bug#33676: GuixSD on eoma68-a20?

[Danny Milosavljevic <dannym@scratchpost.org>]

[-- Attachment #1: Type: text/plain, Size: 1283 bytes --]

Hi Mark,

On Sat, 08 Dec 2018 15:59:58 -0500
Mark H Weaver <mhw@netris.org> wrote:

> This sounds like two distinct bugs:
> 
> * Regarding the lack of disk space: if I'm not mistaken, as things are
>   currently implemented, we must specify the size of the disk image
>   manually.  I guess we need to increase the size.

Ah, you are right!  https://hydra.gnu.org/build/3194622/log/raw (the usb-image
for i686) also fails - but nobody noticed because we only distribute the iso
image (which determines its size dynamically).

I didn't make the connection that this is a fixed-size guest filesystem before.

> * This error that occurs within QEMU is apparently not being propagated
>   to the outer build script.  That should be fixed.

Yes.

> > I'm now setting up my own build server and trying to get the flash image
> > that way - which is ridiculous.  
> 
> You seem to be assuming that the problem building the disk image is a
> Hydra-specific problem.  Do you have reason to believe so?  I would
> guess that these bugs are in the disk image builder.

Probably!

According to the progress bar it's missing an additional 40% ?  That's... large.

I've increased the size for the time being, but according to the logs
it was at 800 MiB in 2015.

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]