Xen users - Feb 2007 - Running Redhat Advanced Server 2.1 on hvm

Hi,

   I was trying to have a working RHEL2.1 domU and found a mail sent  
on January 26th to xen-users by Nelson Chan. I am subscribed to the  
list but didn''t have the original mail in my mailbox, so I''m
sending
this message with the same subject. Sorry if I screw up you mail  
reader''s threading algorithm :)

   Anyway, after successfully installing a RHEL2.1 in hvm mode from  
our kickstart server, I tried two things:

   - running with non-hvm using the 2.4 xenU kernel provided by  
xensource. The system would boot although the console was stuck on  
the last init script launched. The init process would give weird  
messages (sorry, I didn''t wrote them down) but the system could be  
reached by ssh. Some binaries (yum for example) would segfault. The  
upgrade to RHEL3''s glibc worked, but I didn''t want to go that
way...

   - running with hvm and the default RHEL2.1 kernel. This didn''t  
work and gave a kernel panic upon boot (I tried with -52 and the  
latest -71 release level from RH). I then took the vmlinuz and initrd  
files from the installation cd and used them. They are actually built  
from the same srpm and are found in the -BOOT package. I just had to  
tweak the initrd file so that it wouldn''t launch the install after  
booting, but it worked.

   I''m using the system with this setup. As soon as I''ll have
time,
I''ll diff the kernel config files to try and locate the option(s)  
that make the regular kernel crash.

-- 
Eric-Olivier Lamey




_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


Eric-Olivier Lamey wrote:
>   - running with hvm and the default RHEL2.1 kernel. This didn''t
work
> and gave a kernel panic upon boot (I tried with -52 and the latest -71
> release level from RH). I then took the vmlinuz and initrd files from
> the installation cd and used them. They are actually built from the same
> srpm and are found in the -BOOT package. I just had to tweak the initrd
> file so that it wouldn''t launch the install after booting, but it
worked.
This sounds suspiciously like the MXT problem we experienced with RH7.2.
 Did you see a bunch of "Ooops! Too many entries in the memory map!"
and
"MXT page nr start=ffffffff len=ffffffff" errors before the kernel
panic?
> I''m using the system with this setup. As soon as I''ll
have time, I''ll
> diff the kernel config files to try and locate the option(s) that make
> the regular kernel crash.
I''d be interested in hearing about the results.  I don''t know
if RHEL2
has the MXT patches, but if it does, my money is that''s where the
problem is.

- -Karl
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFFwg1Xr6qWhNgoahURAnDDAJ0e/KglX8KubBG+UbCxJInPeNFbawCdEKaQ
YCMWUUeeFmZ26/YtBloO/I0=ABtP
-----END PGP SIGNATURE-----

__ 
    This communication is intended for the use of the recipient to whom it
    is addressed, and may contain confidential, personal, and or privileged
    information. Please contact us immediately if you are not the intended
    recipient of this communication, and do not copy, distribute, or take
    action relying on it. Any communications received in error, or
    subsequent reply, should be deleted or destroyed.
---

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Eric-Olivier Lamey wrote:
>   I did some further testing today and you might actually be right. There
>   are MXT related patches in the RHEL2.1 kernel that are enabled
>   to i686 and i686-smp configs. I''m trying to recompile the SRPM
without
>   those patches to see if it works. I''ll send my results to the
list.
You might be able to just use the most-recent kernel update from Red
Hat, as it appears as though Red Hat dropped support for MXT later on
(at least for RH7.2 - I''d guess that RHEL2.1 would be the same.)  We
used the last vendor-supplied kernel for RH7.2, which fixed the problem.
>   Karl, thank you for the hint, diff''ing the -BOOT and regular
kernel
>   configs would have taken a lot of time...
No problem. :)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFFx4Rzr6qWhNgoahURAp2RAKDWSObs4SVPomwt8j29ZD+vhoAD4ACfVhau
639YLQ1PD0u2VJrcx9eKqOc=xjD8
-----END PGP SIGNATURE-----

__ 
    This communication is intended for the use of the recipient to whom it
    is addressed, and may contain confidential, personal, and or privileged
    information. Please contact us immediately if you are not the intended
    recipient of this communication, and do not copy, distribute, or take
    action relying on it. Any communications received in error, or
    subsequent reply, should be deleted or destroyed.
---

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

On Feb 5, 2007, at 8:24 PM, Karl Stevens wrote:
> You might be able to just use the most-recent kernel update from Red
> Hat, as it appears as though Red Hat dropped support for MXT later on
> (at least for RH7.2 - I''d guess that RHEL2.1 would be the same.) 
We
> used the last vendor-supplied kernel for RH7.2, which fixed the  
> problem.
   Hi,

   I _think_ I tried the latest RHEL2.1 kernel, but I''m not sure :)
   Anyway, you were absolutely right: I recompiled the kernel
   from the source RPM (revision e.71) removing the MXT stuff
   in the config and it works, with and without smp.

   So, once again, thanks a lot for the suggestion.

-- 
Eric-Olivier Lamey




_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users