Ubuntu Considers Replacing initramfs-tools WIth Dracut

[Luke]

I've used dracut on Debian with a working plymouth setup for nearly a decade. At first a lot of customization was required to make it work, now very little. I use a custom Plymouth theme based on the "script" plugin and have had just one issue: I have yet to figure out how to render text in a script-based theme while in the initramfs. I thus have to use premade labels, since things like text for boot passphrase promptd won't render. I had it working at first but that didn't last as other things changed. Some kind of issue with including a font and all the stuff needed to render it.

[ahrs]

The problem with Ubuntu changing internal stuff are the hidden maneuvers to force those changes upstream on Debian...

Ubuntu is downstream of Debian, not upstream. It's always nice when the two can align though.

[Danielsan]

The problem with Ubuntu changing internal stuff are the hidden maneuvers to force those changes upstream on Debian...

[Anux]

What is their intention, are there concrete problems with initRFS that dracut solves? Why are those reasons not listed here? Is it just a change for fun?

[ahrs]

I hate dracut it is hard to configure.

Is there an initramfs that's easy to configure? I have only ever built my own custom Busybox based initramfs because that always seemed easier to me than wrangling a distros initramfs to do something.

[ahrs]

Ok so is dracut shit now too? Usually when Canonical/Ubuntu starts using something, it's already shit or becomes shit. See e.g. https://www.phoronix.com/news/Fedora-39-mkosi-initrd

I've seen my fair sure of "My made up problem only I have is handled badly by Dracut" but in the common case it's good. I don't use an initramfs on any of my systems though, I don't need it.

I'd rather see Canonical do more on unified kernel images and possibly even prebuilt initramfs images built and signed in their infrastructure like Fedora was going to look into in the past (not sure what happened to that). The idea of building and assembling an initramfs on end-user systems is broken.

[Matt Taggart]

dracut has been available in debian since 2010 https://tracker.debian.org/news/4453...-1-source-all/

[siyia]

I hate dracut it is hard to configure.

[klh_io]

Good. I'm tired of the countless flavors that could be summarized as a meta package or an alternative ISO. I'm tired of software that exist because of NIH Syndrome. If we want software to improve, we need more collaboration.

You don't seem to not understand what a distro is? People are still collaborating on the software included in them, the only difference is the default tooling and branding. It's not like GNOME on Ubuntu is that different from GNOME on Fedora, save some compat patches.

[Espionage724]

Ok so is dracut shit now too? Usually when Canonical/Ubuntu starts using something, it's already shit or becomes shit. See e.g. https://www.phoronix.com/news/Fedora-39-mkosi-initrd

Heh, that's interesting correlation; I didn't hear about Fedora's proposal but it looks like it's F42 so I'll look at it again then. I'm all-in on whatever Fedora does for the most part and Dracut seemingly still works fine on F41 beta