| 1 |
On Tuesday, 22 October 2019 10:09:53 BST Neil Bothwick wrote: |
| 2 |
> On 22 October 2019 08:58:01 BST, Mick <michaelkintzios@×××××.com> wrote: |
| 3 |
|
| 4 |
> >With manual copying/naming of kernels I can overwrite any non-booting |
| 5 |
> >kernels |
| 6 |
> >with the latest compiled example, without moving links around. What is |
| 7 |
> >the |
| 8 |
> >recommended solution to the above problem? |
| 9 |
> |
| 10 |
> Be more careful when configuring your kernels ;} |
| 11 |
|
| 12 |
Yes, that would be advisable. :-) |
| 13 |
|
| 14 |
TBH I have not had a non-booting kernel for some time now, but you never know |
| 15 |
what might bite you at the next turn. On new machines it could take a few |
| 16 |
kernel rebuilds before I end up with an optimal configuration, but once the |
| 17 |
kernel recipe for a particular system settles I do not as a rule experience |
| 18 |
any major problems. |
| 19 |
|
| 20 |
|
| 21 |
> I haven't used the symlink approach for some time. I use a script to |
| 22 |
> generate the entries for GRUB or systemd-boot. As I use a script to build |
| 23 |
> and install the kernel in the first place, there are no extra steps as one |
| 24 |
> script calls the other. |
| 25 |
|
| 26 |
I think I'll carry on with my manual kernel copying and naming approach. On |
| 27 |
modern UEFI machines I run efibootmgr to add/delete kernels and do not use a |
| 28 |
separate boot manager, or initrd images. So, it's not as if I have to |
| 29 |
automate what is already a rather simpler approach to upgrading kernels. |
| 30 |
-- |
| 31 |
Regards, |
| 32 |
|
| 33 |
Mick |
Archives