Bug#305230: Fwd: Bug#305230: [sparc64] [rc3] [netinstall] RC3 does not work on Ultra5 with both kernels 2.4 and 2.6
Hi,
On Thu, Apr 21, 2005 at 11:23:37PM -0700, Blars Blarson wrote:
> In article <20050421142818.GA3820@melusine> Frederic.Lehobey@free.fr writes:
> >> > We tried to install RC3 on an Ultra5 machine. It failed with both kernels.
>
> >> > VFS: Cannot open root device "hda2" or 03:02
>
> I've seen this on one of my ultra 5's before they both decieded to die.
> I think it's a reoccurance of the silo initrd loading bug (or something
> related) that can be worked around by changing the memory configuration.
>
> What memory configuration do you have? What CPU speed? If 4 sticks,
> try removing 2. If 2, try moving to the other pair of slots. (Mine
> had 4x32M, d-i worked with 2x32M in slots closest to disk. After
> install the system booted fine with all 4 sticks.)
Thanks a lot. I have performed (with David, the initial submitter) two
new attempts, on Ultra 5 and Ultra 10 both with 384 MB of RAM (2 x 64
+ 2 x 128).
The bug does occur in these configuration in both cases with default
RC3 kernel 2.4.
Your workaround (removing the 2 x 64 MB sticks) worked both times.
Installation is successful with the following steps:
1/ netinstall with RC3 (netboot) default 2.4 kernel and default
partitioning (all system in one partition). Still with 384 MB (4
sticks).
After first reboot failure with:
VFS: Cannot open root device "hda2" or 03:02
2/ Removal of 2 sticks (128 MB), now only 256 MB (2 sticks).
3/ Reboot and successful completion of second step installation.
4/ Switch off the machine and put back of the 2 previously removed
sticks.
5/ After power up, the machine now works properly with 384 MB (kernel
2.4).
I believe this workaround deserves at least an explanation in the
release notes (errata).
Thanks a lot for sharing with us your workaround.
> --
> Blars Blarson blarson@blars.org
> http://www.blars.org/blars.html
> With Microsoft, failure is not an option. It is a standard feature.
Well let's stay modest, they may not be the only ones to experience
standard failure. :)
Best regards,
Frédéric Lehobey
Reply to: