[rabbitmq-discuss] connecting to Rabbit's node

Tue Nov 8 11:13:52 GMT 2011

I understand that, but isn't it against the erlang's philosophy of
creating lightweight processes on demand and let them die after they are
finished?

I am not an erlang coder at all, I just try to get into RabbitMQ
structure to get knowledge about it. However there's no 'RabbitMQ
Internals' documented so I have to study *.erl files :)

thank you

marek

-----Original Message-----
From: Matthias Radestock [mailto:matthias at rabbitmq.com] 
Sent: Tuesday, November 08, 2011 11:26
To: Cermak, Marek
Cc: rabbitmq-discuss at lists.rabbitmq.com
Subject: Re: [rabbitmq-discuss] connecting to Rabbit's node

Marek,

On 08/11/11 10:22, Cermak, Marek wrote:
> looking on the structure of running processes, there is quite a lot of
> delegate supervisors in hibernate state and same for
> rabbit_tcp_client_sup. Why is that?

Most rabbit erlang processes will hibernate when they have been idle for

a while, in order to conserve memory.

Matthias.