Prev: Gnus (Was: ditching mutt)
Next: Complément d'informations Micrologiciel
From: Ionel Mugurel Ciobîcă on 9 Apr 2010 06:10
Hi,

I just migrate my desktop from lenny to squeeze and after installing all
new packages and reboot the system hang at the boot process asking for
the root password because it can't find /dev/sdaX, where X=2,6,7,8,9.
All are valid partitions. root is /dev/sda5 and it is the only partition
mounted. fdisk /dev/sda says that /dev/sda not found. Indeed ls /dev/sd*
shows only /dev/sdY, with Y=c,d,e,f,f1, no /dev/sdZ, where Z=a*,b*. I
try MAKEDEV sda but it says that because .udev is present it must abort.

I can't use the system in this state. Any ideas?

If nothing I will try to boot with knoppix or the net-install testing CD
and chroot into the system to attempt a kernel compilation. But I am not
sure if the kernel is problematic at the moment.

Thank you,
Ionel


--
To UNSUBSCRIBE, email to debian-user-REQUEST(a)lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster(a)lists.debian.org
Archive: http://lists.debian.org/4BBEFA48.8030701(a)tue.nl
From: Stephen Powell on 9 Apr 2010 10:00
On Fri, 9 Apr 2010 05:58:32 -0400 (EDT), Ionel Mugurel Ciobîcă wrote:
>
> I just migrate my desktop from lenny to squeeze and after installing all
> new packages and reboot the system hang at the boot process asking for
> the root password because it can't find /dev/sdaX, where X=2,6,7,8,9.
> All are valid partitions. root is /dev/sda5 and it is the only partition
> mounted. fdisk /dev/sda says that /dev/sda not found. Indeed ls /dev/sd*
> shows only /dev/sdY, with Y=c,d,e,f,f1, no /dev/sdZ, where Z=a*,b*. I
> try MAKEDEV sda but it says that because .udev is present it must abort.
>
> I can't use the system in this state. Any ideas?
>
> If nothing I will try to boot with knoppix or the net-install testing CD
> and chroot into the system to attempt a kernel compilation. But I am not
> sure if the kernel is problematic at the moment.

Migrating from Lenny to Squeeze (or any migration, really) is tricky.
You can't just change your /etc/apt/sources.list file to point to the
squeeze repositories and then do "aptitude update;aptitude full-upgrade".
The problem you are describing above is only the first of your problems
if you have done that. You need to follow the upgrade outline mentioned
in the Release Notes. You didn't say how you did it, and it's too late
in any case, but I just wanted to mention that while we're on the subject.

Assuming that you did the migration properly, this may be a device name
issue. There's been some problems with device names recently, particularly
with IDE hard disks. It hasn't happened to me, but others have reported
device names going from /dev/hdx to /dev/sdx and back again with
subsequent maintenance. You may have to boot from a rescue disk,
edit your /etc/fstab to change device names, according to what they
are now, and rebuild your initial RAM file system.

If that is the issue, an alternative to traditional device names is
to use UUIDs.

--
.''`. Stephen Powell <zlinuxman(a)wowway.com>
: :' :
`. `'`
`-


--
To UNSUBSCRIBE, email to debian-user-REQUEST(a)lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster(a)lists.debian.org
Archive: http://lists.debian.org/371933597.2114181270821502137.JavaMail.root(a)md01.wow.synacor.com
From: Stephen Powell on 9 Apr 2010 16:40
On Fri, 9 Apr 2010 10:43:12 -0400 (EDT), Ionel Mugurel Ciobîcă wrote:
> On Fri, 9 Apr 2010 09:58:22 -0400 (EDT), Stephen Powell wrote:
>>
>> Migrating from Lenny to Squeeze (or any migration, really) is tricky.
>>
>
> I know. I manage fine from bo 2 hamm, from hamm 2 potato, from
> potato 2 woody, from woody 2 sarge and from sarge 2 lenny.

If you've been around that long, then why is it that you didn't know to
reply to the list instead of to me personally? Please post *and* reply
*only* to the list. You can CC someone if they ask for a CC, but always
include the list as one of your recipients.

>> You can't just change your /etc/apt/sources.list file to point to the
>> squeeze repositories and then do "aptitude update;aptitude full-upgrade"..
>> The problem you are describing above is only the first of your problems
>> if you have done that. You need to follow the upgrade outline mentioned
>> in the Release Notes. You didn't say how you did it, and it's too late
>> in any case, but I just wanted to mention that while we're on the subject.
>
> I did this:
> - edit /etc/apt/sources.list (changed lenny/stable to squeeze/testing)
> - aptitude update
> - aptitude install apt dpkg aptitude
> - aptitude full-upgrade (few times)

That is not the proper procedure. Read the release notes.
The most important step you missed is to issue "aptitude dist-upgrade"
*before* the "aptitude full-upgrade", but there are other steps you left
out as well. I speak from experience. I tried to upgrade a system from
Etch to Lenny by "winging it" and not following the upgrade procedure
listed in the release notes, and my system was almost unusable. Fortunately
I had another system on which I had installed Lenny directly, and by comparing
installed packages between the install-Lenny-from-scratch system and the
upgrade-from-Etch-to-Lenny system I was eventually able to recover my system.
But it took many days. Following the upgrade procedure in the release notes
is very important. Take it from one who learned that the hard way.

