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

Re: Rant about installer features (Re: Progeny)



On Sun, 2004-08-15 at 11:44 +1000, Paul Gear wrote:
> Greg Folkert wrote:
> > ...
> >>>Just for laughs, how does FC go on that box?
> >>
> >>Nothing to laugh about - it just works.  Perfectly, straight out of the
> >>box.  No mucking about, just create the partitions i want (RAID 1 on
> >>everything: /, /boot, and swap) and install.  So does SuSE Personal 9.1.
> > 
> > 
> > Too bad. MD Raid is tough for a bootable setup with automated tools.
> 
> This is part of what i don't understand.  As Alvin Oga and i were
> discussing a while back (see archives), it is a supported configuration
> by the kernel, and Red Hat have supported it since 7.3.  I guess that's
> why Progeny decided to port anaconda - it worked out your tough problem.

It not so much the anaconda issue, more over it is the "Debian Way" of
doing things. I guess^WKNOW, Grub has issue with md as root. But, as
with all thing Linux... <accent style="stereotypical crusty German
Colonel"> "We have ways of making it work! Muahahaha!"</accent>

> > I have a workaround to get it to work proper.
> > 
> > If you want I can help you through that.
> 
> Thank you.  That's the first offer i've had (except for John
> Summerfield's offer to consult for a fee).  If you'd like a description
> of my problem, see
> <http://lists.debian.org/debian-boot/2004/07/msg02914.html> and
> <http://lists.debian.org/debian-boot/2004/07/msg00771.html> (although
> the hang is fixed under recent snapshots).
> 
> I've read several HOWTOs on it, but none of them seem to solve the
> problem completely:
> 
> http://lists.debian.org/debian-boot/2004/07/msg01195.html
> http://www.james.rcpt.to/programs/debian/raid1/
> http://members.ferrara.linux.it/calicant/docs/debianraid/debian_raid1.html
> 
> I just did a bit more searching and found
> http://alioth.debian.org/projects/rootraiddoc/ - maybe i'll try that.

I am surprised you haven't gotten it to work. Yes it is a Mickey mouse
way to setup the "Ultimate" Linux Distribution... but, at present, this
Voluntary project doesn't want to "Do it auto-magically" when there is a
way to do it with "other means". Yes, it is asinine, but think about how
Debian really works (well, doesn't work sometimes). It is quite
possible, someone will get a Hair up the arse and make this work... but
until then it'll be a kludge to get proper.

Personally, I have a server with lvm on-top of a "Promise Raid
Mirroring" setup. Works, I see only one drive. But I have / as an LV.

See if you can tell were I patterned my VG names (on machine with more
than one VG)from:

/dev/rootvg
/dev/datavg
/dev/scratchvg

I have bookmarked these and will do the apropos thing. Might be just a
bit though.

> Maybe all of this is pointless if i just wait for the next Progeny beta
> release.  According to Ian Murdock it is fairly close, and as long as it
> can see my SATA drives, i suspect it will work.

Now come on you shouldn't feel that way. I am not a programmer(!!!!).
Well I don't want to be one... but any good Systems Admin/Analyst is.

I'd love to start a project that will do all the work after install. Do
it they way the Admin wants... asking questions, verifying things as you
go, making the needed hardware addressing changes. But, I currently only
have coupla hours a week to start something like that.

Mainly, I value Family Time.

> > ...
> > I jumped ship from RedHat long before that. RH7.3 is the Last version I
> > installed from ANYONE. Customers included.
> 
> There seems to be a large anti-Red Hat/SuSE contingent here.  (I have a
> gift for stating the obvious. :-)  I know people don't like their
> trademark & Enterprise subscription licenses, but until they stopped
> supporting Red Hat Linux, they were a genuinely free, useful, and stable
> option.  Everything "just worked" for me, including RPM, which everyone
> who hasn't take the time to understand seems to think is fundamentally
> broken.  (It's not - saying RPM is broken because it doesn't
> automatically resolve dependencies is like saying dpkg is broken for the
> same reason.  It's a low-level tool for the job of package management -
> the smarts are in the upper levels like yum & apt-get, just like they
> are on Debian.)

I never said I was RedHat/SuSE/RPM Based hostile. I personally, don;t
like the costs they are tagging things like Libraries and Schools with
Such HUGE costs, that they have to do the wrong right thing
(http://www.whiteboxlinux.org/)

I don't have any problem with RPM. To me dpkg and rpm are equivalent in
function and process. The problems usually arise from not-finite enough
packages that HUGELY overlap, and therefore that is proof of bad a QA
process.

The thing that makes Debian so distinct from other Linux "Vendors" is
Policy. Policy, Policy, Policy. Oh yes of course -- Policy. Policy
demanded by a contract that defines HOW Debian can and cannot proceed.
Not some board room telling you to get it out the door as fast as
possible. Thereby also being Voluntary... you can't *MAKE* anyone do
anything in Debian.

That stated, Policy is also Debian's biggest Achilles heel. For obvious
reasons

> > ...
> > I am supporting the existing 7.3- with yum. Using www.fedoralegacy.org
> > awesome. But not quite Debian... :)
> 
> And not quite updated regularly, either.

But regular enough. Mainly I had to build my own RPMs for quite a while
until I found that.

> > ...
> >>Sorry for the venting, but i imagine i'm not alone - there are plenty of
> >>Red Hat refugees around, and i was almost sold on Debian before i even
> >>installed it.  You can make some significant new converts by just taking
> >>our concerns seriously.
> > 
> > We aren't called Snobbians fer nothing. Another common term I ave seen:
> > Dweebians.
> 
> I found that out the hard way...

Sorry, 'bout that.

> > ...
> > Don't let us get you down.
> 
> Too late for that.  ;-(

Bummer
-- 
greg, greg@gregfolkert.net

The technology that is
Stronger, better, faster:  Linux

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: