On Tue, Aug 23, 2016 at 08:31:28AM +0100, Neil Williams wrote: > On Fri, 19 Aug 2016 09:02:30 +0000 > Riku Voipio <riku.voi...@iki.fi> wrote: > > On Tue, Aug 16, 2016 at 11:56:12AM -0400, Sam Hartman wrote: > > > We'd probably have to give up some of the tweaks we have, and add > > > support either for plugins for some of the more basic tweaks > > > directly into vmdebootstrap. As an example, vmdebootstrap would > > > almost certainly need to support raw images without a partition > > > table. > > I don't see what benefit that provides.
Paravirtualization when running Xen, EC2, etc. eg. $ ls -l /dev/xvd* brw-rw---T 1 root disk 202, 1 Jul 10 23:48 /dev/xvda1 brw-rw---T 1 root disk 202, 2 Jul 10 23:48 /dev/xvda2 $ I'm not sure if it would be possible to use pygrub if your assigned volume is partitioned, which may be a problem if you don't control the Dom0. > This is why I'm unsure about the whole plugin request - if the build > tool needs special knowledge to handle your special snowflake device, > it is *your device* which is broken. It's not your device if you don't own it. I understand where you're coming from, but do you really want to forego support for various IaaS environments (and probably a significant number of other devices which might have legitimate reasons for behaving differently which we cannot anticipate)? I do agree that snowflake changes should be packaged wherever possible, but I don't agree that we shouldn't support something just because constraints of it dictate uncommon requirements. -Adam
signature.asc
Description: Digital signature
