On Thursday 15 June 2006 16:07, David Härdeman wrote: > I'm guessing you meant this thread: > http://lists.debian.org/debian-boot/2005/10/msg01082.html Yes, that's the one (though I thought it was more recent; time flies...). > > One major reason seems to be to allow resume swap after > > suspend-to-disk. > > Using initramfs-tools, I resume from a swap-on-lvm partition daily. After a default installation or after tweaking things? Problem is that we cannot only consider the i386/grub/initramfs-tools case. Other arches/bootloaders/initrd generators have to be considered too. Although of course creating swap within LVM can be activated on a per-arch basis by only modifying the relevant recipes. > The reasons I could find in the above mentioned thread were: > > suspend: not a problem > > lowmem: if you're not able to have swap on lvm due to lowmem, > are you really able to have root on lvm? > > overhead: do we have any proof at all for this claim? > harddrive transfer rates should be the bottleneck, not the > (theoretical) lvm overhead. Bastian seems to agree with you on the last one. I would suggest discussing this somewhere (d-devel maybe?) and see what arguments are brought forward.
Attachment:
pgpviwhWE2HbM.pgp
Description: PGP signature