[RHEL kernel bug] XCP vm fails to boot after newest kernel applied.
-
How many of these that are failing have been upgraded from a previous version? Could it be something left over from EL7 or early EL8?
My Alma 8.10 base install is still running fine, did a yum update to apply a few more things and reboot and still working with the same kernel version above. But again, this was a clean fresh install, not something that's been running for a while.
-
FYI, there's a patch submitted to linux-stable (6.6 and earlier) but not yet in a stable release:
https://lore.kernel.org/stable/20250411160833.12944-1-jason.andryuk@amd.com/I guess we'll have to wait until this is picked up by Linux, then Red Hat will have to pick that as well.
-
@Greg_E said in XCP vm fails to boot after newest kernel applied.:
Ok, that might explain the difference.
Would a LEAPP from Rocky 8 up to Alma 9 be possible and solve the issue?
I did a (not LEAPP, but a migration script from Alama) from Rocky 8 to Alma 8, and it died. None of my Rocky 9's have had a problem so far, and of course end up with a completely different kernel.
-
@Greg_E They are all running for a while but none where upgrades from RHEL 7 to RHEL 8. I don't do LEAPPs that gives me more often errors than it works. What I did though on one is replacing the System VDI with a newer one, but that was Alma 8 to Alma 9 and that one has no issues.
-
@bberndt you Rocky/Alma 9 will be fine since it's a bug in the kernel for 8 only.
-
This is just a "me too" reply to indicate that I am also experiencing the boot failure immediately after upgrading AlmaLinux 8.10 to the latest kernel 4.18.0-553.50.1. Hopefully, the kernel fix will get integrated soon into the affected and popular RedHat 8 derivatives so that Alma and Rocky 8 et al. can continue to run on our favorite hypervisor. This is a bad one.
-
Strange that my fresh Alma 8.10 is still working after updates, makes me think k there is more to this than just the kernel.
-
@Greg_E Wondering what might be the difference? I am using the XenServer Linux tools version 8.4.0-1 and BIOS boot firmware.
-
I'm running the same guest tools.
Could it be a processor related issue? I'm running AMD v1756b in my lab, haven't had time to try in production that is Intel Silver.
-
@Greg_E Maybe it only affects Intel? That would be a new wrinkle. Perhaps other AMD users can confirm if the latest AlmaLinux 8 kernel runs fine on their AMD XCP-ng 8.3 hosts? Just read the above referenced RedHat bug report here: https://access.redhat.com/solutions/7116307 Sure enough, it says Xen + Intel in problem description. Check it out.
-
@Greg_E according to the Knowledge base of Redhawks it is den Witz Intel CPUs. So you AMD isn‘t affected
-
According to https://access.redhat.com/solutions/7116307 the bug is solved, but I don't get newer Kernels on Rocky yet.
-
Sad that even this documentation is locked behind a Redhat account, I can only read part of the problem description. Not going to make an account because it really doesn't affect me right now.
-
@Greg_E
I made a free account.
Resolution
The issue has been resolved with the errata: RHBA-2025:4337. Hence, update the kernel to kernel-4.18.0-553.51.1.el8_10 to fix the issue.id assume it arrives soon? I as well, haven't seen it yet.
-
It's now available, yes
After update, reboot will workedit: ha answered for RH, Alma and Rocky will follow soon I suppose
-
I just updated an Alma 8 and it has the .51 kernel now.
-
@Greg_E said in XCP vm fails to boot after newest kernel applied.:
I just updated an Alma 8 and it has the .51 kernel now.
Tried one of my Rocky's and still .50. I wonder where we can see the status of downstream fixes like that? anyone know? thanks.
-
This may sound stupid, but here goes...
Does Rocky need to update to .50 before it can update to .51?
I'd guess no, but I've been away from RHEL based stuff since the IBM rug pull. Only coming back because I may need it in the future as I jockey around trying to improve my paycheck.
-
@Greg_E said in XCP vm fails to boot after newest kernel applied.:
This may sound stupid, but here goes...
Does Rocky need to update to .50 before it can update to .51?
I'd guess no, but I've been away from RHEL based stuff since the IBM rug pull. Only coming back because I may need it in the future as I jockey around trying to improve my paycheck.
https://bugs.rockylinux.org/view.php?id=9307
so, prolly soon-ish. -
Alma must have dropped everything to push this out. I wonder what Oracle is doing, guessing they were hit by this too. I'd guess Rocky has this out pretty fast, they know it is important.
