Re: [REDHAT] Webmin

[Rick Stevens <rstevens vitalstream com>]

David Kramer wrote:
> On Wed, 29 May 2002, Mugleston, Brad wrote:
>
>
> > There was some talk about Webmin a few days ago.  I looked at it and it
> > looks nice but I have a question.
> >
> > My setup is working OK and I'm a little gun shy about doing anything after
> > my experiences with Linuxconf.   If I start up Webmin will what I have
> > change or will it go out and discover what I've already got going and report
> > that and let me change that - i.e. does it use the existing files or make up
> > it's own.
>
>
> I feel your pain.  It will end with Webmin.  Let me explain.
>
> You know how much every geek hates the Windows system registry database, 
> with all this obscure stuff in it you need special tools to edit?  Well 
> linuxconf and yast and all those tools are just one small step away from 
> that.  That's why they get out of sync.
>
> > From the webmin home page:
> "Webmin consists of a simple web server, and a number of CGI programs 
> which directly update system files like /etc/inetd.conf and /etc/passwd".
>
> In other words, they edit the same files you would if you were an 
> ubergeek, so there's nothing to get out of sync.  You can use webmin one 
> day and edit the files directly the next.

Provided the CGI scripts know what files to mung and what their format
is.  There ARE occasions where the config file formats change between
releases and webmin will have the same problem.  As an example,
witness the change in RedHat when inetd was abandoned and xinetd was
used in its place.  All the futzing in the world with the inetd.conf
file wouldn't help you and unless your webmin knew about the change,
you would be stuck.

The only way out for automated administration of systems that can be
as complex as Linux is to use some form of config file version
numbering scheme for apps that use one and making the maintainers of
the programs managed by webmin release new CGIs when they make a
change.  Many may balk at that idea as it somewhat violates their
concept of unfettered development.

A registry is a valid CONCEPT, but M$'s implementation is the absolute
pits.  We had things like registries in VAX/VMS, MVS/OS, AmigaDOS and
a host of other OSes--they just weren't called registries.  You might
also note that most of those OSes weren't supported much by free
software, so the authors had a bit more control.
----------------------------------------------------------------------
- Rick Stevens, SSE, VitalStream, Inc.      rstevens vitalstream com -
- 949-743-2010 (Voice)                    http://www.vitalstream.com -
-                                                                    -
-           "I understand Windows 2000 has a Y2K problem."           -
----------------------------------------------------------------------