Re: Potato->Woody kernel upgrade problems
On Sun, 21 Apr 2002 01:21, Brian May wrote: > On Fri, Apr 19, 2002 at 01:29:24PM +0200, Russell Coker wrote: > > Another possibility is bug 142916. This resulted in one of my machines > > becoming non-bootable. > > Interesting. > > I complained loudly about a similar problem ages ago, when > I installed a broken version of libc6 without ldconfig, > and it broke my entire computer... > > When I mentioned it on debian-devel, the initial response was: > that can't happen, it should have detected the error before > installing the package. Eventually that response changed to: > that is a known bug that has just been fixed. > > hmmm doesn't seem very fixed to me... Yes, well I've still got that broken package for anyone who wants to play with it... -- If you send email to me or to a mailing list that I use which has >4 lines of legalistic junk at the end then you are specifically authorizing me to do whatever I wish with the message and all other messages from your domain, by posting the message you agree that your long legalistic sig is void. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
On Fri, Apr 19, 2002 at 01:29:24PM +0200, Russell Coker wrote: > Another possibility is bug 142916. This resulted in one of my machines > becoming non-bootable. Interesting. I complained loudly about a similar problem ages ago, when I installed a broken version of libc6 without ldconfig, and it broke my entire computer... When I mentioned it on debian-devel, the initial response was: that can't happen, it should have detected the error before installing the package. Eventually that response changed to: that is a known bug that has just been fixed. hmmm doesn't seem very fixed to me... -- Brian May <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
On Fri, 19 Apr 2002 13:40, Andreas Schuldei wrote: > * Russell Coker ([EMAIL PROTECTED]) [020419 13:31]: > > > Did the initrd load at all? If it did then it could be a bug in > > > initrd-tools. Please show me the boot messages. > > an other reason for failiour was that the space in the /boot > partion was used up. I made my boot partition just big enough for > a few kernels in the old days. today that could be not enough. Another possibility is bug 142916. This resulted in one of my machines becoming non-bootable. -- If you send email to me or to a mailing list that I use which has >4 lines of legalistic junk at the end then you are specifically authorizing me to do whatever I wish with the message and all other messages from your domain, by posting the message you agree that your long legalistic sig is void. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
* Russell Coker ([EMAIL PROTECTED]) [020419 13:31]: > > Did the initrd load at all? If it did then it could be a bug in > > initrd-tools. Please show me the boot messages. an other reason for failiour was that the space in the /boot partion was used up. I made my boot partition just big enough for a few kernels in the old days. today that could be not enough. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
On Fri, 19 Apr 2002 11:47, Herbert Xu wrote: > Michael Piefel <[EMAIL PROTECTED]> wrote: > > No Disk > > The old kernel was handcrafted. The new one makes an initial RAM disk > > and loads the modules mentioned in /etc/modules. Of course, the disk > > driver had been compiled in before. I'm not sure what to do about this. > > kernel-image _could_ have warned about the missing SCSI driver, but I > > guess it's difficult. Perhaps an even bigger note about the perils of > > installing a new kernel? > > Did the initrd load at all? If it did then it could be a bug in > initrd-tools. Please show me the boot messages. Also it could simply be a matter of not having a configuration line in the lilo.conf file to load the initrd... The latest version of lilo gives messages such as the following when run with the -v option, so a kernel package could check that the initrd was mapped for an initrd kernel... Boot image: /vmlinuz -> boot/vmlinuz-2.4.18-lsm Mapping RAM disk /initrd.img -> boot/initrd.img-2.4.18-lsm -- If you send email to me or to a mailing list that I use which has >4 lines of legalistic junk at the end then you are specifically authorizing me to do whatever I wish with the message and all other messages from your domain, by posting the message you agree that your long legalistic sig is void. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
Michael Piefel <[EMAIL PROTECTED]> wrote:
> No Disk
> The old kernel was handcrafted. The new one makes an initial RAM disk
> and loads the modules mentioned in /etc/modules. Of course, the disk
> driver had been compiled in before. I'm not sure what to do about this.
> kernel-image _could_ have warned about the missing SCSI driver, but I
> guess it's difficult. Perhaps an even bigger note about the perils of
> installing a new kernel?
Did the initrd load at all? If it did then it could be a bug in
initrd-tools. Please show me the boot messages.
