Jan et al,

I'd support the idea of a truly minimal install, particularly for the
'"truly_minimal"+"my_server"' use.  We do something like this for our
disk servers by installing the 'server' config then removing all the
stuff we don't want.  Sometimes painful to rework for each new SL
variant.

Regards,
	Martin.
-- 
Martin Bly
RAL Tier1 Fabric Team 

> -----Original Message-----
> From: [log in to unmask] 
> [mailto:[log in to unmask]] On 
> Behalf Of Jan Iven
> Sent: 15 May 2007 17:36
> To: [log in to unmask]
> Cc: [log in to unmask]
> Subject: OT: "truly minimal" (yum) group?
> 
> >>  By the way the "Minimal" option has gone too which we 
> never used in
> >> practice but I imagine could be useful.
> >>
> > 
> > Actually, I have found that I can do a much more minimal 
> install than
> > before.  Before when you did minimal you got core and base. 
>  Now, you
> > can actually uncheck things out of base.  (Does my server 
> really need
> > pcmcia, I think not, out it goes)  So I now have a much 
> more trim server
> > than I used to get with minimal.
> 
> Would it be worthwhile sharing ideas for a "truly minimal" YUM group
> (i.e. something that easily allows to add things, but is unlikely to
> ever need something taken out of)? This idea has been touted 
> a number of
> times by various CERN users:
> * to easily set up secure-by-default servers (take
> "truly_minimal"+"my_server", let YUM sort out the rest)
> * to create diskless images (i.e http://cern.ch/ahorvath/u-boot or
> NFS-root clients)
> * to create virtual machine images (chroot/Xen)
> 
> If we ever get a good common candidate we could add it to the 
> YUM group
> definition (yumgroups.xml) even after release...
> 
> As Troy has pointed out, TUV "core" wasn't (until now). The proposed
> group should ideally be self-contained (i.e. no dependencies 
> on packages
> not in the group). While YUM certainly handles this better 
> than anaconda
> did, it makes for easier auditing.
> 
> 
> Regards
> Jsn
>