It really depends if you need the new tools to debug your system as you boot from the rescue image.
Yours is quite old so it might be a good idea to change it:
The rescue kernel includes all the device support. Before integrated graphics: your graphics card died and the replacement (either an old one from the junk drawer or newly purchased model) isn’t supported by the latest installed kernel. Some recent questions in this forum dealt with moving a Fedora install on an external NVME drive in a USB case to a system with nothing in the internal NVME slot. When the NVME drive was moved into the system, kernel failed to boot because it did not include NVME support.
This is tricky to automate – old kernels may not support newer devices, and newer kernels may not support older devices. What you do depends on your
Yes, if the current kernel is working well and you don’t anticipate needing legacy device support.
See: Update rescue image?. If you have a junk drawer with devices you are keeping for emergencies, you can check the Linux Hardware Database, which tracks when linux.org drops support for a device.
If the files /boot/initramfs-0-rescue....img and /boot/vmlinuz-0-rescue.... already exist then they never get updated. If one removes those 2 files then does a kernel upgrade or reinstall the system should automatically upgrade them to the newer kernel version. The linked info on updating the rescue image is useful.
Advantages to upgrading the images include having all the system upgrades to drivers, etc., in the newer images.
what I didn’t fully understand in that post is that it is recommended to usekernel-install add, which apparently adds kernel and initramfs images to /boot. At the same time it’s saying “don’t reinstall current kernel”. Does that mean kernel-install add ... is fine but I shouldn’t be using dnf reinstall ...?
Is this really the reason? I agree that older kernels may not support newest devices but if a newer kernel doesn’t support old devices (like Intel i386 or i486), it wouldn’t make sense to run Fedora or Linux anyways. Also, if you install Fedora from scratch, you get a recent kernel as your rescue kernel, right?
I feel like the rescue kernel should get updated when one upgrades to the next fedora release. Regarding the NVME example, even if the rescue kernel has plenty of kernel modules compiled, it seems much better to have the latest.
@gnwiii refers with devices to “computer components” or usb network adapter etc.
If you change to a newer rescue kernel and you use an older usb network adapter, where is not supported anymore, you can access the rescue image but you do not have network when wifi and your old adapter not works anymore. To make them work you need a network connection, right?
That is what he pointed out. So before updating the rescue image check your tools/adapters you use to manage your system in emergency, if they really also work with the kernel you want update too.
to see the exact command it will be using. You can run these commands directly without re-downloading the kernel package. Thus it may be faster this way.
Some old devices have been usable on 64-bit linux systems via USB adapters for SCSI and other legacy interfaces.
Adding an NVME drive via a PCIe card or a different (older or newer) GPU are probably the most common cases where a rescue kernel boots while the latest installed kernel fails because the device support was omitted during installation/. Early WiFI and Bluetooth adapters may have hasd linux.org drivers in past, but not in recent kernels.