Re: reboot proceedure
According to Russell Coker:
> If I have a number of programs specified in /etc/inittab when in which order
> will they receive SIGTERM during suchdown?
At the same time. Sysvinit scans the entries and sends SIGTERM
to all of them in a tight loop.
> A portslave user has the following in his inittab:
> fr:12345:respawn:/usr/local/sbin/radiusd -f -s &> /dev/null
> s0:2345:respawn:/usr/local/sbin/portslave 6
> s8:2345:respawn:/usr/local/sbin/portslave 9
> s9:2345:respawn:/usr/local/sbin/portslave 10
> His problem is that the radiusd gets killed before portslave, so the end of
> call data can't be sent to radiusd from portslave and he misses billing
Run radiusd from /etc/init.d/radiusd. Then the radius server will
be killed after the switch to the new runlevel, not before.
At least Cistron Radius doesn't need to be run from inittab,
it simply doesn't crash. I have radiusd processes that have
run for months without any memory leak or crash.
"Only two things are infinite, the universe and human stupidity,
and I'm not sure about the former" -- Albert Einstein.