At 02:25 PM 3/10/2002 +0100, Matijs van Zuijlen wrote: >> This seems counter intuitive to me. Can someone explain why it works that >> way?
>To me this makes perfect sense. The K's tell you what you don't want in >a runlevel, the S's tell you what you do want. In your interpretation, >if level 2 had service foo started (i.e., it has: S<number>foo), and 3 >didn't (so it has K<number>foo), it would never get stopped going from 2 >to 3. If you follow the manual, foo will always be running in level 2, >and never in level 3. Maybe I'm looking at the run levels wrong. I must be. Still it seems backwards. Say run level 3 is for X, graphical login and such on a machine. I have a service I want running only in level 3, perhaps a font server. So the current design means to start the service I need to add a S script (link) in rc3.d. That makes sense. But to stop that service when switching to other run levels means that a Stop script must be added to *all* the other rcN.d directories. It would seem easier to add the Stop one place --- the directory where it was started. Or another way to look at it, say I'm switching from run level 4 to level 2 -- that is I've never had X running at run level 3. Why should "K" scripts be run to stop services related to run level 3 (or any other run levels)? Why does switching from run level 4 to 2 need to worry about stopping services not started in run level 4? If both run level 4 and level 2 are text-based modes, why bother trying to stop the font server if it was never running in the first place? It just seems more logical to stop things you *know* are running then just trying to stop things that *might* be running. This is Linux, though, so there must be a reason I'm missing. Bill Moseley mailto:[EMAIL PROTECTED]
