[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Defect #129885 addressed



One of my action items was to resolve defect #129885, Clarification for
Run Level 2.   To complaint was made that "multiuser with no remote
networking" was ambiguous; did this mean no network devices enabled at
all, or no network services?  I've addressed this by changing the
definition of run level 2 to read:

multiuser with no remote services exported

This is most in keeping with existing practice (no one uses the
semantics of disabling eth0 in run level 2), and historical
interpretations of run level 3 by other Unix/Linux-like systems.  It is
also in keeping with the original intentions when this was discussed,
given a review of the mailing list archives.  

However, I will note that this is by no means universal in the Linux
world.  Debian for example makes no distinctions between run-level 2-5,
and treats them all equivalently; it also uses run level 2 as the
default run level (most other Linux distributions use run level 3 as the
default run level).  

Red Hat 7.0 does something very strange and unusual.  Run level 2
exports normal network services such as ntp, lpd, ssh, etc., but it skips
mounting remote filesystems (i.e., as a *client*), and it also avoids
starting ncsd, the name server cache daemon.  I tried to figure out the
rationale behind this, but at the moment I am mystified.  :-)

Hence we may wish to include some notes about the fact that historical
implementations treat run level 2 inconsistently in a rationale section
(do we have a standard way of placing this in the SGML files at this
point?).


I also changed the definition of run level 5 from 

xdm or equivalent

to 

multiuser with xdm or equivalent

The intent here was to clarify that run level 5 is a multiuser run
level.  (It should be obvious, but the spec should be explicit about
such things, so that software vendors who want their services active in
all multiuser run levels DTRT.)

At this point I believe defect #129895 has been addressed; George, could
you please close this in SourceForge?

							- Ted



Reply to: