no boot if raid drive is mounted at boot from kernel 5.3.2 (and up)

I have a software raid 0 installed (4TB + 10TB) and the last kernel that boot fine my raid drive is 5.3.1-1-default @tumbleweed
From kernel 5.3.2 (and up) gives no access to the RAID0, the boot keep waiting for the drive.
If not in fstab then i don’t see the/dev/md/* anymore in Yast.

Who noticed this fault also?

First,
I hope you still have your older kernel that supports your RAID.
I don’t know about your problem, so my guess is that you should probably submit a bug report to https://bugzilla.opensuse.org.
There have been substantial changes recently to how RAID is implemented, i’d be speculating what happened.

Side note, I hope you’re not really running RAID0, that’s highly risky.
RAiD0 is striping without fault tolerance, so if any single disk in your array fails, your entire set is killed.
RAID0 should be implemented only where you might want speed but could care less if the files are lost completely… like most log files.

RAID1 is what most people typically set up with 2 disks, it provides fault tolerance by mirroring data to both drives so that if one disk fails, the other can continue to run until your faulty disk is replaced.

TSU

For the problem i found this information:
https://bbs.archlinux.org/viewtopic.php?id=249969Side note: I know what i’m doing with raid0, i’m using raid0 now 12 years with no fails happens for data loss. And i have a good backup solution for that if needed :wink:
I just want to have a big drive for my media data :-o

Bug opend
https://bugzilla.opensuse.org/show_bug.cgi?id=1155160

Then why did not you simply use linear? That allows easy growing of array in future as well.

LVM? maby in the futher, but raid0 in 12 years give me now “on a rolling os” the first trouble.

question, If you have LVM with 4TB + 10TB, and if the 4TB is in hardware fault, what data loss i have then, the both drives our just the 4TB?
How can i be 100% sure that i will have no data loss. Why is lVM the best choose then?

No, mdadm linear.

So who can tell me if mdadm (or other) can setup and define the raid layout that you have to load into the kernel parameter automatic in the futher?

for me kernel paramater “raid0.default_layout=2” works on kernels 5.3.3 later but i have my doubts in this whole story that the user have to know and manually load the correct raid layout…