On Sun, Jan 12, 2025 at 07:17:19PM +0100, Pascal Hambourg wrote: > On 12/01/2025 at 16:29, Holger Wansing wrote: > > Pascal Hambourg <pas...@plouf.fr.eu.org> wrote (Sun, 12 Jan 2025 16:13:21 > > +0100): > > > > > > I just wanted to make sure everyone is aware that the revert fixes > > > only this specific use case (guided partitioning using LVM with or > > > without encryption). The issue already existed and remains when using > > > guided partitioniong without LVM (ext4 root, no separate /boot). > > > > Ok, so that's fine. > > There is too much inconsistency here for me to say it is fine. > > On one hand, the recipes for ppc64el create an optional ext2 /boot partition > only with LVM, which means that booting with Petitboot will fail if the user > chooses guided partitioning without LVM. >
ppc64el - I think this is the first time I've heard this mentioned in a long time? It's one of those architectures that the media releases team have no hardware for, for example (having just done a point release yesterday). Also, as noted, Petitboot isn't packaged in Debian as far as I can see and the Power 8 probably works fine with grub? > On the other hand, the recipes for arm64 uselessly create a mandatory ext4 > /boot partition even without LVM. > > In between, the recipes for amd64 create an optional ext4 /boot partition > only with LVM and the default recipes create a mandatory ext2 /boot > partition in all cases. > > Either a common boot loader for a given architecture does not support /boot > on ext4 and then the recipes for this architecture should create a mandatory > ext2 /boot partition in all cases (like default recipes), or all common boot > loaders are known to support /boot on ext4 and then the recipes should > create an optional ext4 /boot partition only with LVM (like recipes for > amd64). > Petitboot suggests that it *should* boot on all filesystems supported by Linux kexec. Although consistency is desirable across architectures, maybe this really is an edge case? All the very best, as ever, Andy
