On Tuesday 09 September 2008 10:44, Vladimir Dronnikov wrote:
> >
> > > > and stage 3 is not needed at all - "killall5 -TERM" etc still works,
> > > > no need to talk to init.
> 
> So we do not rely on runsvdir (pid 1) to shut children runsv's down on
> getting HUP?

Well, yes. If TERM is supposed to shut down any program cleanly,
why do we need yet another convention at all?

> > Even killall5 -KILL won't. Process with PID 1 is not killed,
> > even if it is a shell. Kernel will not panic.
> 
> Accidental TERM to runsvdir will cause system to just die in a _very_
> unclean way. Since runsvdir dies suddenly and kernel just panics.

I do not run runsvdir as process with PID 1. Thus when it exits,
kernel does not panic.

> There should be something behind runsvdir to flush the system gracefully --
> so called stage 3.

Why? Are you saying you have programs which do not exit correctly on TERM?
Those programs are broken.

> And if so, we can not just exec runsvdir at the end of 
> init. M.B. patch runsvdir so it could take a command to execute upon exit?

Or execute this command before killall5 -TERM. This way no patching
is required.
--
vda
_______________________________________________
busybox mailing list
[email protected]
http://busybox.net/cgi-bin/mailman/listinfo/busybox

Reply via email to