simple_one_for_one start_child() returns already_started

Question

I have a supervisor which should start simple_one_for_one workers. When I call start_child() for the first time, everything goes excellent. But, when I do it the second time, I get {error,{already_started,<0.71.0>}}. Why would simple_one_for_one supervisor return me a already_started? What am I missing?

Here is the code: supervisor, worker.

Bisulfate

You're going to be awfully confused with your gen server handling messages by spawning a child to send a message to itself and then blocking while waiting to receive that message. – Wisnicki

Why? I do the work in separate child to prevent gen_server from crashing. Is that a bad thing to do? Advice would be much appreciated :-) – Bisulfate

Would it be a catastrophe if the gen_server crashed? It will be restarted by the supervisor. If you are worried about errors then wrap the call instead with a try ... catch ... end. While processes are cheap there is no gain in creating them unnecessarily and here there is no concurrency. – Libido

Can you post the code here? If that link ever breaks, this question will be useless. – Fateful

Answer 1

you are registering a (local) name for your gen_server. once you start one, you can not start another one with the same name.

if you use gen_server:start_link/3 instead, removing the first argument from your current gen_server:start_link/4 call, you should be able to start up more than one.

Thilde