Batching world-rebuilding changes for the core-updates-frozen branch

Batching world-rebuilding changes for the core-updates-frozen branch

[Maxim Cournoyer]

Hello Guix!

Since there's going to be at least this change [0] causing a world
rebuild of the core-updates-frozen branch, I'd like to know if there are
other world-rebuilding but low-risk changes you'd like to see integrated
into the branch.

If you do, feel free to 'block' the 50358 issue with the changes, so
that I can review them and push them at the same time as [0].

Thank you!

Maxim

[0]  https://issues.guix.gnu.org/50358

Re: Batching world-rebuilding changes for the core-updates-frozen branch

[Guillaume Le Vaillant]

[-- Attachment #1: Type: text/plain, Size: 817 bytes --]

Maxim Cournoyer <maxim.cournoyer@gmail.com> skribis:

> Hello Guix!
>
> Since there's going to be at least this change [0] causing a world
> rebuild of the core-updates-frozen branch, I'd like to know if there are
> other world-rebuilding but low-risk changes you'd like to see integrated
> into the branch.
>
> If you do, feel free to 'block' the 50358 issue with the changes, so
> that I can review them and push them at the same time as [0].
>
> Thank you!
>
> Maxim
>
> [0]  https://issues.guix.gnu.org/50358

Hi,

I just pushed a world-rebuilding fix for a bug in binutils (see [1])
2 hours ago.
It's evaluation 15282 on ci.guix.gnu.org. Maybe someone with admin
access to the CI could cancel it, in order not to rebuild everything
twice.

[1] https://lists.gnu.org/archive/html/guix-devel/2021-09/msg00077.html

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 247 bytes --]

Re: Batching world-rebuilding changes for the core-updates-frozen branch

[Thiago Jung Bauermann]

Hello Maxim,

Em quarta-feira, 8 de setembro de 2021, às 12:00:34 -03, Maxim Cournoyer 
escreveu:
> Hello Guix!
> 
> Since there's going to be at least this change [0] causing a world
> rebuild of the core-updates-frozen branch, I'd like to know if there are
> other world-rebuilding but low-risk changes you'd like to see integrated
> into the branch.

Not a code change for the branch, but it would be nice if powerpc64le-linux 
emulation would be removed from the workers so that we can have more 
reliable CI results on that platform by making it use the OSUOSL POWER9 
build machine that has been recently added (guixp9). Some packages can be 
built with emulation, but some very early in the dependency tree can’t 
(e.g., libsigsegv) so I don’t see much benefit in trying.

Mathieu made the change with commit c0a9022d8234 (“hydra: cuirass: Disable 
powerpc64le emulation.”) on the maintenance repo (thanks!), so what is left 
is deploying it on every hydra build machine for it to take effect.

-- 
Thanks,
Thiago

Re: Batching world-rebuilding changes for the core-updates-frozen branch

[Tobias Geerinckx-Rice]

[-- Attachment #1: Type: text/plain, Size: 678 bytes --]

Thiago Jung Bauermann 写道：
> Mathieu made the change with commit c0a9022d8234 (“hydra: 
> cuirass: Disable 
> powerpc64le emulation.”) on the maintenance repo (thanks!), so 
> what is left 
> is deploying it on every hydra build machine for it to take 
> effect.

I just reconfigured all berlin build nodes, with no errors 
reported.

Note that /proc/sys/fs/binfmt_misc/qemu-ppc64le still exists on 
all of them and is still listed in berlin-nodes.scm.  I don't know 
if there are still ways for emulated builds to ‘leak’ into 
happening behind Cuirass's back, and/or whether it's now unused 
cruft that should be removed.

Kind regards,

T G-R

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 247 bytes --]

Re: Batching world-rebuilding changes for the core-updates-frozen branch

[Thiago Jung Bauermann]

Hello,

Em quinta-feira, 9 de setembro de 2021, às 12:47:56 -03, Tobias Geerinckx-
Rice escreveu:
> Thiago Jung Bauermann 写道：
> 
> > Mathieu made the change with commit c0a9022d8234 (“hydra:
> > cuirass: Disable
> > powerpc64le emulation.”) on the maintenance repo (thanks!), so
> > what is left
> > is deploying it on every hydra build machine for it to take
> > effect.
> 
> I just reconfigured all berlin build nodes, with no errors
> reported.

