jozzsi

Since this is just a warning message and since it usually spits out the following information message during initramfs generation in a container, we might want to consider removing these...

> Also, I'd suggest we stop adding {/var,usr}/lib/portables/*.raw to the initramfs automatically: I agree. I updated this PR and I created a separate PR https://github.com/dracut-ng/dracut-ng/pull/1758 for systemd-sysext

closing due to lack of enough feedback over the last 3 months.

Not much of an answer, but to stimulate conversation: Related: https://github.com/dracut-ng/dracut-ng/pull/561, https://github.com/dracut-ng/dracut-ng/pull/461 NO_XATTR introduced: 93be2bc6a CC @cgzones, @HuaxinLu Dracut can create initramfs with squashfs/erofs (packaged inside cpio initramfs) - see...

> So while it'd be an incompatible change, I think it would make sense to consider upstream switching to requiring people doing that to actually set `DRACUT_ENABLE_XATTR` (or have a...

> I am concerned that relying on the grub configuration to set a Kernel Parameter will be less robust than having the module be part of dracut's initramfs itself. @JustTNE...

I still hope to address the 'we are required to mount the top btrfs volume and then remount it as read-write explicitly to avoid all submodules being RO' part here...

> Without this, I don't think `rd.live.overlay.overlayfs=1` will work for us. We mount quite a number of subvolumes, and they have to be mounted rw. I suppose it could be...

1./ > It wouldn't be typical to explicitly mount all the subvolumes `rw` in `/etc/fstab`. I don't think many(any?) distros are doing that. This is very useful to know, but...

> I suppose it could be added to the snapshot-overlay.sh changes you suggested? Yes, this is the most obvious choice for now, but I am also trying to fully understand...