Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #190737 > unrolled thread

Kernel problem?

Back to article view | Back to linux.debian.user

Contents

  Kernel problem? Rob Hurle <rob1940@gmail.com> - 2018-01-07 04:00 +0100
    Re: Kernel problem? Sven Joachim <svenjoac@gmx.de> - 2018-01-07 10:10 +0100
    Re: Kernel problem? Felix Miata <mrmazda@earthlink.net> - 2018-01-10 00:20 +0100
      Re: Kernel problem? David Wright <deblis@lionunicorn.co.uk> - 2018-01-10 22:20 +0100
        Re: Kernel problem? Felix Miata <mrmazda@earthlink.net> - 2018-01-10 23:00 +0100
          Re: Kernel problem? Sven Joachim <svenjoac@gmx.de> - 2018-01-10 23:40 +0100
    Re: Kernel problem? David Wright <deblis@lionunicorn.co.uk> - 2018-01-10 21:50 +0100
      Re: Kernel problem? deloptes <deloptes@gmail.com> - 2018-01-11 01:20 +0100
        Re: Kernel problem? Felix Miata <mrmazda@earthlink.net> - 2018-01-11 01:50 +0100
          Re: Kernel problem? David Wright <deblis@lionunicorn.co.uk> - 2018-01-11 20:00 +0100
            Re: Kernel problem? Felix Miata <mrmazda@earthlink.net> - 2018-01-11 21:40 +0100
              Re: Kernel problem? David Wright <deblis@lionunicorn.co.uk> - 2018-01-12 02:50 +0100
                Re: Kernel problem? deloptes <deloptes@gmail.com> - 2018-01-12 08:40 +0100
    Re: Kernel problem? Jimmy Johnson <field.engineer@gmail.com> - 2018-01-11 03:00 +0100

#190737 — Kernel problem?

[Multipart message — attachments visible in raw view] — view raw

Hi All,

  I'm running Stretch and yesterday I did my normal:

sudo apt-get update
sudo apt-get upgrade

It seemed to install vmlinuz-4.9.0-5-686-pae (and associated config and
image files, etc) in place of 4.9.0-4-686-pae versions.  Now the system
won't boot at all.  I have reverted to 4.9.0-4-686-pae and all is well.  My
questions are:

1.  Does anyone else see this?

2.  How can I revert without losing my working 4.9.0-4-686-pae system?  Can
I just change the soft links for initrd.img and vmlinuz at / to point to
the 4.9.0-4-686-pae versions instead of the 4.9.0-5-686-pae ones?  Will
this break something else for a future upgrade?

Any help much appreciated.  Thank you.

Cheers, Rob Hurle

-----------------------------
Rob Hurle
e-mail:    rob1940@gmail.com
Mobile:   +61 417 293 603 (Australia)
Telephone:  (02) 6236 3895
28 Mirrormere Rd, Burra, NSW 2620, Australia

[toc] | [next] | [standalone]

#190739

On 2018-01-07 13:58 +1100, Rob Hurle wrote:

>   I'm running Stretch and yesterday I did my normal:
>
> sudo apt-get update
> sudo apt-get upgrade
>
> It seemed to install vmlinuz-4.9.0-5-686-pae (and associated config and
> image files, etc) in place of 4.9.0-4-686-pae versions.  Now the system
> won't boot at all.  I have reverted to 4.9.0-4-686-pae and all is well.  My
> questions are:
>
> 1.  Does anyone else see this?

Bug #886485[1] talks about a hang at boot which may be the same what you
are seeing.

> 2.  How can I revert without losing my working 4.9.0-4-686-pae system?  Can
> I just change the soft links for initrd.img and vmlinuz at / to point to
> the 4.9.0-4-686-pae versions instead of the 4.9.0-5-686-pae ones?

Most bootloaders, including the default bootloader on PCs (grub), do not
use these symlinks at all.  To boot with the 4.9.0-4-686-pae kernel by
default, you can fiddle around with the grub configuration or uninstall
the 4.9.0-5-686-pae kernel.  In the latter case you would have to
downgrade or uninstall the linux-image-686-pae metapackage as well.

