Trivial PR, fix shutdown of rc services started with onestart (fwd)

[Dominic Fandrey]

On 14/04/2010 22:18, Doug Barton wrote:
> This issue has come up several times in the past, and previously I've
> been opposed to making changes since I don't regard typing "onestop" an
> overwhelming burden. :)  However there is a reasonable argument to be
> made for having things work when the system is shut down.

Think of this like a laptop user. You've got to leave quickly, so you
press the power button.

10 minutes later you power back on and want to log back into VPN.
Darn, you forgot to log out and now you've got to wait for for another
20 minutes, before your old session times out and you're permitted
back in.

There's a good reason stop_cmds normally do more than 'kill $pid'. And
the results of not executing them can be much more dire than having to
wait for a 30 minutes time out to complete (if you've deal with fuse,
severe data loss can be the result).

Also, do you keep track of what you stared if your laptop is running
for a couple of days non-stop? Did you start that PostgreSQL DB? Did you
start MySQL? Did you start Apache? Tor? Privoxy? ...
Is it really your job to remember?

> I'm not sure the patch in the PR is right (mostly for lack of
> examination or testing), but I'm wondering what people think about the
> general idea of having the shutdown action ignore whether _enable is on?

I'm running with variations of this for more than a year and not once
have I been able to detect unwanted side effects.
However, I can assure you it has spared me a lot of problems.

Regards

-- 
A: Because it fouls the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?