Nginx crash frequency 连续崩溃问题

[lystormenvoy]

Sometimes the worker process would crash continuously in one minute.
有时候Nginx工作进程会短时间内连续崩溃。

这个在多个地方遇到过（这次抓到了core dumped，预计之前其他地方遇到的同样问题也是这个导致的）

gdb bt print:

#0  ngx_rtmp_prepare_merge_frame (s=0x44ab2b0) at ../nginx-rtmp-module/ngx_rtmp_shared_module.c:104
#1  0x0000000000531199 in ngx_rtmp_send (wev=0x7fd6974ed6d0) at ../nginx-rtmp-module/ngx_rtmp_handler.c:728
#2  0x0000000000482a7b in ngx_event_process_posted (cycle=0x372c1d0, posted=0x23a62b0 <ngx_processes+144>)
    at src/event/ngx_event_posted.c:34
#3  0x00000000004827a6 in ngx_process_events_and_timers (cycle=0x372c1d0) at src/event/ngx_event.c:263
#4  0x000000000048ac23 in ngx_worker_process_cycle (cycle=0x372c1d0, data=<optimized out>) at src/os/unix/ngx_process_cycle.c:721
#5  0x0000000000489310 in ngx_spawn_process (cycle=0x372c1d0, proc=0x48ab90 <ngx_worker_process_cycle>, data=0x0, 
    name=0x74f23d "worker process", respawn=0) at src/os/unix/ngx_process.c:199
#6  0x000000000048b6ab in ngx_reap_children (cycle=0x372c1d0) at src/os/unix/ngx_process_cycle.c:598
#7  ngx_master_process_cycle (cycle=0x372c1d0) at src/os/unix/ngx_process_cycle.c:174
#8  0x0000000000462f1c in main (argc=<optimized out>, argv=<optimized out>) at src/core/nginx.c:383

source code at ../nginx-rtmp-module/ngx_rtmp_shared_module.c:104 :

cacf = ngx_rtmp_get_module_app_conf(s, ngx_rtmp_core_module);

    if (cacf == NULL){
        return NGX_OK;
    }
    ngx_rtmp_free_merge_frame(s);

    ln = &s->out_chain;

    #line 104 ↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓:
    for (n = 0; n < cacf->merge_frame && s->out_pos != s->out_last; ++n) {