Re: guix pull takes forever

[ng0]

ng0 transcribed 232K bytes:
> ng0 transcribed 1.3K bytes:
> > Leo Famulari transcribed 0.9K bytes:
> > > On Wed, Mar 08, 2017 at 09:38:58AM +0000, ng0 wrote:
> > > > Hi,
> > > > 
> > > > with current master and using 'guix pull' instead of a git checkout
> > > > update, the process takes more than 9 hours.
> > > 
> > > Unfortunately, I can't reproduce this.
> > > 
> > > > I restarted it this morning, it's the same. Basically what I see is
> > > > just:
> > > > 
> > > > Starting download of /tmp/guix-file.foooo
> > > > From http://urllbla.tar.gz
> > > > [status line of 11.1MiB transfered]
> > > > substitute: updating list of substitutes from bayfront 100%
> > > > substitute: updating list of substitutes from bayfront 100%
> > > > unpacking '/gnu/store/aaaaaaaaaa-guix-latest.tar.gz'
> > > > substitute: updating list of substitutes from bayfront 100%
> > > > substitute: updating list of substitutes from bayfront 100%
> > > > substitute: updating list of substitutes from bayfront 100%
> > > 
> > > What happens if you try using the mirror <https://mirror.hydra.gnu.org>
> > > instead? Bayfront is not online, as far as I know.
> > > 
> > I have 3 mirrors in my list, and up until yesterday all were being
> > picked up (except for the onion, where I still need clearer advice on how
> > to configure this - my current advice is to replicate the hydra build
> > farm and provide an substitute server myself I can test with).
> > 
> > 'guix pull' round 1: 8 hours, round 2: ~12 hours, round 3: ~4 hours.
> > None of them succeded.
> > 
> This commit happened after I started this with machine one, but I have
> two others which face similar issues now.
> For both I can reproduce the error after Ludovic's last commit to
> master, full error output appended.
> 
> 

In my last hope to at least come to the error I have on the two other
systems, I ran some guix gc repairing, optimizing, general cleaning, and
now binutils and others are being rebuild upon 'guix pull' which I see
as some progress.

I leave this for now and will watch for new commits which will fix the
more current issue.