Migrating from centos 7.9.2009 to almalinux - STUCK

Hi all,
I have a current install of Centos 7.9.2009 that was working perfectly - but of course, now EOL. So, I am following the upgrade guide: https://wiki.almalinux.org/elevate/ELevating-CentOS7-to-AlmaLinux-9.html#upgrade-centos-7-to-almalinux-8

All seemed to be pretty good, after running the ```
sudo leapp preupgrade

There were NO errors except this one:

============================================================
ERRORS

2024-07-28 07:38:46.553248 [ERROR] Actor: check_installed_kernels
Message: Cannot find any installed kernel signed by Red Hat.

============================================================
END OF ERRORS

So, now I’m stuck as I’ve searched google and the only answer seem to be that I am not running a proper Centos Kernel? Not sure how that can be?

Can someone please point me in the right direction to fix this error as so I can upgrade?
thank you

you may need to install kernel-core not just kernel.

maybe you’re running elrepo kernel or something, what does this say?:

rpm -qa | grep kernel

CentOS 7 did not have separate kernel-core package.

And uname -r (for currently running kernel, although I suspect that the script looks at installed packages).


Personally, I would rather do a fresh install and transfer user data. (Granted, some setups make it a challenging path.)

I get:

kernel-tools-3.10.0-1160.119.1.el7.x86_64
kernel-4.9.241-37.el7.x86_64
kernel-xen-release-7-13.noarch
kernel-tools-libs-3.10.0-1160.119.1.el7.x86_64
php-symfony-http-kernel-2.8.12-2.el7.noarch
abrt-addon-kerneloops-2.1.11-60.el7.centos.x86_64
kernel-xen-headers-5.4.69-3.el7xen.x86_64

Uname -r shows:

4.9.241-37.el7.x86_64

I can’t do a fresh install - have to upgrade in place!

These are for latest CentOS 7 kernel, but they are not a kernel:

kernel-tools-3.10.0-1160.119.1.el7.x86_64
kernel-tools-libs-3.10.0-1160.119.1.el7.x86_64

This is the (only) kernel that you have and run:

kernel-4.9.241-37.el7.x86_64

and then there is:

kernel-xen-release-7-13.noarch
kernel-xen-headers-5.4.69-3.el7xen.x86_64

Websearch hints that there has been a Xen kernel 4.9.241-37.el7.x86_64, also built by CentOS Virtualization SIG. Since that was built by “third-party”, we got an explanation for the error message.

I don’t recall ever using Xen virtualization. I think I did try to grok it, failed, and chose libvirt/KVM instead. IIRC, both the host and guest did use a Xen-version of kernel. If true, then (A) reverting to CentOS kernel in the guest would be an issue, and (B) one has to get a Xen-kernel for AlmaLinux. As said, I don’t know a thing about those.


Note: One can check with yum where did packages were installed from. For example:

yum --disablerepo=* list installed kernel\*

Thought so. However, do consider a hypothetical worst case: something fatal happens to the server – flat tire, earthquake, a terrible flood, locusts. Do you have any contingency plan for such scenarios?

That’s quite a mess - so you’re running a xen kernel, but have headers/libs/tools for 2 completely different kernels!

Maybe install the proper “redhat” kernel 3.10 and boot into that?

https://vault.centos.org/7.9.2009/updates/x86_64/Packages/kernel-3.10.0-1160.119.1.el7.x86_64.rpm

thanks - will that break my setup though if I try that please?

not sure, i don’t know if xen needs a specific kernel, that’s what test environments are for.

anyone else have an idea on this please? I’m stuck and really need to do this upgrade.

It should be possible to install the regular kernel. Then you would have kernel-4.9.241-37.el7 and kernel-3.10.0-1160.119.1.el7

However, that gets us to GRUB. It loads one kernel. Install of 3.10.0 may change the default from 4.9.241 into the 3.10.0.

If the 3.10.0 cannot boot – if a Xen kernel is required – then you do need recovery.

GRUB may (or may not) be set to briefly show the installed kernels to choose from. If you cannot access that menu, then you cannot select a different (non-default) kernel.
I.e. cannot boot with 4.9.241. Unbootable system counts as “broken”?

GRUB has some means to “try entry once”, but I have no experience with that (and it has to be done in whichever way it would have been done in CentOS 7).


Can you access the GRUB menu during boot?

thanks. No, I can’t access GRUB during boot - but… I can get a full restoration of my server if it plays up. So, I guess I can install the regular kernel on a sunday morning, reboot server and it if messes up - restore it back to saturday night build again.
So, which rpm should I install and where would I download it from please?

See above: Migrating from centos 7.9.2009 to almalinux - STUCK - #7 by sej7278