Re: Tale of woes: hard copy
> %%% Can't open server boot script /dev/hd1s1/boot/servers.boot: 2502
> %%% Server boot script ? [/dev/hd1s1/boot/servers.boot]
> /dev/hd2s1/boot/servers.boot
> %%% hd24: bad access: block = 28, count = 2
> %%% end_request: I/O error, dev 16:04, sector 28
> %%% (serverboot): Running without any paging
> %%% Hurd server bootstrap: ext2fs.static[hd2s1] exec init proc auth
> %%% Single-user environment: /hurd/init: /dev/console: No such file or
> directory
> %%% /hurd/init: Using temporary console /tmp/console
> %%% /hurd/init: Cannot execute /bin/sh: No such file of directory
> %%% init: rebooting Mach (flags 0) ...
[...]
> Main question: Does any immortal have any clue to what is happening ?
Who wants to live forever?
I believe the message about "hd24: bad access" is a known problem that
perhaps has been fixed in current gnumach; it has been discussed here,
check the mailing list archives. But anyway, I think it's harmless, just a
spurious message.
The fact that you got past the "Hurd server bootstrap" line suggests that
the disk and filesystem are working OK. init, proc, and auth are loaded
using the Hurd filesystem in the normal way, so seeing those start up
generally indicates that the disk and the filesystem are working.
The stuff about the console is normal to see when /dev/console is not set
up, which is normal when you haven't run native-install yet.
The "Cannot execute" error indicates either that /bin/sh is missing, or
that the dynamic linker (/lib/ld.so.1 or /lib/ld.so) is missing.
I'm not sure that this is your problem, but is there a reason you specified
the wrong disk in the kernel command line?
Have you tried it with "root=hd2s1"? (The "/dev/" is optional.)
> I have some questions ..
> 1: Is it necessary for the second IDE HDD to be a a Primary Slave ?
There should be no requirements to set your disks up any differently than
you do for Linux. GRUB uses the BIOS (as LILO does) to access the disk,
and then the Mach
> 2: GRUB, Hurd and Mach all use different partition naming styles; could
> we have them use the same naming style, like the Linux style, which
> almost all of us are used to ?)
The Hurd and Mach do not using different styles. The names you give to the
Hurd are passed directly through to Mach. GRUB indeed does use an
annoyingly different style (0-origin slice names), and that will be fixed.
For the Hurd/Mach syntax (which is the FreeBSD syntax), the partition part
of the Linux name (e.g. "1") is the same as the part after the s (e.g. "s1").
Reply to: