to more than one and there
are listening sockets configured with the reuseport flag. Fix is to
disable cloning when master process is disabled.

[Maxim Dounin]

Fix committed, thanks for reporting this.