How can I upgrade the kernel without losing my precious uptime?
>>56419814
https://en.wikipedia.org/wiki/KGraft
>>56419814
https://en.wikipedia.org/wiki/Kpatch
>>56419814
https://en.wikipedia.org/wiki/KernelCare
>>56419814
https://en.wikipedia.org/wiki/Ksplice
>precious uptime
there is nothing precious about it
>>56419847
>>56419888
>>56419907
Costs money. No thanks.
>>56419862
Too complicated. I don't understand? I have to write the patches myself?
>>56419918
It's precious to me. I can't show off on /g/ when my uptime is less than 24 hours.
Is there nothing that is both libre and gratis? Can't I just download a new kernel and replace the current one?
>>56419996
kexec?
manually writing to /dev/{k,}mem?
>Linuxfags have to reboot constantly whenever the smallest kernel patch is fed them by the Canonical botnet
>Linux
>Good
Haven't rebooted my OS X in ages
>>56420011
kexec shuts down your complete OS but rather than restarting your hardware, it just starts a new kernel. You'll loose any unsaved data and your uptime.
>>56419814
Do you need to upgrade your kernel?
>>56419996
nothing sed can't fix
>>56420061
It's bad for security to use old kernel?
>>56420064
I'm a honest person. I don't like lying.
>>56419814
I thought they usually provided loadable kernel modules for urgent security issues so that servers could limp along with the module until a good enough time comes along to actually upgrade your kernel.
>>56420038
But Osx is constantly begging you for reboots, and most of them don't even update any core functionality
>>56420207
Falling for a troll.
>>56420185
>It's bad for security to use old kernel?
Only if you are going to be exposed to things that can exploit the newer vulnerabilities that aren't fixed in your kernel, which usually means downloading and running things or web stuff etc.
If your device is a server or something that doesn't see any new software in months then the chances that you'll get exploited due to not upgrading is pretty remote.
I personally find it hard to believe that no one has thought of pausing all running programs in the scheduler and then live-booting into a different kernel with a info list of pids etc attached.
>>56420352
There are some difficulties. And it's hard to do and not essential.