[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: swriteboot vs debian installer wrt Tru64 disklabel

On Wed, May 04, 2005 at 06:56:40AM -0400, Toni L. Harbaugh-Blackford [Contr] wrote:
> On Wed, 4 May 2005, Steve Langasek wrote:

>   > d-i does use swriteboot to install aboot to the disk.  That's not the issue.
>   > The issue is the same that it's always been: Tru64 doesn't cope with
>   > disklabels that don't include a full-disk slice, and the debian-installer
>   > can't reasonably create such a slice for us (for various and sundry reasons
>   > that have been discussed on this list before).

> Sorry to beat a dead horse, but I am trying to understand
> the installer's operation in the context of a PRE-existing
> Tru64 label, where the full-disk slice already exists
> and thus does not have to be created by the installer.

d-i won't be able to recognize a pre-existing Tru64 label as a valid
partition table, because it can't cope with overlapping partitions.  If you
install to such a disk, d-i will initialize a new disklabel.

> Are you saying that the installer modifies the pre-existing
> label so that no partitions cover block 0, which would in turn
> allow swriteboot to be called without '-f'?

Rather, it modifies the label so that no partitions overlap one another.
swriteboot still has to be called with '-f', because we have to create a
dummy aboot partition at the front of the disk in order to reserve the
space.  Creating a full-disk slice, aside from burning through the number of
available partitions more quickly, means swriteboot won't work at all
because it only lets you ignore overlaps with a *single* partition.

I'd love to not have to create a dummy partition for aboot so that we can
create a full-disk slice instead for Tru64 compatibility, but this isn't
going to happen for sarge.

Steve Langasek
postmodern programmer

Attachment: signature.asc
Description: Digital signature

Reply to: