On 17/06/2020 04:03, Leon Fauster via CentOS wrote:
Am 16.06.20 um 22:04 schrieb Fabian Arrotin:
On 16/06/2020 15:06, Leon Fauster via CentOS wrote:
Hi all,
I updated a Dell XPS laptop from CentOS 8.1 (1911) to 8.2 (2004).
Installed kernels are kernel-4.18.0-147.5.1.el8_1.x86_64 kernel-4.18.0-147.8.1.el8_1.x86_64 kernel-4.18.0-193.6.3.el8_2.x86_64
Unfortunately I can not boot into the latest kernel-4.18.0-193.6.3.el8_2.x86_64.
After grub2 screen I only see following line:
EFI stub: UEFI Secure Boot is enabled
Booting into the older kernel is still possible. The above line appears and after that the normal kernel output scrolls over the screen (rhgb quiet disabled).
Is the new kernel correctly signed?
What can I do?
-- Thanks Leon
Hi Leon,
Don't think that it's due to secureboot, as on my work laptop (thinkpad t490s), I have secureboot on, and kernel working fine.
OTOH, on my family laptop (also in secureboot mode), when I updated from 8.1.1011 to 8.2.2004, laptop became unresponsive during the microcode_ctl update (in scriptlet) and after that it auto-reset itself , so in the middle of the whole rpm transaction. I tried to recover it but it was to a point where it was faster to just reinstall from scratch with 8.2.2004, which I did ... and in gnome, everything was fine, etc (adding repo, pkgs) but then on the *same* kernel it was installed with, just tried a reboot, and nothing : grub shows menu, you select kernel and on upper left there is only cursor (fixed) and nothing happens ..
I'll try to diagnose what's the issue as actually that means troubles with family using that laptop :)
Hi Fabian,
an earlyprintk=efi kernel option shows a slowly executed kernel (at least the output). I disabled the early_microcode dracut option and rebuilded the initramfs image but no success in booting the kernel 4.18.0-193.6.3.el8_2.x86_64. Unfortunately no more time for more heuristics ...
-- Leon
I finally had reinstalled the laptop over pxe at home *but* pointing to kickstart repo (so GA content without updates, and so local mirror of http://mirror.centos.org/centos/8/BaseOS/x86_64/kickstart/), to ensure that microcode_ctl wouldn't be installed, and in some minutes laptop was back in action. Excluding it from updates and updated the rest and all is ok.
I've seen some people mentioning strange problems like this due to microcode, and it seems Ubuntu even had a second update a in row to fix issues : - https://usn.ubuntu.com/4385-1/ (introducing issue) - https://usn.ubuntu.com/4385-2/ (fixing the introduced issue)
All that was reported for centos 7 as we had the same issue there too (see https://bugs.centos.org//view.php?id=17452)
So for people impacted, I guess we have to wait for a new update to land, and excluding it from updates for now