On Fri, 30 Jan 2015, John Darrington wrote:

> I kindof presumed it was the bandwidth to hydra which was the problem.  Not the
> machine itself.
>
> Are you sure which problem needs to be solved before we spend $$$ on it?

Most of the time I was seeing that hydra is unresponsive and guix will 
fetch the source instead a substitute, but ATM I am getting a message:

guix substitute-binary: warning: while fetching 
http://hydra.gnu.org/nar/nax41ygqk14w23nbyzfh4i23vxl3qf01-python-3.3.5: 
server is somewhat slow

but it is still fetching the substitute from there. The hydra admin should 
find what is the cause.


> On Tue, Jan 27, 2015 at 11:11:36PM +0100, Ludovic Courtès wrote:
>
>     I think we have just the right size in terms of x86 build machines.  The
>     front-end itself shouldn’t also be a build machine, because it already
>     has enough to do (evaluations + HTTP requests, which includes providing
>     compressed archives on demand.)
>
>     So I would like to keep hydra.gnunet.org as a build machine and to
>     switch to a dedicated box for the front-end.
>
>     The FSF might be able to help with that, otherwise we’ll go ahead and
>     get ourselves another machine, I think.  But I think it’s time to fix
>     that issue now.
>
>     Thanks,
>     Ludo’.