>> Assuming that you did the migration properly, this may be a device name
>> issue. There's been some problems with device names recently, particularly
>> with IDE hard disks. It hasn't happened to me, but others have reported
>> device names going from /dev/hdx to /dev/sdx and back again with
>> subsequent maintenance. You may have to boot from a rescue disk,
>> edit your /etc/fstab to change device names, according to what they
>> are now, and rebuild your initial RAM file system.
>
> That is not the issues. The booting process manage to use /dev/sda5 as root
> but then it can't find the other /dev/sdaX. Indeed, ls /dev/sd* /dev/hd*
> shows no /dev/hd* files and onlu /dev/sdY, with Y=c,d,e,f,f1, not even
> /dev/sda5 which
> is already mounted as root.
>>
>> If that is the issue, an alternative to traditional device names is
>> to use UUIDs.
>
> In /dev/.udev/ are some entry with uuid, but they are for the external hdd
> (also sata) which is there as /dev/sdf.
>
> I don't understand why /dev/sda5 is mounted as root, then all /dev/sdaX are
> no longer available. dmesg | grep sda shows all partitions.
>
> This is not a kernel issue. I manually install a kernel (via the memory
> stick)
> from debian site (linux-image-2.6.32.1-3_i686.deb) and it is doing the same.
>
> I can boot fine with the rescue CD and chroot into the installed
> squeeze, but
> I don't know any more how to fix it.

Neither do I. Maybe someone else has some ideas, but I'm out. I don't know
about you, but I know what I would do if it were me. I'd reinstall from
scratch. (I always keep /home in a separate partition for just such
occasions as this.) I'd reinstall from scratch and reformat the / partition.
If you don't have /home as a separate partition then back up the /home directory
somewhere first so you won't lose your personal files. I strongly recommend
that you make /home a separate partition in your new install, if it isn't now.

I also recommend that you use the latest "daily build" development version
of the Debian installer if you want to install Squeeze directly.
That's the only version of the D-I that I trust to install Squeeze at this
point. Good luck.

If I had known in advace how much work it was going to be to recover my damaged
system from a bad upgrade attempt, I wouldn't have even tried. I would have
reinstalled from scratch. You've got your system hosed up much worse than
mine was. I recommend that you reinstall from scratch. And next time,
read (and carefully follow) the release notes when you do an upgrade.

--
.''`. Stephen Powell <zlinuxman(a)wowway.com>
: :' :
`. `'`
`-


--
To UNSUBSCRIBE, email to debian-user-REQUEST(a)lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster(a)lists.debian.org
Archive: http://lists.debian.org/1972509979.2211501270845101316.JavaMail.root(a)md01.wow.synacor.com
From: Tom H on 9 Apr 2010 16:40
On Fri, Apr 9, 2010 at 4:31 PM, Stephen Powell <zlinuxman(a)wowway.com> wrote:
> On Fri, 9 Apr 2010 10:43:12 -0400 (EDT), Ionel Mugurel Ciobîcă wrote:
>> On Fri, 9 Apr 2010 09:58:22 -0400 (EDT), Stephen Powell wrote:
>>>
>>> Migrating from Lenny to Squeeze (or any migration, really) is tricky.
>>
>> I know. I manage fine from bo 2 hamm, from hamm 2 potato, from
>> potato 2 woody, from woody 2 sarge and from sarge 2 lenny.
>
> If you've been around that long, then why is it that you didn't know to
> reply to the list instead of to me personally?  Please post *and* reply
> *only* to the list.  You can CC someone if they ask for a CC, but always
> include the list as one of your recipients.
>
>>> You can't just change your /etc/apt/sources.list file to point to the
>>> squeeze repositories and then do "aptitude update;aptitude full-upgrade".
>>> The problem you are describing above is only the first of your problems
>>> if you have done that.  You need to follow the upgrade outline mentioned
>>> in the Release Notes.  You didn't say how you did it, and it's too late
>>> in any case, but I just wanted to mention that while we're on the subject.
>>
>> I did this:
>> - edit /etc/apt/sources.list (changed lenny/stable to squeeze/testing)
>> - aptitude update
>> - aptitude install apt dpkg aptitude
>> - aptitude full-upgrade (few times)
>
> That is not the proper procedure.  Read the release notes.
> The most important step you missed is to issue "aptitude dist-upgrade"
> *before* the "aptitude full-upgrade", but there are other steps you left
> out as well.

You must have meant:
issue "aptitude safe-upgrade" *before* the "aptitude full-upgrade"


--
To UNSUBSCRIBE, email to debian-user-REQUEST(a)lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster(a)lists.debian.org
Archive: http://lists.debian.org/r2i6d4219cc1004091335p9b2d0ac1uf7b7a280cbee290(a)mail.gmail.com
From: Stephen Powell on 9 Apr 2010 17:10
On Fri, 9 Apr 2010 16:35:43 -0400 (EDT), Tom H wrote:
> On Fri, Apr 9, 2010 at 4:31 PM, Stephen Powell wrote:
>> That is not the proper procedure.  Read the release notes.
>> The most important step you missed is to issue "aptitude dist-upgrade"
>> *before* the "aptitude full-upgrade", but there are other steps you left
>> out as well.
>
> You must have meant:
> issue "aptitude safe-upgrade" *before* the "aptitude full-upgrade"

Correct. My mistake. Good catch!

--
.''`. Stephen Powell <zlinuxman(a)wowway.com>
: :' :
`. `'`
`-


--
To UNSUBSCRIBE, email to debian-user-REQUEST(a)lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster(a)lists.debian.org
Archive: http://lists.debian.org/1251236643.2217511270847170483.JavaMail.root(a)md01.wow.synacor.com
 |  Next  |  Last
Pages: 1 2 3
Prev: Gnus (Was: ditching mutt)
Next: Complément d'informations Micrologiciel