Cheers,
       Sven


1. https://bugs.debian.org/886485

[toc] | [prev] | [next] | [standalone]

#190864

Rob Hurle composed on 2018-01-07 13:58 (UTC+1100):

>   I'm running Stretch and yesterday I did my normal:

> sudo apt-get update
> sudo apt-get upgrade

> It seemed to install vmlinuz-4.9.0-5-686-pae (and associated config and
> image files, etc) in place of 4.9.0-4-686-pae versions.  Now the system
> won't boot at all.  I have reverted to 4.9.0-4-686-pae and all is well.  My
> questions are:

> 1.  Does anyone else see this?

> 2.  How can I revert without losing my working 4.9.0-4-686-pae system?  Can
> I just change the soft links for initrd.img and vmlinuz at / to point to
> the 4.9.0-4-686-pae versions instead of the 4.9.0-5-686-pae ones?  Will
> this break something else for a future upgrade?

> Any help much appreciated.  Thank you.                        


Here's a portion of /boot/ on one I just updated minutes ago:

lrwxrwxrwx  1       26 Jan  9 17:47 initrd -> initrd.img-4.9.0-5-686-pae
lrwxrwxrwx  1       26 Jan  9 17:47 initrd-cur -> initrd.img-4.9.0-5-686-pae
-rw-r--r--  1 17388979 Jan  9 17:45 initrd.img-4.9.0-4-686-pae
-rw-r--r--  1 17067283 Oct 21 03:00 .initrd.img-4.9.0-4-686-pae1
-rw-r--r--  1 17388194 Oct 21 04:51 .initrd.img-4.9.0-4-686-pae2
-rw-r--r--  1 17388979 Jan  9 17:45 .initrd.img-4.9.0-4-686-pae3
-rw-r--r--  1 17392772 Jan  9 17:44 initrd.img-4.9.0-5-686-pae
-rw-r--r--  1 17392772 Jan  9 17:44 .initrd.img-4.9.0-5-686-pae1
lrwxrwxrwx  1       26 Oct 21 03:18 initrd-prv -> initrd.img-4.9.0-4-686-pae
lrwxrwxrwx  1       23 Jan  9 17:47 vmlinuz -> vmlinuz-4.9.0-5-686-pae
-rw-r--r--  1  3643920 Dec 22 19:39 vmlinuz-4.9.0-4-686-pae
-rw-r--r--  1  3645296 Jan  4 06:12 vmlinuz-4.9.0-5-686-pae
lrwxrwxrwx  1       23 Jan  9 17:46 vmlinuz-cur -> vmlinuz-4.9.0-5-686-pae
lrwxrwxrwx  1       23 Oct 21 03:18 vmlinuz-prv -> vmlinuz-4.9.0-4-686-pae

Note the additional symlinks, and the in-place initrd backups. If what happened
to you happened to me, and I was unable to rebuild the 4.9.0.5 initrd or
anything else to solve the problem, I would purge the 4.9.0.5 kernel, delete the
-cur symlinks, and copy the -prv symlinks back to vmlinuz and initrd. If and
when an update brings a 4.9.0.6 or newer kernel, the 4.9.0.4 would be retained,
as 4.9.0.5 would have if it hadn't been purged.
-- 
"Wisdom is supreme; therefore get wisdom. Whatever else you
get, get wisdom." Proverbs 4:7 (New Living Translation)

 Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata  ***  http://fm.no-ip.com/

[toc] | [prev] | [next] | [standalone]

#190911

On Tue 09 Jan 2018 at 18:18:27 (-0500), Felix Miata wrote:

> Here's a portion of /boot/ on one I just updated minutes ago:
> 
> lrwxrwxrwx  1       26 Jan  9 17:47 initrd -> initrd.img-4.9.0-5-686-pae
> lrwxrwxrwx  1       26 Jan  9 17:47 initrd-cur -> initrd.img-4.9.0-5-686-pae
> -rw-r--r--  1 17388979 Jan  9 17:45 initrd.img-4.9.0-4-686-pae
> -rw-r--r--  1 17067283 Oct 21 03:00 .initrd.img-4.9.0-4-686-pae1
> -rw-r--r--  1 17388194 Oct 21 04:51 .initrd.img-4.9.0-4-686-pae2
> -rw-r--r--  1 17388979 Jan  9 17:45 .initrd.img-4.9.0-4-686-pae3
> -rw-r--r--  1 17392772 Jan  9 17:44 initrd.img-4.9.0-5-686-pae
> -rw-r--r--  1 17392772 Jan  9 17:44 .initrd.img-4.9.0-5-686-pae1
> lrwxrwxrwx  1       26 Oct 21 03:18 initrd-prv -> initrd.img-4.9.0-4-686-pae
> lrwxrwxrwx  1       23 Jan  9 17:47 vmlinuz -> vmlinuz-4.9.0-5-686-pae
> -rw-r--r--  1  3643920 Dec 22 19:39 vmlinuz-4.9.0-4-686-pae
> -rw-r--r--  1  3645296 Jan  4 06:12 vmlinuz-4.9.0-5-686-pae
> lrwxrwxrwx  1       23 Jan  9 17:46 vmlinuz-cur -> vmlinuz-4.9.0-5-686-pae
> lrwxrwxrwx  1       23 Oct 21 03:18 vmlinuz-prv -> vmlinuz-4.9.0-4-686-pae
> 
> Note the additional symlinks, and the in-place initrd backups.

I understand keeping backups, but what are the symlinks for?
Your backups are already renamed with suffixes, and also hidden
with dots.

> If what happened
> to you happened to me, and I was unable to rebuild the 4.9.0.5 initrd or
> anything else to solve the problem, I would purge the 4.9.0.5 kernel, delete the
> -cur symlinks, and copy the -prv symlinks back to vmlinuz and initrd. If and
> when an update brings a 4.9.0.6 or newer kernel, the 4.9.0.4 would be retained,
> as 4.9.0.5 would have if it hadn't been purged.

They're actually 4.9.0-4 and 4.9.0-5. I don't know how stable the
4.9 kernels are as I can't see into the future, but just to point
out that jessie was stuck on 3.16.0-4 from at least November 2014
(five months before its release) until this week. Wheezy was even
longer in the tooth; my 7.1 installation disk has 3.2.0-4 images
dated 9 June 2013 and I upgraded yesterday, so at that rate you'd
be waiting nearly five years.

Cheers,
David.

[toc] | [prev] | [next] | [standalone]

#190913

David Wright composed on 2018-01-10 15:12 (UTC-0600):

> On Tue 09 Jan 2018 at 18:18:27 (-0500), Felix Miata wrote:

>> Here's a portion of /boot/ on one I just updated minutes ago:

>> lrwxrwxrwx  1       26 Jan  9 17:47 initrd -> initrd.img-4.9.0-5-686-pae
>> lrwxrwxrwx  1       26 Jan  9 17:47 initrd-cur -> initrd.img-4.9.0-5-686-pae
>> -rw-r--r--  1 17388979 Jan  9 17:45 initrd.img-4.9.0-4-686-pae
>> -rw-r--r--  1 17067283 Oct 21 03:00 .initrd.img-4.9.0-4-686-pae1
>> -rw-r--r--  1 17388194 Oct 21 04:51 .initrd.img-4.9.0-4-686-pae2
>> -rw-r--r--  1 17388979 Jan  9 17:45 .initrd.img-4.9.0-4-686-pae3
>> -rw-r--r--  1 17392772 Jan  9 17:44 initrd.img-4.9.0-5-686-pae
>> -rw-r--r--  1 17392772 Jan  9 17:44 .initrd.img-4.9.0-5-686-pae1
>> lrwxrwxrwx  1       26 Oct 21 03:18 initrd-prv -> initrd.img-4.9.0-4-686-pae
>> lrwxrwxrwx  1       23 Jan  9 17:47 vmlinuz -> vmlinuz-4.9.0-5-686-pae
>> -rw-r--r--  1  3643920 Dec 22 19:39 vmlinuz-4.9.0-4-686-pae
>> -rw-r--r--  1  3645296 Jan  4 06:12 vmlinuz-4.9.0-5-686-pae
>> lrwxrwxrwx  1       23 Jan  9 17:46 vmlinuz-cur -> vmlinuz-4.9.0-5-686-pae
>> lrwxrwxrwx  1       23 Oct 21 03:18 vmlinuz-prv -> vmlinuz-4.9.0-4-686-pae

>> Note the additional symlinks, and the in-place initrd backups.