--
Debian GNU/Linux 2.2 is out! ( http://www.debian.org/ )
Email: Herbert Xu ~{PmV>HI~} <[EMAIL PROTECTED]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
#include
Michael Piefel wrote on Thu Apr 18, 2002 um 03:32:59PM:
> The Crash
> Well, simply, it was the wrong kernel. The guide recommends to install
> a new kernel with: apt-get install kernel-image-2.4.18-{386,586tsc,686}
> However, the system has a AMD K6. Of course it's his fault for choosing
We could rewrite it, with:
- mentioning -k6 and -k7
- recomendation to use 2.4.18-bf2.4 if you don't want to touch the boot
loader's config for now
> and loads the modules mentioned in /etc/modules. Of course, the disk
> driver had been compiled in before. I'm not sure what to do about this.
> kernel-image _could_ have warned about the missing SCSI driver, but I
The kernel-image's postinst shows a warning and simple hints about
needed changes in the boot loader. If you did not see a such screen,
file bugs. If you did, but ignored, then nobody can help you.
> No Net
> The appropriate alias for eth0 was missing. This is a case of getting
What dou you mean with "missing"?
> tired with all those "config file was changed by you or a script"
/etc/network/interfaces is not provided by a package. Though it must be
updated by ifupdown upgrade. If it was NOT upgraded, please file a bug.
Gruss/Regards,
Eduard.
--
Nutzt Du Win9x belächeln Dich die WinNT Helden, benutzt Du NT belächelt
Dich die Linux Front. Benutzt Du SuSE belächeln Dich die Debian Helden,
benutzt Du Debian, belächeln Dich die Selbstkompilierer, die wiederum
von den Unix-Helden belächelt werden. (irgendjemand in dcoulm)
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
Le ven 19/04/2002 à 09:57, Michael Piefel a écrit : > He just went for the best: 686. It seems that one doesn't work on k6 > anymore. That one wasn't ever meant for k6, which is not a 686 cpu. -- .''`. Josselin Mouette/\./\ : :' : [EMAIL PROTECTED] `. `'[EMAIL PROTECTED] `- Debian GNU/Linux -- The power of freedom signature.asc Description: PGP signature
Re: Potato->Woody kernel upgrade problems
Am 18.04.02 um 16:21:55 schrieb Bob Nielsen:
> > > a new kernel with: apt-get install kernel-image-2.4.18-{386,586tsc,686}
> > > However, the system has a AMD K6. Of course it's his fault for choosing
> > 386, 486, and 586 kernels work fine on a K6. When in doubt use a 386
> > kernel,
> > it'll run on any x86 system that runs Linux.
He just went for the best: 686. It seems that one doesn't work on k6
anymore.
> 2.4.18 has a k6 kernel-image, as well (also k7).
Of course, yes, but it wasn't mentioned in the docs. That's why I think
it'd be good to mention either all flavours or say that there are many
and you should look closely.
Bye,
Mike
--
|=| Michael Piefel
|=| Humboldt-Universität zu Berlin
|=| Tel. (+49 30) 2093 3831
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
Am 19.04.02 um 00:51:24 schrieb Josip Rodin: > > The old kernel was handcrafted. > he chose to trash all that and instead follow some random > newbie instructions? :) Well, there was a time when I compiled all my kernels myself, but then one day I got intrigued by our high-quality kernel-packages with their initrd... Since that time I always recommend the package to everyone, this makes me the one who is to blame. ;( Bye, Mike -- |=| Michael Piefel |=| Humboldt-Universität zu Berlin |=| Tel. (+49 30) 2093 3831 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
Martijn van [EMAIL PROTECTED]@Fri, 19 Apr 2002 09:50:42 +1000: > The solution to this is to stuff all your own aliases under something like > /etc/modules/mine or something like that. Then they will never been > overwritten (except if a package named "mine" decides it needs some modules > :) > Hmmm local-mine? -- *=-+-__ |[EMAIL PROTECTED]: _ Ugh! Nio2f says something: __ : http://www.lintux.cx/ |/ on the to somethine decider ste e \ ~~-+-=-+~+-=* -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
On Thu, Apr 18, 2002 at 03:32:59PM +0200, Michael Piefel wrote: > No Net > The appropriate alias for eth0 was missing. This is a case of getting > tired with all those "config file was changed by you or a script" > messages when very often you are sure that you didn't touch it. In this > case, a split to "common aliases" and "additional user aliases" in > seperate files would have helped, but I'm not sure about how sensible > this is. Perhaps three-way-diffs in dpkg would be helpful. The solution to this is to stuff all your own aliases under something like /etc/modules/mine or something like that. Then they will never been overwritten (except if a package named "mine" decides it needs some modules :) -- Martijn van Oosterhouthttp://svana.org/kleptog/ > Ignorance continues to thrive when intelligent people choose to do > nothing. Speaking out against censorship and ignorance is the imperative > of all intelligent people. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
On Thu, Apr 18, 2002 at 11:52:53PM +0200, Russell Coker wrote:
> On Thu, 18 Apr 2002 15:32, Michael Piefel wrote:
> > The Crash
> > Well, simply, it was the wrong kernel. The guide recommends to install
> > a new kernel with: apt-get install kernel-image-2.4.18-{386,586tsc,686}
> > However, the system has a AMD K6. Of course it's his fault for choosing
>
> 386, 486, and 586 kernels work fine on a K6. When in doubt use a 386 kernel,
> it'll run on any x86 system that runs Linux.
