[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

RE: big trouble after apt-get upgrading my HP9000 d360/2



Hello Jürgen,

I just discover udev (while devfs was never investigate on this hppa arch), so
I am far from being an expert but I have a install which look like yours: a
d380/2, excepted that I use soft raid1 for boot, swap, root and raid1+lvm2 for
some more system stuff and even raid5+lvm2 for a data array ;-)

---------- Initial header -----------

>From      : "Jürgen Leibner" juergen.leibner@t-online.de
To          : debian-hppa@lists.debian.org
CC          :
Date      : Tue, 21 Feb 2006 18:44:02 +0100
Subject : RE: big trouble after apt-get upgrading my HP9000 d360/2

> > -----Original Message-----
> > From: Frans Pop [mailto:aragorn@tiscali.nl]
> > Sent: Tuesday, February 21, 2006 6:30 PM
> > To: debian-hppa@lists.debian.org
> > Subject: Re: big trouble after apt-get upgrading my HP9000 d360/2
> >
> >
> > On Tuesday 21 February 2006 18:06, Jürgen Leibner wrote:
> > > udevd-event[3324]: run_program: exec of program
> > > '/lib/udev/udev_run_hotplugd' fa
> > > iled
> > > udevd-event[3325]: run_program: exec of program
> > > '/lib/udev/udev_run_devd' failed
> >
> > All these messages are harmless. The latest version of udev (0.084-5)
> > fixes this again.
> >
> > > Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
> > > ide: Assuming 50MHz system bus speed for PIO modes; override with
> > > idebus=xx Done.
> > > Begin: Mounting root file system ...
> > > Begin: Running /scripts/local-top ...
> > > Done.
> > > ALERT! /dev/sda5 does not exist. Dropping to a shell!
> >
> > This is the only real problem and is also a known issue fixed
> > in latest
> > udev. To work around it:
> > - check if 'ls /dev/sd*' actually shows /dev/sda5 or not;
> > - if it does, try typing exit at the shell prompt and the boot will
> >   probably complete normally
> > - if not, try 'cat /proc/modules'; is any module you need missing?
> >   if it is, try modprobing it, see if /dev/sda comes available and
> >   again type exit to continue the boot
> >
mmm, according to your dmesg the disk driver (ncr53c8xx) is builtin, so you
would have to modprob it and kernel also knows sda: just /dev is not well
populated.

I have a look on my testing d380 and see that:
# ll /dev/sda*
lrwxrwxrwx 1 root root 33 Feb 21 17:38 /dev/sda ->
scsi/host0/bus0/target5/lun0/disc
lrwxrwxrwx 1 root root 34 Feb 21 17:38 /dev/sda1 ->
scsi/host0/bus0/target5/lun0/part1
lrwxrwxrwx 1 root root 35 Feb 21 17:38 /dev/sda10 ->
scsi/host0/bus0/target5/lun0/part10
lrwxrwxrwx 1 root root 34 Feb 21 17:38 /dev/sda2 ->
scsi/host0/bus0/target5/lun0/part2
lrwxrwxrwx 1 root root 34 Feb 21 17:38 /dev/sda3 ->
scsi/host0/bus0/target5/lun0/part3
lrwxrwxrwx 1 root root 34 Feb 21 17:38 /dev/sda4 ->
scsi/host0/bus0/target5/lun0/part4
lrwxrwxrwx 1 root root 34 Feb 21 17:38 /dev/sda5 ->
scsi/host0/bus0/target5/lun0/part5
lrwxrwxrwx 1 root root 34 Feb 21 17:38 /dev/sda6 ->
scsi/host0/bus0/target5/lun0/part6
lrwxrwxrwx 1 root root 34 Feb 21 17:38 /dev/sda7 ->
scsi/host0/bus0/target5/lun0/part7
lrwxrwxrwx 1 root root 34 Feb 21 17:38 /dev/sda8 ->
scsi/host0/bus0/target5/lun0/part8
lrwxrwxrwx 1 root root 34 Feb 21 17:38 /dev/sda9 ->
scsi/host0/bus0/target5/lun0/part9

well for you:
[...]
scsi2 : ncr53c8xx-3.4.3g
SCSI device sda: 4165272 512-byte hdwr sectors (2133 MB)
SCSI device sda: drive cache: write through
SCSI device sda: 4165272 512-byte hdwr sectors (2133 MB)
SCSI device sda: drive cache: write through
/dev/scsi/host1/bus0/target5/lun0: p1 p2 p3 < p5 p6 >
[...]

have a look in /dev/scsi/host1/bus0/target5/lun0
if part[123] and part[56] exist then you could try to create manualy links:
/dev/sda1 -> /dev/scsi/host1/bus0/target5/lun0/part1
/dev/sda2 -> /dev/scsi/host1/bus0/target5/lun0/part2
/dev/sda3 -> /dev/scsi/host1/bus0/target5/lun0/part3
/dev/sda5 -> /dev/scsi/host1/bus0/target5/lun0/part5
/dev/sda6 -> /dev/scsi/host1/bus0/target5/lun0/part6

then try to complete the boot (I am not familiar with initramfs, so I have no
clue if Ctrl-D or init 2 will do the drill?)

If that works, updating udev would also rebuild an initrd.

That said, even thought you have a 2-way system, I would advise you to install
and use temporarely a up kernel (right now there are still an issue with
udev-0.084-5 for hppa smp which, tx to Kyle, would be soon fixed).

Hth,
    Joel

---------------------------------------------------------------
A free anti-spam and anti-virus filter on all Scarlet mailboxes
More info on http://www.scarlet.be/



Reply to: