Segmentation fault in special nginx mode

[eendru@…]

Use config in attachment
I think, that keyword - master_process off;

Run nginx, send some requests via curl or something else. After, send SIGHUP to nginx, send again some requests, you will recive something like

* Rebuilt URL to: 127.0.0.1:9999/

• Trying 127.0.0.1...
• Connected to 127.0.0.1 (127.0.0.1) port 9999 (#0)

  GET / HTTP/1.1
  Host: 127.0.0.1:9999
  User-Agent: curl/7.47.0
  Accept: */*
  

  2017/04/18 13:27:42 [alert] 10307#0: 65536 worker_connections are not enough
  Connection #0 to host 127.0.0.1 left intact
  

And after send another SIGHUP, and you catch

Program received signal SIGSEGV, Segmentation fault.
0x00000000004105bb in ngx_clean_old_cycles (ev=0x69f280 <ngx_cleaner_event>)

[eendru@…]

nginx.conf

[eendru@…]

As I checked, on new nginx

nginx version: nginx/1.11.13
built by gcc 4.9.4 (Ubuntu 4.9.4-2ubuntu1~14.04.1)
configure arguments: --with-debug

same problem.

• use master_process off;
• send requests
• send sighup
• send requests(responses won't be received)
• send sighup
• send requests

wait less than 20 seconds

Program received signal SIGSEGV, Segmentation fault.
0x000000000041cbda in ngx_clean_old_cycles (ev=0x6bdb80 <ngx_cleaner_event>) at src/core/ngx_cycle.c:1303
1303	            if (cycle[i]->connections[n].fd != (ngx_socket_t) -1) {

[eendru@…]

Is duplicate, sorry

[Maxim Dounin]

Duplicate of #945.