I am going to use nginx in production mode.
I normally used nginx -s reload when nginx.config requies to be updated.
But if I call this command several times, I could see strange issue.
As you can see, some nginx processes are in shutting down status.
And also this is error.log file.
I assume that only I will see only signal process started message.
What’s the next merge message? why this is happening?
I want to discuss this kind of issue more widely in our community.
Thanks
Hello everyone
I finllay find the reason for current issue.
I used timer_event in my nginx sub module.
because nginx won’t let a worker exit while it still has any non-cancelable timers pending.
Actually, nginx is waiting for that timer stops
When you run nginx -s reload, old workers go into graceful shutdown: they stop accepting new work and then wait until all connections are closed and the timer queue is empty. Since you add a periodic timer in init_process(), the old worker keeps at least one future timer scheduled. Unless that timer fires and your handler returns without re-arming, nginx keeps the worker in “shutting down”.
So the process is working strangely
