clone 360699 -1 reassign -1 discover1 retitle -1 Entries in exception lists should overrule default lists tags -1 + d-i thanks Hi Javier, Thanks for your extensive, well researched and well written report On Tuesday 04 April 2006 11:17, Javier Fernández-Sanguino Peña wrote: > Network config issue > -------------------- > The first problem throughout the installation was the DHCP > configuration step, it goes through all the other steps without showing > any error but this step fails as it is not able to get an assigned IP > address. [...] > Notice that the Harware MAC address is all '0's, which is clearly > wrong. If you use 'dmesg' you see the following errors: [...] > Google searching turns out that this is a known bug of the 'dmfe' > driver which replaces the 'tulip' driver. If you run 'lsmod' in console > you will see that both drivers are loaded: [...] > http://040.digital-bless.com/texts/installing_debian_linux_on_sun_netra >_x1.html also mentions issues with the 'dmfe' network driver. I wonder > why the installer tries to install both drivers when the 'dmfe' is not > recommended. This looks like a bug in discover. It has different lists it uses as a hardware database, the relevant ones are the pci.lst and the pci-sparc.lst. pci.lst: 12829102 ethernet dmfe 21x4x DEC-Tulip compatible 10/100 Ethernet pci-sparc.lst: 12829102 ethernet tulip 21x4x DEC-Tulip compatible 10/100 Ethernet As you can see, tulip _is_ listed as recommended for sparc, and IIUC this should overrule the default in the pci.lst. However, from your report for some reason both are being loaded. That said, this won't be fixed for Sarge anymore and the installer for Etch no longer uses discover for hardware detection, but udev. I'd be interested to know how the Etch Beta 2 release deals with this NIC. > SILO issue > ---------- > SILO installation failed > Running "/sbin/silo" failed with error code "1". [...] > /etc/silo.conf appears to be valid > Fatal error: Cannot open /dev/hda3 This is the first time I see that error and Google has nothing on it either. The only strange thing I can see about your partition table is the fact that 0-16765 seems unused. Is that free space? > However, if you restart the installer (without rebooting the system) > and rerun the system (either from the installer or from a chroot) > it *does* work. ? Unless you can reprocuce the problem on a new install, I don't see any real possibility of tracing this. I don't see much point in reassigning this to silo-installer to be honest. Let's hope it may prove useful sometime having this in the archive... > RAID issue > ---------- > Software RAID can not be configured throughout the installation since > the partitions cannot be configured as a RAID volume. > The option: "Use as: physical volume for RAID" is not available > in the 'partman' module. Known issue; needs someone to work on adding RAID support for sparc. There's quite a few little issues that could be improved for sparc... > Reboot issue > ------------ > After the installation, the system does not boot and returns the > following error message: > > Allocated 8 Megs of memory at 0x40000000 for kernel > Uncompressing image... > Loaded kernel version 2.6.15 > Illegal Instruction > > If the system is hard-reset (soft-reset won't do) then it is able to > reboot properly and load the Linux kernel. Could this be related to the silo problem somehow? I think I do remember earlier mails or reports that talked about needing a power off before rebooting into the installed system. Cheers, FJP
Attachment:
pgpQh0uImasLj.pgp
Description: PGP signature