kernel panic after upgrade to kernel 4.4.136-56-default

I got the latest upgrade today (2018-06-18) for my opensuse LEAP 42.3, which contained a kernel upgrade to 4.4.136-56. However this doesn’t boot but produces a kernel panic and halts. Fortunately, I could boot into the previous 4.4.132 version.

Charles

This has happened to a colleague’s workstation, but OTOH mine upgraded and rebooted fine. The two setups differ in some major respects though. The one where the boot succeeded is a new-style default setup: the whole system (including /boot) is on a btrfs filesystem created on a primary partition, with a separate vfat-formatted partition for /boot/efi. The one that didn’t boot seems to be configured for a legacy (i.e. non-EFI) boot. It has a dedicated partition for /boot and the rest of / is on an LVM logical volume (both formatted ext4).

My system is also an EFI system where / includes the /boot and /boot/efi is on a separate filesystem which is vfat formatted. The root (including /boot) is an ext4 system and is a logical volume where the whole volume group is based on a s/w raid1. The first lines of the panic are

  2.264548] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
  2.264548]
  2.264645] CPU: 0 PID: 1 Comm: init Not tainted 4.4.136-56-default #1
  2.264709] Hardware name: System manufacturer System Product Name/Z170-P, BIOS 3301 02/10/2017
.
.
.

above typed from picture taken of screen, so maybe contains typos.

Charles

I have exactly this problem too after applying the latest kernel update, this is my setup:

/dev/sda8        30G   11G   18G  38% /
/dev/sda1       496M   66M  431M  14% /boot/efi
/dev/sda9       419G  316G  103G  76% /home

just a pretty standard install but choosing EXT4 as the default FS. After some frantic searching to see whether the hardware is still under warranty I did eventually think to check whether the older kernels would boot, and indeed I’m able to use this system once again with the older kernel.

I guess if not already done this should be put into bugzilla, unfortunately bugzilla.opensuse.org seems inaccessible at present.

Andy

Bugzilla seems to be working now, but I don’t know how to report this problem (never used bugzilla before). If it hasn’t reported yet (I think), could somebody do report this?

Another related question: after the update, it appears there is only 1 older kernel version still available. I am afraid that in case a correction would also fail, I will be stuck with 2 non-bootable kernels. Is there a way to retain the working kernel (by making it sticky, or to allow more than 2 kernels)?

Charles

There appears an existing bug, I added a note linking to this forum thread:

https://bugzilla.opensuse.org/show_bug.cgi?id=1094203

With regard to keeping older kernels, I’ve not tried myself but this post describes how to do it:

https://en.opensuse.org/SDB:Keep_multiple_kernel_versions

and still seems relevant despite being ‘tested on 12.1’.

Personally, I shall be avoiding all updates until the problem is at least claimed to be fixed.

thanks. I have modified my /etc/zypp/zypp.conf file to retain more kernels and to retain the current working 4.4.132 kernel. Will see if that works with the next upgrade.

I have just installed a new upgrade which has installed the 4.4.138-59-default kernel and this one seems to be running fine with no boot problems.

I checked my grub entries and it has retained indeed 3 different versions: the new one, the faulty 4.4.136 and the working 4.4.132.

Charles