Open darkmaxx1 opened 3 years ago
The problem with EFI partition detection is that in some cases device is not in the form of sdXY I can suggest following solution:
If i will prepare a test fix, do you volunteer to test it ? ;)
OL8 is up and running, but yeah, I can do a Rear recover to return back to CentOS8 and test your fix :P
Oops, the rear recover isn't working anymore. Duh :(
I noticed this issue as well. After the switch, efibootmgr still showed an entry for CentOS, which was pointing at the now-nonexistent EFI/centos/shimx64.efi
instead of EFI/redhat/shimx64.efi
. Removing it with -B
and then using the command as per @darkmaxx1 (altering the boot device and the partition number in -p
) works.
This issue should be solved, otherwise it may be quite annoying to those who are doing the switch remotely, as it may require physical access to the machine...
similar issues for me on migration
also see here: https://github.com/oracle/centos2ol/issues/70
I've been trying to migrate CentOS 8 to Oracle Linux, everything went fine when disabling the non-repo, but when rebooting, the server couldn't start on the EFI boot. It says failed without showing the grub menu.
Did a recover with ReaR after each failed time to find out why. I figured the EFI Boot wasn't on sda, but sdb.
It fixed my issue efibootmgr -c -d /dev/sdb -p 1 -L "Oracle Linux" -l "\EFI\redhat\shimx64.efi"
I'm wondering if the script can be changed to detect the location of the EFI partition ?