Hi everybody,

I searched but can't find anything about.

Can I use UUID when I am creating the initrd? Like "-r UUID=xxxxxx -h UUID=xxxxxx"

Because even using UUID in my fstab, the script mkintrd suggest /dev/sda2 to root and /dev/sda5 and to resume/swap it shows the UUID of partition but it can't find the partition.

I don't do hibernation, so I've never tried that, but I have successfully used -r "UUID=xxxxxxx" in my mkinitrd commands before.

Instead of "/dev/sdXX", try "/dev/disk/by-uuid/your-disk-uuid-here".

I guess which it shouldn't be necessary because both mkinitr and mkinitrd_command_generator can understand when we use UUID.

I was reading both scripts and what I can imagine:

- mkinitrd_command_generator and mkinitrd understand UUID when it's used to root partition BUT mkinitrd_command_generator only use UUID to translate from UUID to /dev/sdx
- mkinitrd_command_generator doesn't understand UUID to swap partition because it ask what is your swap device.
- mkinitrd work flawlessly with UUID. It uses UUID to find the block devices.

But against about what's on change log of 06/08/2016, there is nothing in both scripts about UUID be used to resume partitions. I can't found.

You can still use /dev/disk/by-uuid/XXXXX to reference devices as it is just a symlink to /dev/sdX. If the script doesn't have support for UUIDs or labels, it just means you can't use "UUID=XXXXXX" and have to use the location /dev/disk/by-uuid/XXXXXX.

You can still use /dev/disk/by-uuid/XXXXX to reference devices as it is just a symlink to /dev/sdX. If something doesn't have proper support for UUIDs or labels, it just means you can't use "UUID=XXXXXX" and have to use the full location... /dev/disk/by-uuid/XXXXXX.

Basically, anywhere you see /dev/sdX, you can safely replace it with /dev/disk/by{id,uuid,label}/XXXXXX.