Skip to main content
Topic: MDADM RAID1 Array Broke when Upgrading Kernel 5.18 (Read 757 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

MDADM RAID1 Array Broke when Upgrading Kernel 5.18

I setup a software (mdadm) RAID1 array on kernel 5.15 following the archwiki guide. Since upgrading to kernel 5.18 the virtual drive fails to mount on boot and the 'md#' partitions are gone when I run lsblk.

Predictably,
# mdadm --assemble --scan
prints: mdadm: unexpected failure opening /dev/md127

I thought maybe adding 'mdadm_udev' to the mkinitcpio.conf hooks and regenerating initramfs might help but it did nothing.

I'd expect I could downgrade my kernel and modules but was wondering if it looks like I did anything wrong or there was something else I could do.

Any help is appreciated.

Re: MDADM RAID1 Array Broke when Upgrading Kernel 5.18

Reply #1
I'm not sure if this is relevant, Ive removed nvidia-bl and regenerated initramfs.

From '/var/log/pacman.log'
[2022-06-01T18:08:30-0400] [ALPM] running '70-dkms-install.hook'...
[2022-06-01T18:08:30-0400] [ALPM-SCRIPTLET] ==> dkms install --no-depmod nvidia-bl/18.11.0 -k 5.18.0-artix1-1
[2022-06-01T18:08:32-0400] [ALPM-SCRIPTLET] Error! Bad return status for module build on kernel: 5.18.0-artix1-1 (x86_64)
[2022-06-01T18:08:32-0400] [ALPM-SCRIPTLET] Consult /var/lib/dkms/nvidia-bl/18.11.0/build/make.log for more information.
[2022-06-01T18:08:32-0400] [ALPM-SCRIPTLET] ==> WARNING: `dkms install --no-depmod nvidia-bl/18.11.0 -k 5.18.0-artix1-1' exited 10
[2022-06-01T18:08:32-0400] [ALPM-SCRIPTLET] ==> dkms install --no-depmod nvidia/470.129.06 -k 5.18.0-artix1-1
[2022-06-01T18:09:49-0400] [ALPM-SCRIPTLET] ==> depmod 5.18.0-artix1-1

Re: MDADM RAID1 Array Broke when Upgrading Kernel 5.18

Reply #2
clearly dkms doesn't like the new kernel
Cat Herders of Linux

Re: MDADM RAID1 Array Broke when Upgrading Kernel 5.18

Reply #3
I should have checked before posting, unfortunately I was getting those errors in 5.17 while RAID was working.

Re: MDADM RAID1 Array Broke when Upgrading Kernel 5.18

Reply #4

If it is any help, I have a similar RAID 1 setup and report no issues with the 5.18.0 kernel, just the standard kernel that is,

Re: MDADM RAID1 Array Broke when Upgrading Kernel 5.18

Reply #5
Unless you have time to spare trying to debug / investigate this I'd suggest just downgrading for now and trying again later. 
I've downgraded back to 5.17 myself for now. For me it's nvidia-vulkan-dkms (AUR) which failed to build with the new kernel.
The reason I'm using that is it's the only way I could find to get Far Cry 6 running with my Nvidia card. 

Saw the problem and just downgraded straight away and added linux and linux-headers to IgnorePkg in /etc/pacman.conf
Code: [Select]
IgnorePkg   = linux linux-headers
You get a warning every time you do a system update so eventually I'll get round to trying the newer kernel again. I imagine nvidia-vulkan-dkms just needs a patch so I'll monitor that.

TLDR:
Downgrade for now. Life's too short.