13.1 kernel 3.11.10-7.1 update caused serious boot problems

[andromeda]

Today I accepted to install a few security updates, amongst others the 3.11.10-7.1 kernel update that was announced yesterday:
http://lists.opensuse.org/opensuse-security-announce/2014-02/msg00003.html

During the updates I got no error messages. A message said a reboot was required so I rebooted.

Problems:
I can’t boot with kernel 3.11.10-7.1, I just get some error messages amongst one that sais “Kernel panic”. It doesn’t help to reboot again, it gives error messages every time.
I can still boot with the old kernel 3.11.6-4.1 in the GRUB menu, but things don’t work like they should.
The ethernet connection seems to be working, I get an IP address and can access the router, but the internet connection doesn’t seem to be working at all! I can’t use firefox or konqueror to connect to anything on the internet, the graphical autoupdate service doesn’t show up at all and YAST can’t connect to the repositories. I can use the internet on my windows computer, my android phone, and my Ubuntu computer, so the internet problem is with the OpenSUSE computer only.
YAST complains about the /boot being almost full. The size of /boot is 148MB, usage 97%. /boot contains both new and old kernel.

Background:
I recently switched from Ubuntu because of incompatibilities between FDE and their upgrade process which made the computer unusable twice, and I got fed up with their incompetence related to disk encryption and looked for alternatives. I have used SUSE or OpenSUSE before at many different occasions and had only good memories when it came to stability of the OS. Now this happens, and I’m very disappointed as I thought OpenSUSE was a bit better than Ubuntu when it came to stability and security. This was just a security update within the 13.1 version, not even an upgrade like in Ubuntu.

Possible cause:
If this problem is entirely related to the fact that /boot is too small, then I have to blame the OpenSUSE 13.1 installer that gave me this /boot partition size. I used automatic partitioning to not risk creating anything that was incompatible with OpenSUSE. In that case there would need to be a new DVD 13.1.1 created that doesn’t create too small /boot partitions during automatic partitioning of a full harddrive to correct this problem, rather than a patch to the kernel update.

Recovery:
How can I get a fully functional OpenSUSE after today’s disaster? What files can I erase, edit or manually replace from a repo to a usb to get the system fully functioning again?
Do I need to reinstall OpenSUSE again?
In that case how do I do a good partitioning of the hard drive that doesn’t break Linux when there are kernel updates?
The current partitioning created by the OpenSUSE installer is:
/dev/sda1 156MB EFI boot FAT /boot/efi
/dev/sda2 157MB Linux native Ext4 /boot
/dev/sda3 465GB with encrypted LVM with a lot of different things under it (created manually by me or automatically by the OpenSUSE installer)

Please help me recover my installation after the mess this involuntary kernel update created. I would be really grateful if anyone helped me.

[ratzi]

@andromeda

Today I accepted to install a few security updates, amongst others the 3.11.10-7.1 kernel update …

Hi !

I wonder in which way/how you did manage to remember the number, 3.11.10-7.1 …
You don’t seem to be an unexperienced user.
My advice:
Install openSUSE 13.1 again,
wait a couple days (while enjoing your system),
and then try to do the updates again.

Good luck

[nrickert]

@andromeda

YAST complains about the /boot being almost full. The size of /boot is 148MB, usage 97%. /boot contains both new and old kernel.

I’m guessing that is the problem.

Here’s what I would do. Your decision whether to try it.

1. Delete vmlinuz-3.11.10-7-desktop and initrd-3.11.10-7-desktop from “/boot”. That’s to free up space.
2. Run “mkinitrd”. That should rebuild the “initrd” for the older kernel.
3. Reboot to the older kernel.
4. If that all works, then use Yast software manager
5. delete plymouth. That way, you will need less space in “/boot”
6. Run “mkinitrd” again, and notice the smaller “initrd” size.
7. Start Yast again, and reinstall the 3.11.10.7 kernel.

Hopefully, that will fix all.

If you instead choose to reinstall, then don’t install plymouth. Hmm, that’s not a choice with a live installer, but it is a choice in the software selection section for the DVD installer. In the software selection, search for “plymouth” and uncheck.

I’m using less than around 70M for “/boot” with the two kernels, but without plymouth.

[ratzi]

@ratzi

… wait a couple days (while enjoing your system),
and then try to do the updates again.

Sorry for that! I was a bit tired.

@andromeda

I used automatic partitioning to not risk creating anything that was incompatible with OpenSUSE.

Have a look at http://forums.opensuse.org/showthread.php/494513-Looking-for-advice-on-installing-on-UEFI-desktop?p=2617350#post2617350
and the whole thread there.

Increasing the size of your EFI boot partition does not lead to incompatibilities.

And that is common to all OSs that are UEFI booting - even windows.

@andromeda

I have used SUSE or OpenSUSE before at many different occasions and had only good memories when it came to stability of the OS.

That is my experience as well, for several years already.

@andromeda

Possible cause:
If this problem is entirely related to the fact that /boot is too small, then I have to blame the OpenSUSE 13.1 installer that gave me this /boot partition size. I used automatic partitioning to not risk creating anything that was incompatible with OpenSUSE. In that case there would need to be a new DVD 13.1.1 created that doesn’t create too small /boot partitions during automatic partitioning of a full harddrive to correct this problem, rather than a patch to the kernel update.

I would say you’re not wrong.

But you’re not absolutely right either.

A larger /boot partition will help.

But if there are enough kernel updates, then even that larger /boot (or EFI boot partition) may get bust again.

To create boot entries for a new kernel, however, helps to keep the system going even if a new kernel
wouldn’t work for your system. That’s the advantage (you experienced the disadvantage).

OK, read through the thread linked above, that was started by oldcpu.

Then have another look at the suggestions of nrickert above

@nrickert

I’m guessing that is the problem.

Here’s what I would do. Your decision whether to try it.

1. Delete vmlinuz-3.11.10-7-desktop and initrd-3.11.10-7-desktop from “/boot”. That’s to free up space.
2. Run “mkinitrd”. That should rebuild the “initrd” for the older kernel.
3. Reboot to the older kernel.
4. If that all works, then use Yast software manager
5. delete plymouth. That way, you will need less space in “/boot”
6. Run “mkinitrd” again, and notice the smaller “initrd” size.
7. Start Yast again, and reinstall the 3.11.10.7 kernel.

Hopefully, that will fix all.

Even if that works, then, in my view, the next kernel update still may cause the same problems as you experienced now.

So perhaps better:

Enlarge your /boot (EFI boot partition).

And re-install openSUSE to get a clean system that is prepared for a few more kernel updates.

Good luck!
Mike

[ratzi]

sh… again

with /boot you didn’t mean the UEFI boot partition, right?

Still the argument remains the same: if you don’t increase the size of /boot
then the next kernel updates will lead to the same problems that you already encountered.

[gogalthorp]

You need space for 3kernels + misc stuff on a separate /boot Unless you change how zypper keeps kernels

I’d say you need a minimum of 250 meg. I’d go 500 meg just to be on the safe side

You need to understand there is a difference from a /boot and a /boot/efi partition