ac1-qboot.iso testing

[Balint Cristian]

On Monday 18 April 2005 02:42, Michal Jaegermann wrote:
> On Tue, Apr 05, 2005 at 03:44:43PM -0400, Jay Estabrook wrote:
> > Thanks to Sergey and Cristian, the ISO image "ac1-qboot.iso" was made
> > available for anon FTP at ftp.rdsor.ro:

Thanks Michal  for report.

	I try figure out as much as possible to see whats wrong. Right now i make testings over
final 1.0 iso's.


> 
> I tried a bit more thoroughly the latest test boot image, i.e.
> ac1-qboot1.iso from April 14th, on UP1500 (Nautilus).  In general it
> boots and appears to work but here are various things which failed.
> 
> 1. When booting into a "rescue" a code which mounts filesystems
> expects to find /mnt/sysimage/sys after a root file system from a disk
> is mounted.  If this is an older installation, like RH7.2-Alpha, then
> this mount point does not exist and you are getting "some error
> occurred" message.  Nothing more illuminating in logs or other screens
> and some file systems may end up not mounted.  It is hard to guess
> what really happened.  If you will add that "/sys" then everything
> mounts just fine.  I would think that either this error should be
> ignored or "mount -p /mnt/sysimage/sys" should be done once there is a
> place for this.  The later can be difficult as file systems are
> possibly mounted read-only and maybe partially scrambled (which
> may make some mountpoints to vanish even if they were there before).
> 
> 2. Some scripts, possibly from a freshly mounted filesystems, expect
> /bin/bash.  There is a /bin/sh link and /usr/bin/bash but you will
> get "missing interpreter" if a /bin/bash link is not added.
> 
> 3. Attempts to configure a network interface via DHCP failed.  It
> appears that this is using /usr/bin/pump written in Python.  That
> returns very quickly but no address is ever configured.  Running
> 'pump -i eth0' directly has the same effect.  OTOH I can make eth0 work
> by reconfiguring it with  'ifconfig' from BusyBox which is available.
> Also specyfing a static address instead of DHCP, when asked, seems to
> work (once I had to take eth0 down and reconfigure it with 'ifconfig'
> again before it started moving traffic but maybe I made a typo?).  My
> network card is one of Tulips (PCI id 1011:0019).  I do not know if a
> behaviour with de4x5 driver would be different.  Although the last one
> is available on a boot image a file 'pcitable' is not listing that
> driver for any of network cards.
> 
> 4. When I tried an installer in a text mode then every time, to my
> great surprise, I got as far as "Running anaconda, the AlphaCore
> system installer - please wait" and the next line was "install exited
> abnormally" with no chance to check what really happened.  Expected
> for now?  I am rather used to troubles in a graphics mode. :-)
> Moreover "rescue" seems be rather close to "text" and the former works.
> 
> 5. In a graphics mode an installer starts fine and goes through
> all preparatory stages.  There is a catch.  It does "Searching for
> Fedora Core installations" and apparently is not finding anything,
> or results are ignored, so there is no option to upgrade but only
> to install.  This is in a marked difference with a rescue mode search
> which correctly identified all possible candidates for a root file
> system (I happen to have here more than one).  The surprise is
> that I would think that the same search code would be used.  Logs
> indicate that various file systems are mounted but if anything peeks
> there it is hard to tell.  I have copies of anaconda.log and syslog
> if anybody would like to have a look but they do not seem to be
> telling very much.
> 
> What a text mode would do with that I have no idea.  See point 4.
> 
>   Michal
> 
> _______________________________________________
> axp-list mailing list
> axp-list at redhat.com
> https://www.redhat.com/mailman/listinfo/axp-list
>