> I understand keeping backups, but what are the symlinks for?                    

Multiboot (menu.lsts and grub prompts). Installations retaining more than two
kernels have -prv2, -prv3, etc., and occasionally, -tst or -orig.

> Your backups are already renamed with suffixes, and also hidden
> with dots.

>> If what happened
>> to you happened to me, and I was unable to rebuild the 4.9.0.5 initrd or
>> anything else to solve the problem, I would purge the 4.9.0.5 kernel, delete the
>> -cur symlinks, and copy the -prv symlinks back to vmlinuz and initrd. If and
>> when an update brings a 4.9.0.6 or newer kernel, the 4.9.0.4 would be retained,
>> as 4.9.0.5 would have if it hadn't been purged.

> They're actually 4.9.0-4 and 4.9.0-5.

Dots, underscores and hyphens in package versions exist mainly to bloat
opportunities for human errors and break alphanumeric sorting in file manager
displays. :-)

In the case of stable Debian release kernel versioning at least, the .0 in both
seem to be entirely superfluous. Does any other integer ever appear in place of
that 0? Upstream according to https://www.kernel.org/ 4.9 is at 4.9.75.
-- 
"Wisdom is supreme; therefore get wisdom. Whatever else you
get, get wisdom." Proverbs 4:7 (New Living Translation)

 Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata  ***  http://fm.no-ip.com/

[toc] | [prev] | [next] | [standalone]

#190914

On 2018-01-10 16:58 -0500, Felix Miata wrote:

> In the case of stable Debian release kernel versioning at least, the .0 in both
> seem to be entirely superfluous. Does any other integer ever appear in place of
> that 0?

No.  The trailing .0 is only there to appease programs that do not work
with a two-digit kernel version, which is also the reason why Linus
numbers his releases as x.y.0(-rcN) rather than just x.y(-rcN) .

You can learn more about the numbers used in Debian kernels and what
they mean in the Debian Kernel Handbook[1].

> Upstream according to https://www.kernel.org/ 4.9 is at 4.9.75.

The upstream version of the Debian kernel is available with "uname -v":

,----
| $ uname -v
| #1 SMP Debian 4.14.12-2 (2018-01-06)
`----

Here 4.14.12 is the upstream version, and the -2 indicates the Debian
revision, not to be confused with the ABI version that appears in
"uname -r":

,----
| $ uname -r
| 4.14.0-3-amd64
`----

HTH,
    Sven


1. https://kernel-handbook.alioth.debian.org/ch-versions.html

[toc] | [prev] | [next] | [standalone]

#190908

On Sun 07 Jan 2018 at 13:58:25 (+1100), Rob Hurle wrote:
>   I'm running Stretch and yesterday I did my normal:
> 
> sudo apt-get update
> sudo apt-get upgrade
> 
> It seemed to install vmlinuz-4.9.0-5-686-pae (and associated config and
> image files, etc) in place of 4.9.0-4-686-pae versions.  Now the system

                    ↑↑↑↑↑↑↑↑↑↑↑ really? It's a different package so
it should install alongside the old one.

> won't boot at all.  I have reverted to 4.9.0-4-686-pae and all is well.  My
> questions are:
> 
> 1.  Does anyone else see this?
> 
> 2.  How can I revert without losing my working 4.9.0-4-686-pae system?  Can
> I just change the soft links for initrd.img and vmlinuz at / to point to
> the 4.9.0-4-686-pae versions instead of the 4.9.0-5-686-pae ones? Will
> this break something else for a future upgrade?

With this being a new package with a different name, both versions 4
and 5 should be in place in /boot.

Assuming you're booting with grub, it could be as simple as just
telling grub which kernel you want to be your default. The steps are:

Edit /etc/default/grub and change GRUB_DEFAULT=0 to GRUB_DEFAULT=saved
and run

# grub-mkconfig > /boot/grub/grub.cfg

which inserts a juggling trick into grub.cfg. That's a once-only
step. You can repeat the rest of this process at will.

