Description of problem: Exact same problem as described in: https://bugzilla.redhat.com/show_bug.cgi?id=1147328 Would expect the updated kernel to be the default after a kernel upgrade: [root@localhost ~]# cp /boot/grub2/grub.cfg{,.orig} [root@localhost ~]# dnf update kernel -y ... [root@localhost ~]# diff /boot/grub2/grub.cfg{.orig,} | grep default < set default="0" > set default="1" Where 0th entry is the latest kernel and the 1 entry is the older kernel. Version-Release number of selected component (if applicable): [root@localhost ~]# rpm -q grubby grubby-8.40-1.fc22.x86_64 [root@localhost ~]# rpm -q kernel kernel-4.0.4-301.fc22.x86_64 kernel-4.0.7-300.fc22.x86_64 How reproducible: Always Steps to Reproduce: 1. -> see description
Please attach /var/log/grubby
Created attachment 1051469 [details] /var/log/grubby
Created attachment 1113138 [details] Compressed /var/log/grubby like requested from reporter by the bug maintainer
The same thing is annoying me since years, could be since Fedora 21 like noted in 1147328. I attached by /var/log/grubby. Hope it helps.
I figured out my problem here. Somehow /etc/sysconfig/kernel did not exist on my machine. The contents of this file should have been: ``` # UPDATEDEFAULT specifies if new-kernel-pkg should make # new kernels the default UPDATEDEFAULT=yes # DEFAULTKERNEL specifies the default kernel package type DEFAULTKERNEL=kernel-core ``` I believe once I created the file the default gets updated appropriately. Albert, Can you check to see if /etc/sysconfig/kernel exists for you? If not put the above contents in there and retry.
Indeed it had to do with the contents of this file. In mine i found: DEFAULTKERNEL=kernel-PAE No idea, where this came from. I have a 64 bit x86_64 AMD processor. PAE should not make any sense here. I modified to kernel-core and now with new kernel installed it became the default as i'd expect. Thank you !
The reason why mine was messed up is because I'm not using anaconda to install my system. It would be nice if the contents of that file were written by an rpm rather than anaconda itself, but I won't complain too much about that because I know I'm doing something way outside of the norm. Closing this now as the problem seems to be found.