Trying to install rawhide: madness and where do we hope to go ?

[Jeremy Katz]

On Wed, 2005-09-07 at 06:21 -0400, Daniel Veillard wrote:
>   First trying the installer, I have a local copy of the rawhide mirror,
> any attempt to bout from the boot image results in anaconda failing
> not able to import os or sys at the python level, i.e. not the kind of
> bug you can work around by yourself if you're not intimate with anaconda:

And as Paul sent mail about a few weeks ago, things are very much in
flux right now for the installer.  Does it suck a bit?  Yes.  But
unfortunately the installer is very much "the sum of the parts of the
OS" and so we have to be testing in a live environment.  And there's
some real work that has to be done this time instead of just touchy
feely stuff.  And that's going to involve breaking various things.
We're trying to avoid doing it as much as possible, but sometimes it
happens.

>   it had been that way for months, general feedback is installer does
> not work, and that the expected way is to upgrade from a FC4 setup.

For months?  Things have mostly been working except for the past week or
so.  Caveats of: minimal package set, LVM needs a remake of the initrd.
Neither of which is impossible things to deal with.

>   So okay it's takes way more time but I reinstall a FC4 partition from
> scratch, then modify the yum.conf to point to my local rawhide repository.
> I upgrade yum, I upgrade the kernel and one hours and a half later I try
> to reboot. Dies as all init level respawns too fast. 2.6.13-1.1538_FC5 won't
> boot but the old 2.6.11-1.1369_FC4 still allows to boot.

See the (several) threads about the kernel being broken the past few
days.  Poor timing. ;)  As to testing -- hard to do across the variety
of arch and hardware combinations.  The worst tends to be right after a
new kernel version since there are lots of things going into the
upstream kernel and davej does a good job of making sure we follow that
closely :)  

>   Then reboot ... 2.6.13-1.1538_FC5 still fails to boot all process forked
> by init just seems to crash, but booting the old 2.6.11-1.1369_FC4 works.
> Half a day later I still don't have a setup to test inotify or recent Xen.

FWIW, you're not going to have luck testing recent Xen with rawhide
anyway.  There are other known broken bits there ;)

>   Do we really expect to get significant feedback on a system that is so
> hard to install or run ? 

I actually don't think the picture is as grim as you paint it here.  A
few days have been worse off, but things are definitely getting better.

> Delaying the release of FC5 raises the risk of
> decoupling from our user base and people who test bleeding edge, with a
> HEAD that is so hard to install and get running people who want to test
> new stuff have an easier upgrade path by installing Ubuntu (and maybe
> openSuse) than trying to get Rawhide going. I am afraid the current state
> of Rawhide just means silent exodus of the people who really help building
> the distro. The fact that the installer has been broken for months in my
> experience is a very significant threat, and increase the risk associated
> to delaying FC5 to an extend we didn't anticipated. 

Actually, we delayed FC5 partially so this work *COULD* happen.  The
installer needs to be mostly working by the time test1 comes out.  And
some of the changes just need more time than the usual 2.5 months
between a release and test1.

I do think that we want to consider having four test releases with the
first one earlier than the current schedule.  That will help to avoid
some of the problems you're afraid of.

Jeremy