You would normally now type

# grub-set-default 0

which would make grub boot by default the first entry: it writes

saved_entry=0

into /boot/grub/grubenv. This little file is the other part of the
trick that allows grub to remember something across reboots.

However, you don't want the first entry (0) to boot as this will
normally be the latest kernel. So you need to type

# grub-set-default 'foo>bar'

where foo and bar are, unfortunately, rather long strings¹.

Look in /boot/grub/grub.cfg for the submenu which is probably called
'Advanced options for Debian GNU/Linux'.
The string that you should write instead of "foo" is the
menuentry_id_option at the end of the line, which probably looks
something like gnulinux-advanced-7ccc1c1c-a690-418b-96c0-edcce6ebd3c1.

Now look under that submenu for the first menuentry that boots the
version 4 kernel. It's probably called
'Debian GNU/Linux, with Linux 4.9.0-4-686-pae'.
Again, the string that you should write instead of "bar" is the
menuentry_id_option which probably looks something like
gnulinux-4.9.0-4-686-pae-advanced-7ccc1c1c-a690-418b-96c0-edcce6ebd3c1.

After you have run grub-set-default with those strings, you can
examine the result with

$ cat /boot/grub/grubenv. The file is padded with # characters.
For example

$ cat /boot/grub/grubenv 
# GRUB Environment Block
saved_entry=gnulinux-advanced-7ccc1c1c-a690-418b-96c0-edcce6ebd3c1>gnulinux-4.9.0-4-686-pae-advanced-7ccc1c1c-a690-418b-96c0-edcce6ebd3c1
############################################################ … … …

Whenever you reboot, grub reads grubenv to find which menuentry
to boot. The code is in the first 40 lines or so of grub.cfg.

Because you've kept both versions of the kernel, you can retry
booting it whenever you come up with ideas as to what's wrong.
You'll also get any security updates etc as normal. If you
remove version 5, you may be unaware of any changes they make.

¹ disk partition UUIDs are rather long.

Cheers,
David.

[toc] | [prev] | [next] | [standalone]

#190916

David Wright wrote:

>> It seemed to install vmlinuz-4.9.0-5-686-pae (and associated config and
>> image files, etc) in place of 4.9.0-4-686-pae versions.  Now the system
> 
> ↑↑↑↑↑↑↑↑↑↑↑ really? It's a different package so
> it should install alongside the old one.

no, this is one and the same package - just a different debian revision - so
the previous gets replaced AFAIK

regards

[toc] | [prev] | [next] | [standalone]

#190917

deloptes composed on 2018-01-11 01:12 (UTC+0100):

> David Wright wrote:

>>> It seemed to install vmlinuz-4.9.0-5-686-pae (and associated config and
>>> image files, etc) in place of 4.9.0-4-686-pae versions.  Now the system

>> ↑↑↑↑↑↑↑↑↑↑↑ really? It's a different package so
>> it should install alongside the old one.

> no, this is one and the same package - just a different debian revision - so
> the previous gets replaced AFAIK                 

I think you missed this same thread post from yesterday:
https://lists.debian.org/debian-user/2018/01/msg00372.html
-- 
"Wisdom is supreme; therefore get wisdom. Whatever else you
get, get wisdom." Proverbs 4:7 (New Living Translation)

 Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata  ***  http://fm.no-ip.com/

[toc] | [prev] | [next] | [standalone]

#190953

On Wed 10 Jan 2018 at 19:48:57 (-0500), Felix Miata wrote:
> deloptes composed on 2018-01-11 01:12 (UTC+0100):
> 
> > David Wright wrote:
> 
> >>> It seemed to install vmlinuz-4.9.0-5-686-pae (and associated config and
> >>> image files, etc) in place of 4.9.0-4-686-pae versions.  Now the system
> 
> >> [spaces inserted]  ↑↑↑↑↑↑↑↑↑↑↑ really? It's a different package so
> >> it should install alongside the old one.
> 
> > no, this is one and the same package - just a different debian revision - so
> > the previous gets replaced AFAIK                 

Sorry, but evidently you don't.

