Core-updates

Core-updates

[Andreas Enge]

Hello,

one more thing that would be nice to have is gtk-doc:
   http://hydra.gnu.org:3000/build/1345571
I already tried to upgrade to 1.25, but that did not change the error.
And maybe node for those who are interested in it:
   http://hydra.gnu.org:3000/build/1326189

And there are a few failures of python packages.

Two more cases of non-determinism: The updated versions of dealii and zsh
built without problem on my x86_64 laptop, but fail on hydra (on x86_64
and i686).

Andreas

core-updates

[Ricardo Wurmus]

Hey,

I’ve requested a new evaluation of core-updates on hydra because sbcl
and a few other packages have since been fixed.  This should show us
fewer packages that we need to worry about before merging core-updates.

--
Ricardo

GPG: BCA6 89B6 3655 3801 C3C6  2150 197A 5888 235F ACAC
https://elephly.net

Re: core-updates

[Ludovic Courtès]

Hi!

Ricardo Wurmus <rekado@elephly.net> skribis:

> I’ve requested a new evaluation of core-updates on hydra because sbcl
> and a few other packages have since been fixed.  This should show us
> fewer packages that we need to worry about before merging core-updates.

As of this writing hydra.gnu.org is still building 1.8K jobs (mostly
armhf):

  https://hydra.gnu.org/jobset/gnu/core-updates

berlin.guixsd.org is roughly done, building mostly for aarch64:

  https://berlin.guixsd.org/status/

The weather is OK, though it’s surprisingly lower than what I’d expect
given that everything has been built according to Hydra/Cuirass:

--8<---------------cut here---------------start------------->8---
$ ./pre-inst-env guix weather --substitute-urls="https://berlin.guixsd.org https://mirror.hydra.gnu.org"
computing 7,047 package derivations for x86_64-linux...
looking for 7,348 store items on https://berlin.guixsd.org...
https://berlin.guixsd.org
  55.2% substitutes available (4,057 out of 7,348)
  7,756.8 MiB of nars (compressed)
  23,011.1 MiB on disk (uncompressed)
  0.000 seconds per request (2.1 seconds in total)
  3,553.4 requests per second
looking for 7,348 store items on https://mirror.hydra.gnu.org...
updating list of substitutes from 'https://mirror.hydra.gnu.org'... 100.0%
https://mirror.hydra.gnu.org
  37.3% substitutes available (2,744 out of 7,348)
  9,299.7 MiB of nars (compressed)
  23,535.4 MiB on disk (uncompressed)
  0.266 seconds per request (1,953.6 seconds in total)
  3.8 requests per second
--8<---------------cut here---------------end--------------->8---

FWIW, the closure of my system as reported by:

  guix size $(readlink -f /run/current-system) | tail -1

has shrunk from 2.7G on master to 2.2G on ‘core-updates’, which is nice
(it includes QEMU among other things, which is why it’s so big.)

Ludo’.

Re: core-updates

[Ricardo Wurmus]

Ludovic Courtès <ludo@gnu.org> writes:

> The weather is OK, though it’s surprisingly lower than what I’d expect
> given that everything has been built according to Hydra/Cuirass:

The patch for GHC 8 is not yet part of core-updates, which will
invalidate a lot of packages.  We should get that commit from master and
apply it to core-updates, and then start another evaluation.

Have there been any big failures that could explain the low coverage?

--
Ricardo

GPG: BCA6 89B6 3655 3801 C3C6  2150 197A 5888 235F ACAC
https://elephly.net

Re: core-updates

[Mark H Weaver]

ludo@gnu.org (Ludovic Courtès) writes:

> Ricardo Wurmus <rekado@elephly.net> skribis:
>
>> I’ve requested a new evaluation of core-updates on hydra because sbcl
>> and a few other packages have since been fixed.  This should show us
>> fewer packages that we need to worry about before merging core-updates.
>
> As of this writing hydra.gnu.org is still building 1.8K jobs (mostly
> armhf):
>
>   https://hydra.gnu.org/jobset/gnu/core-updates
>
> berlin.guixsd.org is roughly done, building mostly for aarch64:
>
>   https://berlin.guixsd.org/status/
>
> The weather is OK, though it’s surprisingly lower than what I’d expect
> given that everything has been built according to Hydra/Cuirass:
>
> $ ./pre-inst-env guix weather --substitute-urls="https://berlin.guixsd.org https://mirror.hydra.gnu.org"
> computing 7,047 package derivations for x86_64-linux...
> looking for 7,348 store items on https://berlin.guixsd.org...
> https://berlin.guixsd.org
>   55.2% substitutes available (4,057 out of 7,348)
>   7,756.8 MiB of nars (compressed)
>   23,011.1 MiB on disk (uncompressed)
>   0.000 seconds per request (2.1 seconds in total)
>   3,553.4 requests per second
> looking for 7,348 store items on https://mirror.hydra.gnu.org...
> updating list of substitutes from 'https://mirror.hydra.gnu.org'... 100.0%
> https://mirror.hydra.gnu.org
>   37.3% substitutes available (2,744 out of 7,348)
>   9,299.7 MiB of nars (compressed)
>   23,535.4 MiB on disk (uncompressed)
>   0.266 seconds per request (1,953.6 seconds in total)
>   3.8 requests per second

Hydra has certainly built far more than 37% of jobs on core updates.
The percentage of built store items should be closer to 90% or more.

I guess the discrepancy is because most of those store items had never
been requested, so the NARs haven't yet been built.  Does that make
sense?

      Mark

core-updates & ‘guix weather’

[Ludovic Courtès]

Mark H Weaver <mhw@netris.org> skribis:

> ludo@gnu.org (Ludovic Courtès) writes:

[...]

>> The weather is OK, though it’s surprisingly lower than what I’d expect
>> given that everything has been built according to Hydra/Cuirass:
>>
>> $ ./pre-inst-env guix weather --substitute-urls="https://berlin.guixsd.org https://mirror.hydra.gnu.org"
>> computing 7,047 package derivations for x86_64-linux...
>> looking for 7,348 store items on https://berlin.guixsd.org...
>> https://berlin.guixsd.org
>>   55.2% substitutes available (4,057 out of 7,348)
>>   7,756.8 MiB of nars (compressed)
>>   23,011.1 MiB on disk (uncompressed)
>>   0.000 seconds per request (2.1 seconds in total)
>>   3,553.4 requests per second
>> looking for 7,348 store items on https://mirror.hydra.gnu.org...
>> updating list of substitutes from 'https://mirror.hydra.gnu.org'... 100.0%
>> https://mirror.hydra.gnu.org
>>   37.3% substitutes available (2,744 out of 7,348)
>>   9,299.7 MiB of nars (compressed)
>>   23,535.4 MiB on disk (uncompressed)
>>   0.266 seconds per request (1,953.6 seconds in total)
>>   3.8 requests per second
>
> Hydra has certainly built far more than 37% of jobs on core updates.
> The percentage of built store items should be closer to 90% or more.

Yeah, definitely.

> I guess the discrepancy is because most of those store items had never
> been requested, so the NARs haven't yet been built.  Does that make
> sense?

No I don’t think so (I checked on hydra and ‘guix publish’ was not
creating new nars.)  And it’s really 404s not 500s that I’m getting.
Not sure what’s going on.

Ludo’.