init problem with Intrepid

Derek Broughton news at pointerstop.ca
Thu Oct 30 23:25:35 UTC 2008


Mario Vukelic wrote:

> On Thu, 2008-10-30 at 16:22 -0400, Rashkae wrote:
>> Why not?
>> 
>> I know that gdm stop is the proper way to do it, no argument there,, but
>> so long as you refrain from using the -9, killall just sends a nice
>> gentle Sigterm, which is a perfectly legit way of stopping a process
> 
> To stop a process yes, but who knows what else the stop script does that
> the process knows nothing about. I am too lazy to check, but anyone who
> is interested can and draw their conclusions :)

Well, I wasn't too lazy, since I was the first to suggest that stopping gdm
that way wasn't any big deal, and figured I'd better be sure of what I was
saying.  And really, there is _nothing_ that should be harmed by stopping
any process with SIGTERM.  Yes, it's sloppy, and _I_ would never do it if
there was another option, but let's not go claiming that the sky is
falling - SIGTERM _is_ the proper way to stop a process, and is the way
that start-stop-daemon is going to stop it in the end, anyway.

With most daemons (because of the use of start-stop-daemon) the major thing
that is going to happen if you use kill to stop the relevant program is
that you will get a message on restart about a pre-existing pid file, which
it will remove (it checks whether there really is a process associated with
that pid first).
-- 
derek





More information about the ubuntu-users mailing list