> I think you missed this same thread post from yesterday:
> https://lists.debian.org/debian-user/2018/01/msg00372.html

I'm not sure that would help. The first half of that post showed
a very idiosyncratic /boot listing which seems to be customised
to support your own multibooting setup. I'm not sure whether it
would help or confuse the OP.

The second half didn't take any account of what the various parts
of the version numbers mean. They're not just there as a joke,
or an accident of punctuation.

The kernel-image maintainers update the Debian versions, and
*their* numbers, as quickly as possible. OTOH they strive to
change the ABI number as *slowly* as possible. Your advice was
wrong: you don't want to wait for an ABI change.

You should only revert to the old ABI version to get the system up
so that you can investigate why the new one won't work.¹

You should leave the new ABI version installed so that any
normal upgrades of it will take place. (It's always possible
that the reason the system didn't boot is because another bug
was present, and gets fixed in the usual manner.)

No one knows when the next ABI change will take place. It could
be tomorrow; it could be years away.

¹ There may be folks, too, who have yet to make kernel module
  changes to suit the new ABI.
  Disclaimer: sorry if you get hacked.

Cheers,
David.

[toc] | [prev] | [next] | [standalone]

#190959

David Wright composed on 2018-01-11 12:52 (UTC-0600):

> On Wed 10 Jan 2018 at 19:48:57 (-0500), Felix Miata wrote:

>> deloptes composed on 2018-01-11 01:12 (UTC+0100):

>> > David Wright wrote:

>> >>> It seemed to install vmlinuz-4.9.0-5-686-pae (and associated config and
>> >>> image files, etc) in place of 4.9.0-4-686-pae versions.  Now the system

>> >> [spaces inserted]  ↑↑↑↑↑↑↑↑↑↑↑ really? It's a different package so
>> >> it should install alongside the old one.

>> > no, this is one and the same package - just a different debian revision - so
>> > the previous gets replaced AFAIK                 

> Sorry, but evidently you don't.

Who didn't get what?

>> I think you missed this same thread post from yesterday:
>> https://lists.debian.org/debian-user/2018/01/msg00372.html

> I'm not sure that would help. The first half of that post showed
> a very idiosyncratic /boot listing which seems to be customised
> to support your own multibooting setup. I'm not sure whether it
> would help or confuse the OP.

I expected it to. Here's the same listing with nonessential lines excised:

-rw-r--r--  1 17388979 Jan  9 17:45 initrd.img-4.9.0-4-686-pae
-rw-r--r--  1 17392772 Jan  9 17:44 initrd.img-4.9.0-5-686-pae
-rw-r--r--  1  3643920 Dec 22 19:39 vmlinuz-4.9.0-4-686-pae
-rw-r--r--  1  3645296 Jan  4 06:12 vmlinuz-4.9.0-5-686-pae

Clearly, 4.9.0-4 and 4.9.0-5 pae kernels are present, the very kernels in the OP
that, as I read, deloptes subsequently claimed were "one and the same package"
(could not coexist).
-- 
"Wisdom is supreme; therefore get wisdom. Whatever else you
get, get wisdom." Proverbs 4:7 (New Living Translation)

 Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata  ***  http://fm.no-ip.com/

[toc] | [prev] | [next] | [standalone]

#190964

On Thu 11 Jan 2018 at 15:35:36 (-0500), Felix Miata wrote:
> David Wright composed on 2018-01-11 12:52 (UTC-0600):
> 
> > On Wed 10 Jan 2018 at 19:48:57 (-0500), Felix Miata wrote:
> 
> >> deloptes composed on 2018-01-11 01:12 (UTC+0100):
> 
> >> > David Wright wrote:
> 
> >> >>> It seemed to install vmlinuz-4.9.0-5-686-pae (and associated config and
> >> >>> image files, etc) in place of 4.9.0-4-686-pae versions.  Now the system
> 
> >> >> [spaces inserted]  ↑↑↑↑↑↑↑↑↑↑↑ really? It's a different package so
> >> >> it should install alongside the old one.
> 
> >> > no, this is one and the same package - just a different debian revision - so
> >> > the previous gets replaced AFAIK                 
> 
> > Sorry, but evidently you don't.
> 
> Who didn't get what?

