From: Stephen Isard (S.IsardDeleteThis@ed.ac.uk)
Date: Wed Dec 15 1999 - 13:36:08 CET
Hello,
I've been having trouble getting a single floppy muLinux to boot on a
five year old Dell 486/25s with 12 Mb ram. The problem occurs at the
stage where the USR section of the floppy is uncompressed into
/dev/ram1. Very occasionally this works and I get the "I feel good"
message, but much more often I get the "This usr have some problems"
message. I can get muLinux to start up correctly and run programs as I
would expect if I go into maintenance mode and copy a version of USR
from the hard disk to /dev/ram1, so I think the ROOT and kernel are
getting loaded ok.
One possibility is just that the floppy itself is bad, but I don't think
so because a) I have tried with several different floppies and b) all of
these floppies run muLinux with no problem on a different (64Mb Celeron)
machine.
In fact, I think the trouble comes in reading the data from
/dev/fd0H1722, rather than in decompressing it, because I can go into
maintenance mode and copy the data to the hard disk with 'fastdd
/dev/fd0H1722 1137' twice in a row and get files that differ according
to cmp and which have bad checksums according to 'bzip2 -stvv'.
So another possibility is that the floppy drive is faulty. Maybe it is,
but why should the fault show itself so reliably in this particular
context and no other that I am aware of? I have used the floppy drive
for file transfer under dos/windows for five years and minix for a few
months and never noticed any corruptions. So maybe the 1.7Mb Linux
driver pushes the hardware harder than dos/windows. But there doesn't
seem to be any problem in reading in the kernel and root from this same
floppy and I have a floppy with tomsrtbt that goes through a very
similar loading procedure (reading a compressed /usr filesystem from a
1.7Mb floppy into a ram disk) and that works perfectly every time.
Does anyone know whether the 1.7Mb floppy driver of tomsrtbt likely to
be any different from the one in the muLinux kernel? The block devices
do have different names - fd0H1722 vs. fd0u1722.
A possible clue is that I sometimes get an error message "fd0 timeout
handler died" while reading USR from the floppy. (I haven't seen it at
any other time.) However, copying can go wrong without the message.
I'd be grateful for any ideas on how to diagnose or fix this problem.
(In particular, I'd rather not buy a new floppy drive only to discover
that the device driver is assuming a faster processor or something.)
On an entirely separate topic:
I find that I can't mail to sunsite.auc.dk with my normal mail program,
plain old unix mail, called mailx under solaris. The mail is eventually
returned with a message saying it couldn't be delivered after 4 or 5
days and something about "451 Open relay". In particular, I couldn't
subscribe to this group until I discovered that mail sent with the
netscape mailer would get through. I have mailed
postmaster@sunsite.auc.dk (using netscape!) enclosing the error messages
I get with the other mailer, but I've had no response so far. Has
anyone else had this problem or know what it is about? I've never had
it with any other site.
Regards,
Stephen Isard
---------------------------------------------------------------------
To unsubscribe, e-mail: mulinux-unsubscribe@sunsite.auc.dk
For additional commands, e-mail: mulinux-help@sunsite.auc.dk
This archive was generated by hypermail 2.1.6 : Sat Feb 08 2003 - 15:27:12 CET