Centos 8 wont boot after yum update

[sbdcunha]

Dear All,

I have upgraded my Centos 7 servers with a new Centos 8.2 Install and after the reboot Centos 8 refuses to boot.
I am presented with the below 3 options at boot

1 vmlinuz-4.18.0-240.1.1.el8_3.x86_64
2 vmlinuz-4.18.0-193.28.1.el8_2.x86_64

the default being no 1

if I select no 2 the older kernel it works fines and no issues

i used grubby to make my default kernel no 2 at all is well

Appreciate if some can help and advise..


Regards
simon

[TrevorH]

Boot the working one and run yum reinstall kernel-4.18.0-240.1.1.el8_3.x86_64 and see if that helps.

[sbdcunha]

Dear TrevorH,

Thanks for your quick reply
will check it out

really apprecite

regards

simon

[sbdcunha]

Dear TrevorH,

I did as you advised but no luck

if I boot from the upgraded kernel that is ... vmlinuz-4.18.0-240.1.1.el8_3.x86_64 i get lots of messages with every line prefixed with dracut and finally i get a dracut prompt

but selecting the vmlinuz-4.18.0-193.28.1.el8_2.x86_64 kernel has not issues

appreciate your advise and help.
if you need any more information please do request

Thanks and regards

simon

[tunk]

The newest kernel is 4.18.0-240.22.1, maybe you
could run yum update and see if this kernel works?

[sbdcunha]

Dear Tunk,

Thanks for your quick reply
I did a upgrade and now my kernel is

vmlinuz-4.18.0-240.22.1.el8_3.x86_64

but still face the same issue

I attach some screen shots of error messages
hope it can be of some help

Also my Server is a Cisco UCS-B200-M3

but as I said be if i choose the installation kernel that is .. vmlinuz-4.18.0-193.el8.x86_64
there is no issue at all

appreciate your advice and help

regards

simon

[jlehtone]

Have you installed some additional (disk) driver for 193?
Something that 240 lacks. Hence the LVM fails to activate volumes.

[TrevorH]

If you use lsinitrd /boot/initramfs-4.18.0-240.22.1.el8_3.x86_64 and the same command against the working kernel's initramfs then compare the two (ignoring dates & times etc) then you should be able to see if the working one includes files/drivers that are not in the new one.

[sbdcunha]

Dear jlehtone, trevorh

Thanks for the quick response

You are right i did install Centos 8.2 x64 using the inst dd option at boot since the Cisco UCS-B200-M3 Server has a SAS megaraid raid controller which I downloaded from the https://elrepo.org/linux/dud/el8/x86_64/ site and copied to USB .
the file name is dd-megaraid_sas-07.710.50.00-1.el8_2.elrepo.iso
and then I did an upgrade

kindly advise how to have this driver inserted into the new kernel

Thanks and regards

simon

[jlehtone]

The dd-*.el8_2.elrepo.iso is a disk image that contains a tiny yum repository of driver packages.
The same RPMs are available in main ELRepo repository.

If you look with:

Code: Select all

dnf list installed kmod\*

Then you probably see that driver package to have been installed from @DD-1, or something like that.

Code: Select all

dnf list --showduplicates $name_of_that_package

Should show what ELRepo has now (if you have 'elrepo' configured&enabled).

Note that there are already el8_4 packages that work only with yet to be released RHEL 8.4-based kernel (305?).

With 240 we need the el8_3 version.


Frankly, I'm not entirely sure how the update of driver package goes. It definitely places the kernel module files in some /lib/modules tree and adds links to other kernel version's trees. Furthermore, since this driver is required before /lib/modules is accessible, it must be added to the initramfs image(s)*.
(The part I don't know is whether the new driver is auto-added to image of old kernel, where it does not work and thus would break that too.)


*When machine boots, UEFI loads bootloader. Bootloader loads two files: kernel and initramfs. The initramfs is contains filesystem; files. The kernel can load files (drivers) from the initramfs. With those drivers kernel can activate and mount the filesystems that are on HDD/SSD/iscsi.