Bug 1227236

Summary:	System not able to correctly read RAID (1) devices with kernel >=6.9
Product:	[openSUSE] openSUSE Tumbleweed	Reporter:	Toomey Timey <timeygaming69>
Component:	Kernel:Storage	Assignee:	Kernel Bugs <kernel-bugs>
Status:	NEW ---	QA Contact:	E-mail List <qa-bugs>
Severity:	Normal
Priority:	P5 - None	CC:	colyli, hare, tiwai
Version:	Current
Target Milestone:	---
Hardware:	x86-64
OS:	openSUSE Tumbleweed
Whiteboard:
Found By:	---	Services Priority:
Business Priority:		Blocker:	---
Marketing QA Status:	---	IT Deployment:	---

Description Toomey Timey 2024-07-01 13:36:10 UTC

Overview:

Updating to a kernel that is >=6.9.x leads to the system being unable to correctly read and assemble MDADM RAID devices, works fine with older Tumbleweed kernels (e.g. 6.8.x kernels)

Steps to Reproduce:

1. Create a RAID 1 device with MDADM, with LVM volumes within, formatted with BTRFS
2. Update to newer kernel version in TW, higher than 6.9

Actual Results:

System fails to correctly assemble MD devices, only some LVs are visible and mountable on 1 disk. Secondary disk does not appear to be part of the MD array, and LVs do not appear.

Replugging the devices cause other LVs that could not be read during boot time to appear, but are not mountable and readable. Original readable LVs are not visible after replugging.

Expected Results:

All created MD devices and LVs should be assembled and mountable correctly.

Build Date & Hardware:

I am unsure about the specific build, but I have tried 2 6.8 kernels (6.8.3, 6.8.5) by rolling back the system, and it functions correctly. Updating to a snapshot with a kernel >6.9 causes the issue. Updating to a newer snapshot with the kernel-default package appears to not trigger the issue.

Additional Builds and Platforms: Whether or not the bug takes place on other platforms (or browsers, if applicable).

Occurs on an installation with MicroOS, and Slowroll with kernel >6.9, have not tested other distributions.

Additional Information:

System is a X1 Carbon with 2x Seagate Ironwolf drives attached via a USB enclosure. Functioned for 2 years, and tested with new, identical USB enclosures and other laptops to confirm the issue, which is replicated identically.

When the devices can be mounted correctly with older kernels, SMART data and BTRFS scrubbing did not identify any issues.

Comment 1 Toomey Timey 2024-07-02 13:14:21 UTC

ADDITIONAL INFO:

Using Fedora 40 with the 6.8.5 kernel does not present the issue either. Updating to 6.9.6 kernel presents the issue as well. 

I have tried the kernel-vanilla package in Tumbleweed, and it seems to have the same issue where the disks are not properly read. Using the kernel-longterm package seems to not present the issue either (since its a 6.6 kernel). 

Additionally, I have found that grub seems to fail to start the system when the disks are plugged in, with the following message:

error: ../../grub-core/disk/efi/efidisk.c:612:failure reading sector 0xffffffe8 from 'hd1'.

Doing my own research suggests that it may be a symptom of a failing disk, but I am not sure why the behaviour changes between snapshots.