2.4.18 has a k6 kernel-image, as well (also k7).
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
On Thu, Apr 18, 2002 at 03:32:59PM +0200, Michael Piefel wrote:
> The guide recommends to install a new kernel with:
>apt-get install kernel-image-2.4.18-{386,586tsc,686}
>
> The old kernel was handcrafted.
Okay, so this person had a nice tailored kernel done, and instead of simply
upgrading the sources and running make oldconfig bzImage modules install
modules_install, he chose to trash all that and instead follow some random
newbie instructions? :)
--
2. That which causes joy or happiness.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Potato->Woody kernel upgrade problems
On Thu, 18 Apr 2002 15:32, Michael Piefel wrote:
> The Crash
> Well, simply, it was the wrong kernel. The guide recommends to install
> a new kernel with: apt-get install kernel-image-2.4.18-{386,586tsc,686}
> However, the system has a AMD K6. Of course it's his fault for choosing
386, 486, and 586 kernels work fine on a K6. When in doubt use a 386 kernel,
it'll run on any x86 system that runs Linux.
> No Disk
> The old kernel was handcrafted. The new one makes an initial RAM disk
> and loads the modules mentioned in /etc/modules. Of course, the disk
> driver had been compiled in before. I'm not sure what to do about this.
> kernel-image _could_ have warned about the missing SCSI driver, but I
> guess it's difficult. Perhaps an even bigger note about the perils of
> installing a new kernel?
It's a pity they de-installed the old kernel. If they'd kept it as
/vmlinuz.old then it would have been a lot less pain.
> No Net
> The appropriate alias for eth0 was missing. This is a case of getting
> tired with all those "config file was changed by you or a script"
> messages when very often you are sure that you didn't touch it. In this
> case, a split to "common aliases" and "additional user aliases" in
> seperate files would have helped, but I'm not sure about how sensible
> this is. Perhaps three-way-diffs in dpkg would be helpful.
It's best to create your own file for such things, then an upgrade won't
touch it. I've been bitten by this too. :(
--
If you send email to me or to a mailing list that I use which has >4 lines
of legalistic junk at the end then you are specifically authorizing me to do
whatever I wish with the message and all other messages from your domain, by
posting the message you agree that your long legalistic sig is void.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Potato->Woody kernel upgrade problems
A friend of mine who's quite well-versed in handling Linux, but not so
experienced with Debian, tried to upgrade his old Potato system to
Woody. It went quite well as he adhered to the instructions on
http://www.debian.org/releases/woody/i386/release-notes/ch-upgrading
Unfortunately, though, the system didn't boot anymore. (Quite a
show-stopper.)
There turned out to be a number of problems. He had installed a new
kernel, and encountered the following phenomena:
- crash
- no disk
- no net.
The Crash
Well, simply, it was the wrong kernel. The guide recommends to install
a new kernel with: apt-get install kernel-image-2.4.18-{386,586tsc,686}
However, the system has a AMD K6. Of course it's his fault for choosing
the wrong kernel, but I still think the documentation could be improved
here. It should either list all flavours (if k6 had been mentioned, it
would have been obvious which one was appropriate), or mention that
there a lot of kernels which you can list with dpkg -l kernel-image-2.4.18\*
No Disk
The old kernel was handcrafted. The new one makes an initial RAM disk
and loads the modules mentioned in /etc/modules. Of course, the disk
driver had been compiled in before. I'm not sure what to do about this.
kernel-image _could_ have warned about the missing SCSI driver, but I
guess it's difficult. Perhaps an even bigger note about the perils of
installing a new kernel?
No Net
The appropriate alias for eth0 was missing. This is a case of getting
tired with all those "config file was changed by you or a script"
messages when very often you are sure that you didn't touch it. In this
case, a split to "common aliases" and "additional user aliases" in
seperate files would have helped, but I'm not sure about how sensible
this is. Perhaps three-way-diffs in dpkg would be helpful.
So, these kernel problems were the only major things, everything else
quite harmless.
Bye,
Mike
--
|=| Michael Piefel
|=| Humboldt-Universität zu Berlin
|=| Tel. (+49 30) 2093 3831
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
