CentOS - Jun 2005 - Re: Reboots -- LSB 2.1 Core Generic Section 8.5

From: Rodrigo Barbosa <rodrigob at
suespammers.org>
> There is nothing wrong with using X11 at runlevel 3. The only thing
> that can't be present is a display manager (KDM, XDM, GDM etc).
That's _exactly_ what I'm talking about.
Run-level 3 starts xdm.
Several Linux distros use 2 for multi-user, 3 for multi-user w/X.

Run-level 2 as multi-user w/o networking or w/o NFS was introduced
by Red Hat, and any Fedora-based (or RHL-based) distro uses it.
> If it is, it is wrong, and doesn't comply with the unix standard
(don't
> remember exactly which, but I studied it when working for Conectiva).
First off, it's not "wrong."  Many distros established many
different
standards well before any standardization efforts.  Even Linus admits
that he based many things off of SunOS 4.1 (retroactively Solaris 1)
and Solaris 2, and Solaris uses 2-3.

[ Remember, Solaris was the original GNU platform. ]

So in that regard, Debian is actually like Solaris in its use of 2-3,
instead of Fedora-based 3-5.

Secondly, you are referring to Linux Standards Base (LSB).
Specifically, Section 8.5 in the General Section of the latest revision:  
http://refspecs.freestandards.org/LSB_2.1.0/LSB-Core-generic/LSB-Core-generic.html#RUNLEVELS

And even this latest version states:
"However, it is not required that LSB-compliant run-time environments
 use these run levels; the distribution-provided install_initd script may
 map the run levels specified below to whatever distribution-specified
 run levels are most appropriate."

It should also be noted that not even Fedora Core's init scripts are
fully LSB compliant, and Debian and SuSE do far better in many areas.

LSB even declares RPM as the package standard.  But at the same time,
one might remember why LSB came into existance -- to try to get 
standards across largely FC/RHL forks and other RPM distros.  As much
as it attempts to be platform agnostic, some of its standards are
heavily defaulting to RPM ones.
> The standard is there. If a distro chooses not to follow it,
> you can be very sure it will have acceptance problems.
Not nit-pick, but your statement was _exactly_:  
  "Also, run-levels are standarized, and should be the same on all
   unix-like platforms."

Not only is there _no_ standard for System-V platforms, but
pretty much LSB says even Linux distros can do what they want.

There is _no_ "run-level standard" for "unix-like
platforms."
> When I have time later today (or tomorrow), I'll give out the
> standard I'm mentoning and will give you a proper reference.
> Sorry I can't provide it from the top of my head.
LSB 2.1 Core Generic Section 8.5.  Again, URL:  
http://refspecs.freestandards.org/LSB_2.1.0/LSB-Core-generic/LSB-Core-generic.html#RUNLEVELS

Just to summarize, this thread was _not_ to "prove you wrong."
It was just a reminder that not only do Linux systems vary wildly,
but "UNIX-like" systems do as well and you should be careful
with assumptions.  ;->



--
Bryan J. Smith   mailto:b.j.smith at ieee.org

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, Jun 02, 2005 at 02:27:38PM -0500, Bryan J. Smith <b.j.smith at
ieee.org> wrote:
> > If it is, it is wrong, and doesn't comply with the unix standard
(don't
> > remember exactly which, but I studied it when working for Conectiva).
> 
> First off, it's not "wrong."  Many distros established many
different
> standards well before any standardization efforts.  Even Linus admits
> that he based many things off of SunOS 4.1 (retroactively Solaris 1)
> and Solaris 2, and Solaris uses 2-3.
> 
> [ Remember, Solaris was the original GNU platform. ]
The original GNU development platform, that is correct. Even tho
you are refering to SunOS (prior to 4.1).
> So in that regard, Debian is actually like Solaris in its use of 2-3,
> instead of Fedora-based 3-5.
> 
> Secondly, you are referring to Linux Standards Base (LSB).
No, not LSB. It is another standard, much older than that.
> > The standard is there. If a distro chooses not to follow it,
> > you can be very sure it will have acceptance problems.
> 
> Not nit-pick, but your statement was _exactly_:  
>   "Also, run-levels are standarized, and should be the same on all
>    unix-like platforms."
> 
> Not only is there _no_ standard for System-V platforms, but
> pretty much LSB says even Linux distros can do what they want.
> 
> There is _no_ "run-level standard" for "unix-like
platforms."
I'll disagree with you on this as soon as I find the standard I'm
talking about.
> > When I have time later today (or tomorrow), I'll give out the
> > standard I'm mentoning and will give you a proper reference.
> > Sorry I can't provide it from the top of my head.
> 
> LSB 2.1 Core Generic Section 8.5.  Again, URL:  
>
http://refspecs.freestandards.org/LSB_2.1.0/LSB-Core-generic/LSB-Core-generic.html#RUNLEVELS
As I said, I was not talking about LSB, let alone 2.1. It is a standard
that even AIX (from IBM) follows, even tho I'm not sure about Solaris.
I would suppose it does, tho. I'll find it ... Eventually.
> Just to summarize, this thread was _not_ to "prove you wrong."
> It was just a reminder that not only do Linux systems vary wildly,
> but "UNIX-like" systems do as well and you should be careful
> with assumptions.  ;->
I know it is not. This is just to clarify a technical point that
is of interest for us all. A point that we don't agree uppon, and
we are both trying to find solid enough arguments to clear a
missunderstanding.

You can rest assured I don't take anything as a personal attack
easily. I always tend to consider the other person is trying to
help me understand something better. Have no worries about it :)

Let me make a even more generic summary. Standarization or no standarization,
Linux Distros do have different runlevel characteristic, and that
should be taken into consideration. I'll conceed defeat on that point.
As for the standarization, I'm still looking for the standard,
so I'll wait until my memory is proven to be wrong, which won't
be the first time.

[]s


- -- 
Rodrigo Barbosa <rodrigob at suespammers.org>
"Quid quid Latine dictum sit, altum viditur"
"Be excellent to each other ..." - Bill & Ted (Wyld Stallyns)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQFCn41npdyWzQ5b5ckRAmBIAKCXqAG1U1B5b0BLSq+6lW379GfaKwCfTYmY
3pHltqviy+okg/OhYFEcfrs=5knv
-----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, Jun 02, 2005 at 02:27:38PM -0500, Bryan J. Smith <b.j.smith at
ieee.org> wrote:
> From: Rodrigo Barbosa <rodrigob at suespammers.org>
> > There is nothing wrong with using X11 at runlevel 3. The only thing
> > that can't be present is a display manager (KDM, XDM, GDM etc).
> 
> That's _exactly_ what I'm talking about.
> Run-level 3 starts xdm.
> Several Linux distros use 2 for multi-user, 3 for multi-user w/X.
> 
> Run-level 2 as multi-user w/o networking or w/o NFS was introduced
> by Red Hat, and any Fedora-based (or RHL-based) distro uses it.
Oh, just one thing. I don't have Fedora here to test, otherwise
I would check myself.

Are you sure it is starting xdm at runlevel 3 ? I noticed several
distros are not definining the runlevel as a kernel paramter on
grub/lilo, which will bypass the one defined on inittab.

[]s

- -- 
Rodrigo Barbosa <rodrigob at suespammers.org>
"Quid quid Latine dictum sit, altum viditur"
"Be excellent to each other ..." - Bill & Ted (Wyld Stallyns)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQFCn5ampdyWzQ5b5ckRAssTAJ4r8okk4Yg5POL/6SPEDjbYkm/KqgCgtbNo
TTiFWpdn+gm1LwGzFilAMvc=JeRf
-----END PGP SIGNATURE-----