Re: better make a standard for /etc/*/*_not_to_be_run
On Sun, 10 Aug 2003, Micha? Politowski wrote:
> On Sun, 10 Aug 2003 16:46:26 -0300, Henrique de Moraes Holschuh wrote:
> [...]
> > It is not broken at all. Due to the way sysv-rc works, no symlink at all
> > means "undefined state". If you go to a runlevel where something is started
> > (It has a S symlink), then back to a runlevel with no symlink, the service
> > is left alone (i.e. started).
> >
> > Likewise, if you go to a runlevel where the service is stopped (K symlink),
> > then back to a runlevel where there is no sysmlink, the service is left
> > alone (i.e. stopped).
> >
> > So, deleting a symlink to mean "do not start it" is an operator error. Put a
> > K symlink in there if you want the service to not be running on that
> > runlevel, OR switch to something less annoyingly error prone, such as
> > file-rc.
>
> Well, I must agree. The state is undefined, so it's probably better
> to use explicit K links.
> Still invoke-rc.d has to decide, and its current decision is inconsistent
> with its documentation.
Well, it can be argued that "invoke-rc.d itself will only pay attention to
the current runlevel, and block any tries to start an init script in a
runlevel it is not configured to be started at." means either "no S
symlink", or "a K symlink". The niceties of the sysv-rc system :-)
Powerful, but rather... picky.
As I said before, I'm happy with the RC=101 change. I didn't do it like
that right away because I didn't want extra opposition at the time.
--
"One disk to rule them all, One disk to find them. One disk to bring
them all and in the darkness grind them. In the Land of Redmond
where the shadows lie." -- The Silicon Valley Tarot
Henrique Holschuh
Reply to: