On 6/7/25 21:13, George N. White III wrote:
On Sat, Jul 5, 2025 at 10:07 PM <fed...@eyal.emu.id.au 
<mailto:fed...@eyal.emu.id.au>> wrote:

    Today I did the usual 'dnf update' on a VM system.
    As part of the update I received kernel-6.15.4-200.fc42.x86_64


"VM System" is too vague: which VM?

QEMU/KVM.
Both the host and the guest are running the latest fedora 42.

    However, after the upgrades were done, the last console message was
             >>> Running post-transaction scriptlet: 
kernel-core-0:6.15.4-200.fc42.x86_64
    after which the system was hard locked (about 45m...).


Does "the system" refer to the VM or the host?

The VM froze hard. No problem on the host.

    Looking at the log (I remote rsyslog to my server) I see:

    Jul  6 09:16:53 e4 dracut[24606]: dracut-105-3.fc42
    [...]

    Jul  6 09:17:07 e4 kernel: block vda: the capability attribute has been 
deprecated.


I see this on my systems -- it seems to have been a change in the supported 
capabilities
for block devices.

Checking old logs shows this message logged since kernel 6.3.4 (Jun 2023) so it 
is a red herring...

    Jul  6 09:17:07 e4 dracut[24609]: *** Including module: 
kernel-modules-extra ***

    This was followed with a hard reset. The boot (6.15.4) failed. So I booted 
to the earlier kernel (6.15.3).


A hard reset you initiated, ot a spontaeous reset?

I left the guest frozen for about 45 minutes, and unable to get any response 
from the guest I forced a reset.

    I then removed and installed the latest kernel which did NOT create a new 
initramfs.

Failure to create the initramfs is a known issue.

Actually, while I can see that dracut did not complete, I do not know what else 
dnf was going to do.
So far I did not notice other problems.

    Solved by running:
             $ sudo dracut -f /boot/initramfs-6.15.4-200.fc42.x86_64.img 
6.15.4-200.fc42.x86_64
    which proceeded to completion:

    [...]Jul  6 10:19:32 e4 dracut[2939]: *** Creating initramfs image file 
'/boot/initramfs-6.15.4-200.fc42.x86_64.img' done ***

    The following 6.15.4 boot succeeded (it did a root fsck).

    1) Is the initial install failure a problem with the package (or dnf)?


Do you have out-of-tree modules?  See 
<https://bugzilla.redhat.com/show_bug.cgi?id=2333757 
<https://bugzilla.redhat.com/show_bug.cgi?id=2333757>>.

No out-of-tree modules. No dkms builds.


    2) Anyway, what is the 'block vda: the capability attribute has been 
deprecated.' about, and how do I correct it?


No correction is needed,

    3) Should the 'dnf remove/install' of the latest kernel run the 
post-transaction for that package?


The answer may depend on the reason for the failure to create the initrd.

Let me be more specific. Following an incomplete update, running 'dnf update' 
claimed there is nothing to do.
Even after having removed the latest (failed) kernel.

What is the correct way to have dnf repeat the kernel update in a complete way, 
including any scriplets and what not,
such that I do not need to find what is missing?

It seems that the package manager is not aware that the update did not complete 
properly. I expected it would know
how far it got.

TIA

--
George N. White III

--
Eyal at Home (fed...@eyal.emu.id.au)
--
_______________________________________________
users mailing list -- users@lists.fedoraproject.org
To unsubscribe send an email to users-le...@lists.fedoraproject.org
Fedora Code of Conduct: 
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: 
https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org
Do not reply to spam, report it: 
https://pagure.io/fedora-infrastructure/new_issue

Reply via email to