Hey everyone! I'm facing a frustrating issue with Debian 12 on several machines where I'm using RAID1 (mdadm) for my root filesystem. When I try to shut down or reboot, I often encounter a kernel panic related to the 'md_notify_reboot' module. I've set up the RAID configuration using the Debian installer, and it's been problematic across various setups: I've tested it on different types of storage—NVMe drives, SATA SSDs—and three different motherboards, including a Z390, Z490, and Z890. Surprisingly, when I tried a similar setup on Fedora and AlmaLinux, everything worked seamlessly. The kernel panic doesn't happen every time, but roughly one out of every five shutdowns or reboots. I've even tried using a VM with emulated UEFI, and the same issue occurred there. I've looked into possible hardware malfunction, but since everything works fine on other distributions, it seems like a Debian-specific problem. Is anyone else experiencing this? Any ideas or solutions?
4 Answers
I've had my share of dealing with mdadm on Debian, and while I've never run into this exact kernel panic, I do sometimes experience longer shutdown times because of busy devices. It might be worth checking if there's something unusual in your setup or if you're dealing with any flaky hardware like drives or RAM. Since you've ruled out defective devices by testing different configurations and even in a VM, maybe try isolating the problem further. Consider doing a minimal installation without the typical setup to see if the panic still occurs. Switching the shutdown method in systemd for a more classic approach might also help you pinpoint the issue.
You're not alone in this; I've experienced kernel panic during reboots with Debian 12 and RAID1 too. I heard there's a patch available for kernel 6.15 that might solve this. I found a bug report linked to this issue, so it might be a kernel problem rather than hardware. Make sure you report your problems on that bug report; more attention might speed up a fix!
I’d love to see that bug report! If you’re testing with the patched kernel, let us know if that resolves your issue. Stuff like this really makes a difference for the community.
I appreciate all the feedback! I'll consider blacklisting the md_notify_reboot module and keep an eye out for any updates. It's reassuring to know I'm not the only one dealing with this weird bug on Debian 12. Cheers, everyone!
Definitely report back on what you find! We're all eager to learn if the blacklist or any patch works for you.
Turning off write caching on the RAID controller in BIOS has worked for some folks with similar issues. It might be worth a shot if you haven't tried it yet.
Thanks for the suggestions! I’ll definitely try a minimal install and see how that goes. Switching shutdown methods sounds interesting too. I'll keep you posted on what I find.