New day...
...new problems! ;-)
But first of all, hi there everybody! :-)
I'm still working on a Sun Sparcstation SLC which is intended to be an X-
terminal soon. :-) The machine will have NO disks attached to it and everything
will be there done via NFS. (At the moment, there IS a disk, of course, with
Debian running on it.)
Now, I've some rather basic problems concerning the boot process, maybe you
can comment on them or give some additional links! :-)
At least (!!) the kernel starts up now, but it then hangs because it's unable
to mount its root device (that's my problem now).
Of course, I've built the kernel with option "root on NFS". :-)
I have NOT included BOOTP support - is this a fault? (I thought the parameters
could be also given as a parameter on startup OR set with the "knl" utility?!)
So, I've to instruct the kernel now where's its root (and swap) filesystem,
right? Am I right if I use the "knl" utility for this purpose, or did I get
something wrong?
Or, just a fundamental question: If the kernel is loaded via NFS and starts up
afterwards, how does it know where (and how) to mount its root filesystem
from?
That's the thing I wanted to use the "knl" utility for, but it only says:
# knl --kernel= /target/boot/vmlinuz-2.2.9-netboot
knl: (3) File is not a kernel image.
The file "/target/boot/vmlinuz-2.2.9-netboot" is mounted via NFS and is the
kernel file. As said before, it is loaded succesfully (but then dies).
Another question: What is the thing about the /dev/nfsroot special file?
Should it be created on the server (/dev/nfsroot) or on the client's
filesystem? (here: /target/dev/nfsroot)
--
cya
ANDI
Reply to: