Kernel Upgrade Attempts to use Old Modules

banderson · 2015-08-25 21:20:23

I recently performed a full system upgrade including a Linux kernel upgrade from 4.1.5 to 4.1.6. Upon reboot, I could not mount my EFI partition due to the fact that the vfat module was missing. I downgraded to 4.1.5 and everything was okay. When I attempt to upgrade the kernel to 4.1.6, this is the mkinitcpio output:

>>> Updating module dependencies. Please wait ...
>>> Generating initial ramdisk, using mkinitcpio.  Please wait...
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'default'
  -> -k /boot/efi/vmlinuz-linux.efi -c /etc/mkinitcpio.conf -g /boot/efi/initramfs-linux.img
==> ERROR: '/lib/modules/4.1.5-1-ARCH' is not a valid kernel module directory
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'fallback'
  -> -k /boot/efi/vmlinuz-linux.efi -c /etc/mkinitcpio.conf -g /boot/efi/initramfs-linux-fallback.img -S autodetect
==> ERROR: '/lib/modules/4.1.5-1-ARCH' is not a valid kernel module directory

The kernel modules previously resided in /lib/modules/4.1.5-1-ARCH, but upon upgrading the kernel they were moved to /lib/modules/4.1.6-1-ARCH. However, it is still trying to find them at the previous location. Again, downgrading to 4.1.5 completes successfully.

I found this topic with a similar problem, but in that case it was a RAID issue, and I am not using RAID.

Any help would be greatly appreciated.

jasonwryan · 2015-08-25 21:22:20

This generally happens when /boot isn't mounted for the upgrade. Running

pacman -Q linux && uname -r

will confirm...

banderson · 2015-08-25 21:46:15

jasonwryan wrote:

This generally happens when /boot isn't mounted for the upgrade. Running
pacman -Q linux && uname -r
will confirm...

Thank you for your answer. However, /boot and /boot/efi are mounted when I attempt to upgrade:

[bryan@anderson /]$ lsblk
NAME               MAJ:MIN RM   SIZE RO TYPE  MOUNTPOINT
sda                  8:0    0 232.9G  0 disk  
├─sda1               8:1    0   450M  0 part  
├─sda2               8:2    0    99M  0 part  /boot/efi
├─sda3               8:3    0    16M  0 part  
├─sda4               8:4    0  48.3G  0 part  
├─sda5               8:5    0   250M  0 part  /boot
└─sda6               8:6    0 183.8G  0 part  
  └─vgroup0        254:0    0 183.8G  0 crypt 
    ├─vgroup0-swap 254:1    0     8G  0 lvm   [SWAP]
    └─vgroup0-root 254:2    0 175.8G  0 lvm   /
sr0                 11:0    1  1024M  0 rom

pacman -Q linux && uname -r

returns:

linux 4.1.5-1
4.1.5-1-ARCH

Upon attempted upgrade, it then returns:

linux 4.1.6-1
4.1.5-1-ARCH

Thanks again. I hope this clarifies the issue.

jasonwryan · 2015-08-25 21:49:36

Your ESP is /boot/EFI and your boot manager is systemd-boot? See https://wiki.archlinux.org/index.php/Sy … t#Updating

banderson · 2015-08-25 21:57:23

jasonwryan wrote:

Your ESP is /boot/EFI and your boot manager is systemd-boot? See https://wiki.archlinux.org/index.php/Sy … t#Updating

Correct. I just updated it as described on the wiki:

bootctl --path=/boot/efi update

And I continue to receive the original error, unfortunately.

Orobouros · 2015-08-25 22:57:12

I had a very similar problem. After updating to 4.1.6-1-ARCH my iptables shut down. System loaded okay but I had zero network access. Tried booting from another arch install (usb) and using arch-chroot and pacman to fix the problem. No avail. Downgrading to 4.1.5-1-ARCH fixed everything without a hassle. Even checked that boot was loaded up correctly; it was.

Was going to file a bug report but (a) guidelines say check the forums first, so here I am. And (b) I don't know if this is a linux kernel issue or an ARCH kernel issue.

is it common for each kernel update to cause somebody some problems? If not, should a bug report go to arch or linux?

Linux XXXXXX 4.1.5-1-ARCH #1 SMP PREEMPT Tue Aug 11 15:41:14 CEST 2015 x86_64 GNU/Linux

jasonwryan · 2015-08-25 23:03:05

Please don't hijack the thread: https://wiki.archlinux.org/index.php/Fo … _hijacking
Your issue is not the same (at least not as you have described it)...

banderson: paste the output of mount immediately before -Syyu.

Raynman · 2015-08-25 23:16:04

You seem to have configured mkinitcpio to use the kernel in /boot/efi. To make that work you would have to move the new kernel from /boot to /boot/efi after pacman extracts the files in the linux package, but before it runs the install script (which calls mkinitcpio). Otherwise, mkinitcpio will determine the kernel version by looking at the old kernel in /boot/efi and fail to find the modules directory.

You want mkinitcpio to look at the newly installed kernel in /boot and then afterwards you can move the files to /boot/efi. (Or make life easier by mounting the ESP at /boot or use a boot manager that supports more complex partitioning.)

Arphel · 2015-08-27 09:34:48

I have also a similar problem. Here's the output of mkinitcpio -p linux :

==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'default'
  -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux.img
==> Starting build: 4.1.6-1-ARCH
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [autodetect]
  -> Running build hook: [modconf]
  -> Running build hook: [block]
  -> Running build hook: [filesystems]
  -> Running build hook: [keyboard]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating gzip-compressed initcpio image: /boot/initramfs-linux.img
==> Image generation successful
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'fallback'
  -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux-fallback.img -S autodetect
==> Starting build: 4.1.6-1-ARCH
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [modconf]
  -> Running build hook: [block]
  -> Running build hook: [filesystems]
  -> Running build hook: [keyboard]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating gzip-compressed initcpio image: /boot/initramfs-linux-fallback.img
==> Image generation successful

But I can't even boot my system, dowgrading to linux to linux-4.1.5-1 does not change anything. Here's the errors :

init started BusyBox v1.21.1 (2013-7-28 11:02:26 EDT)
Failed to execute /init (error -2)
init:can't log to /dev/tty
starting pid 45 tty ``:`/etc/init.d/rcS`
can't open /dev/tty4 : No such file or directory
process `-/bin/sh` exited Scheduling for restart

Is there a way to fix this ?

Edit :

I've finally do a new installation of my system. It's work fine now, event with linux-4.16-1.

Last edited by Arphel (2015-08-27 14:14:51)

Arch Linux

#1 2015-08-25 21:20:23

Kernel Upgrade Attempts to use Old Modules

#2 2015-08-25 21:22:20

Re: Kernel Upgrade Attempts to use Old Modules

#3 2015-08-25 21:46:15

Re: Kernel Upgrade Attempts to use Old Modules

#4 2015-08-25 21:49:36

Re: Kernel Upgrade Attempts to use Old Modules

#5 2015-08-25 21:57:23

Re: Kernel Upgrade Attempts to use Old Modules

#6 2015-08-25 22:57:12

Re: Kernel Upgrade Attempts to use Old Modules

#7 2015-08-25 23:03:05

Re: Kernel Upgrade Attempts to use Old Modules

#8 2015-08-25 23:16:04

Re: Kernel Upgrade Attempts to use Old Modules

#9 2015-08-27 09:34:48

Re: Kernel Upgrade Attempts to use Old Modules

Board footer