On Mon, Oct 25, 2010 at 5:53 PM, David Cantrell <dcantr...@redhat.com> wrote:
> On Wed, 20 Oct 2010, Nico Kadel-Garcia wrote:

>> It's non-standard for RHEL installations, but it's a lot faster and
>> uses a lot less bandwidth for cluster installations. I've been using
>> it to redeploy clusters and have fully defined test environments,
>> without having to rewrite kickstart operations, for years. Unwrap the
>> tarball, chroot to it, make changes, exit the chroot, then bundle the
>> new tarball. Voila, new date stamped OS image.
>
> You say you've been doing this for years, so I'm curious as to what
> particular
> configuration settings and interactions are troublesome for you in anaconda.
> The decisions that anaconda makes are not arbitrary, but rather done to make
> sure we have a consistent experience for all users across all platforms.
> Certain ways of doing things are simply not supported in the installer,
> because we want to make it clear what the intended installation path should
> be.

I gave David a much longer private note with a list of outstanding
Anaconda issues.

In short, many of the "intended installation paths" have been
unfortunately curtailed. Others go the long way around. I'll give one
example of each here.

* Curtailed. Anaconda does not store the actual ks.cfg file used in a
kickstart installation, it stores what it derives to somewhat resemble
the structure it has worked out with you. The result is that for
systems that have multiple '%pre' or '%post' scripts, those are
discarded and the host has no record of what what was actually used.
This is an *old* bug, and one I've complained about, in service
tickets, for years.

* The long way around: the overall package selection system. The base
install is *way* too big, and includes numerous unnecessary packages.
(How many *different* ways to we need to allow RedHat to dynamically
commit updates from RHN? Seriously?) This mandates going back and
yanking out unnecessary daemons and components in a %post script.

_______________________________________________
rhelv6-beta-list mailing list
rhelv6-beta-list@redhat.com
https://www.redhat.com/mailman/listinfo/rhelv6-beta-list

Reply via email to