Re: start-stop-daemon.... for the love of GOD! Why?
On Mon, Dec 12, 2005 at 05:22:27PM +0200, Maxim Vexler wrote:
> On 12/12/05, Andrew Cady <email@example.com> wrote:
> > You could try putting a sulogin immediately before the mythbackend
> > script in your runlevel then running mythbackend manually. If it
> > works you've ruled out ordering and should check for environment
> > problems by stopping mythbackend, putting it last in the runlevel
> > and proceeding to boot; if that works you've ruled out environment
> > and should investigate the race condition.
> > A more direct approach may be to get mythbackend to tell you why
> > it's failing.
> I'm having similar problems to the original poster, but with a
> different application. Could you please clarify your way of debugging
> the subject.
Are you sure it's that similar? The original poster claimed the
application started when envoked manually, but not from rc, although it
was definitely envoked from rc -- a very unusual symptom.
> I mean for the sulogin issue, what should be checked after executing
> the command?
I meant just envoking the script: /etc/init.d/mythbackend start
> In what state (that is relevant for debugging) the system should be
> now, is the environment kept when doing sulogin ?
Actually, the same environment as the rc scripts. So in fact the
method I outlined will not distinguish between environment problems
and ordering if the former is the issue; you would have to use su - to
get that right. Note that an environment problem is quite unlikely,
especially if you are using a different application.
Anyway this sulogin stuff is a bit nasty, although it could save you
from some reboots. You can just put the script last in the runlevel,
reboot, and then carry on from there.