Here are the lines whose prefix-quoting displays the answer:

> >> deloptes composed on 2018-01-11 01:12 (UTC+0100):
> >> > no, this is one and the same package - just a different debian revision - so
> >> > the previous gets replaced AFAIK [As Far As I Know]

     ↑___the extra > indicates which lines deloptes wrote.

IOW

who      deloptes …
what     … was mistaken in distinguishing between "package name"
         and "Debian version" in the excerpt posted by the OP.

> >> I think you missed this same thread post from yesterday:
> >> https://lists.debian.org/debian-user/2018/01/msg00372.html
> 
> > I'm not sure that would help. The first half of that post showed
> > a very idiosyncratic /boot listing which seems to be customised
> > to support your own multibooting setup. I'm not sure whether it
> > would help or confuse the OP.
> 
> I expected it to. Here's the same listing with nonessential lines excised:
> 
> -rw-r--r--  1 17388979 Jan  9 17:45 initrd.img-4.9.0-4-686-pae
> -rw-r--r--  1 17392772 Jan  9 17:44 initrd.img-4.9.0-5-686-pae
> -rw-r--r--  1  3643920 Dec 22 19:39 vmlinuz-4.9.0-4-686-pae
> -rw-r--r--  1  3645296 Jan  4 06:12 vmlinuz-4.9.0-5-686-pae

That's much clearer. But then your post went on to talk about
deleting and copying the (duly noted) additional symlinks.
These (as seen in your earlier post) are things that you've made
for yourself for a purpose that's clear to you, but probably not
clear to most people running a more conventional booting setup.

> Clearly, 4.9.0-4 and 4.9.0-5 pae kernels are present, the very kernels in the OP
> that, as I read, deloptes subsequently claimed were "one and the same package"
> (could not coexist).

That's right: deloptes claim was mistaken, which I pointed out
because errors of fact need correcting.
I didn't mean to mystify you (or anybody else).

It follows that if you don't dist-upgrade (apt-get's sense) when the
ABI is bumped, the new package won't be installed. If you don't have
the generic metapackage installed, you might not even be aware that
the ABI *has* been bumped—you just stop getting kernel upgrades for
no discernible reason. AIUI (do correct me) only the latest ABI's
version gets updated after the change.

Cheers,
David.

[toc] | [prev] | [next] | [standalone]

#190968

David Wright wrote:

> That's right: deloptes claim was mistaken, which I pointed out
> because errors of fact need correcting.
> I didn't mean to mystify you (or anybody else).

Sorry I agree with you, I also learned something, thanks.

[toc] | [prev] | [next] | [standalone]

#190918

On 01/06/2018 06:58 PM, Rob Hurle wrote:
> Hi All,
> 
>    I'm running Stretch and yesterday I did my normal:
> 
> sudo apt-get update
> sudo apt-get upgrade
> 
> It seemed to install vmlinuz-4.9.0-5-686-pae (and associated config and
> image files, etc) in place of 4.9.0-4-686-pae versions.  Now the system
> won't boot at all.  I have reverted to 4.9.0-4-686-pae and all is well.  My
> questions are:
> 
> 1.  Does anyone else see this?
> 
> 2.  How can I revert without losing my working 4.9.0-4-686-pae system?  Can
> I just change the soft links for initrd.img and vmlinuz at / to point to
> the 4.9.0-4-686-pae versions instead of the 4.9.0-5-686-pae ones?  Will
> this break something else for a future upgrade?
> 
> Any help much appreciated.  Thank you.

I had a problem with that kernel on first boot, so I booted the old 
kernel with no problem and then tried the new kernel again to see if I 
could get a hint at what the problem was and it booted and the desktop 
started with no problem and again with no problem, go figure..I'm still 
booting the new kernel(Linux jimmy-1 4.9.0-5-amd64 #1 SMP Debian 
4.9.65-3+deb9u2 (2018-01-04) x86_64 GNU/Linux) with no problem.  So if 
you have not tried to reboot with the new kernel, maybe cross your 
fingers and give it another try or two.

Cheers,
-- 
Jimmy Johnson

Debian Stretch - KDE Plasma 5.8.6 - AMD A8-7600 - EXT4 at sda6
Registered Linux User #380263

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.user

csiph-web