On Mon, Nov 12, 2012 at 04:22:12PM +0100, Zeeshan Ali (Khattak) wrote: > On Mon, Nov 12, 2012 at 11:47 AM, Christophe Fergeau > <cfergeau@xxxxxxxxxx> wrote: > > On Sun, Nov 11, 2012 at 08:14:57PM +0100, Zeeshan Ali (Khattak) wrote: > >> From: "Zeeshan Ali (Khattak)" <zeeshanak@xxxxxxxxx> > >> > >> Instead of hardcoding target disk, lets get this info from application. > > > > Are all OSes OK with being auto installed on something else than the first disk? > > (thinking of older Windows here for example). > > Don't think so but that is no reason to assume that every app wants to > install every OS on the first disk. If some OSes are not OK with being installed some place else than the first disk, we need to be aware of these limitations, and this needs to be conveyed in the API one way or the other (you also mention this below actually). > > Doesn't this also mean that > > we need a way to choose between bootloader on MBR or bootloader on > > partition for linux installations? > > I don't know, does it? I'm not familiar at all with libosinfo installation API, so just asking questions that seem relevant to the API being designed here. > > > I'd rather not make that argument mandatory, but just fallback on vda/sda > > when it's not specified, this would make the API more convenient to use. > > Although its just one more call but yeah. Also we probably should > provide API for app to query if its possible to install on something > other than first disk. Suggestions for that, welcome. Same issues with not being familiar at all with the whole API... Maybe these restrictions are much easier to handle from a higher level, and this would belong to libvirt-designer? Christophe
Attachment:
pgpIWFJz0YGju.pgp
Description: PGP signature
