This is enabled by default since 12.3.
If you haven’t changed the config file manually rpm will replace it with the new version. If you did edit it, it should not get replaced though, zypp.conf.rpmnew will be created instead with the new defaults. It would be your own job to merge that with the existing /etc/zypp.conf if necessary.
There’s also the command “rcrpmconfigcheck” to find/list those *.rpmnew and *.rpmsave files on your system.
Testing does not guarantee perfection. Why? Kernel updates contain fixes which may result in adverse side effects not found during testing. Also, human beings sometimes make mistakes. Without a system backup, a re-install might be the only recovery in some situations.
If a problem results, multiple kernels give the user a simple way back to reboot the previous kernel. I don’t recall anyone complaining before.
I guess another advantage of multiple kernels is that the running kernel’s modules stay on the hard disk.
In earlier versions (12.2 and before) you (mostly) had to reboot immediately after installing a kernel update, because installing the update removed the currently running kernel’s modules. So without rebooting, the kernel couldn’t load any modules from the hard disk anymore.
The effect of this was, that you suddenly could not access any USB sticks/external hard disks you connected f.e.
With multiple kernels this problem doesn’t exist because the current kernel modules don’t get deleted anymore.
>
> Why the multikernel is now enabled by default? Is it really necessary?
>
> It is assumed that minor kernel updates and are tested.
>
Also it is impossible to test all variations of hardware and software
that are out there.
ok I understand the advantage but a problem, when the kernel is updated the previous entries did not appear in the grub, only the new update. So I can not boot previous kernel versions.