Open Stack

> I'm not saying this is the most easy way to do things, in fact, I'm
> suggesting - as Eric did - to replace current backdoor_port with
> enable_backdoor and always pick a random port. In addition, we could
> also support linux sockets.
> 


I actually question the value of having this accessible via TCP at all, but I'm not opposed to it existing. If it does, then yes, I prefer the random port + logging approach.

Also, I'm wondering: if the common use-case will be to launch a service explicitly for debugging purposes (and not to run a 24/7 backdoor), then why not support a command-line argument (--debug or similar) that would disable forking and would put an eventlet backdoor on STDIN / STDOUT?

Regards,
Eric Windisch