Thank you!

Curiously, the odd-numbered workers still advertise powerpc64-linux: 
https://ci.guix.gnu.org/machine/hydra-guix-101

> Note that /proc/sys/fs/binfmt_misc/qemu-ppc64le still exists on
> all of them and is still listed in berlin-nodes.scm.  I don't know
> if there are still ways for emulated builds to ‘leak’ into
> happening behind Cuirass's back, and/or whether it's now unused
> cruft that should be removed.

I don’t know either.

-- 
Thanks,
Thiago

Re: Batching world-rebuilding changes for the core-updates-frozen branch

[Tobias Geerinckx-Rice]

[-- Attachment #1: Type: text/plain, Size: 380 bytes --]

Thiago Jung Bauermann 写道：
> Curiously, the odd-numbered workers still advertise 
> powerpc64-linux: 
> https://ci.guix.gnu.org/machine/hydra-guix-101

…?  Curious indeed!  I definitely remember odd numbers scrolling 
by, but I don't recall .101 specifically, nor did I read the 
(voluminous) output in detail.  I'll take a look tomorrow.

Kind regards,

T G-R

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 247 bytes --]

Re: Batching world-rebuilding changes for the core-updates-frozen branch

[Thiago Jung Bauermann]

Hello,

Em quinta-feira, 9 de setembro de 2021, às 19:33:43 -03, Tobias Geerinckx-
Rice escreveu:
> Thiago Jung Bauermann 写道：
> 
> > Curiously, the odd-numbered workers still advertise
> > powerpc64-linux:
> > https://ci.guix.gnu.org/machine/hydra-guix-101
> 
> …?  Curious indeed!  I definitely remember odd numbers scrolling
> by, but I don't recall .101 specifically, nor did I read the
> (voluminous) output in detail.  I'll take a look tomorrow.

Perhaps there’s a delay in when those web pages are updated. Today for the 
first time I’ve seen a build being routed to guixp9:

--8<---------------cut here---------------start------------->8---
Sep 07 14:51:14 guixp9 cuirass[25700]: Building `/gnu/store/
4rf7myvpzjmnfqksp3hlf2r2qm36gn3k-ublock-origin-chromium-1.37.2.drv' 
derivation.
Sep 07 14:51:14 guixp9 cuirass[25700]: ;;; (("https://ci.guix.gnu.org" 
"https://bordeaux.guix.gnu.org"))
Sep 08 18:27:47 guixp9 cuirass[25700]: Building `/gnu/store/
4rf7myvpzjmnfqksp3hlf2r2qm36gn3k-ublock-origin-chromium-1.37.2.drv' 
derivation.
Sep 08 18:27:47 guixp9 cuirass[25700]: ;;; (("https://ci.guix.gnu.org" 
"https://bordeaux.guix.gnu.org"))
Sep 09 04:47:10 guixp9 cuirass[25700]: Derivation `/gnu/store/
4rf7myvpzjmnfqksp3hlf2r2qm36gn3k-ublock-origin-chromium-1.37.2.drv' build 
failed: build of `/gnu/store/4rf7myvpzjmnfqksp3hlf2r2qm36gn3k-ublock-
origin-chromium-1.37.2.drv' failed
--8<---------------cut here---------------end--------------->8---

So it seems to have worked, even though https://ci.guix.gnu.org/metrics 
still says there were 0 builds there in the past 24h.

-- 
Thanks,
Thiago

Re: Batching world-rebuilding changes for the core-updates-frozen branch

[Thiago Jung Bauermann]

Hello again,

Em quarta-feira, 8 de setembro de 2021, às 12:00:34 -03, Maxim Cournoyer 
escreveu:
> Since there's going to be at least this change [0] causing a world
> rebuild of the core-updates-frozen branch, I'd like to know if there are
> other world-rebuilding but low-risk changes you'd like to see integrated
> into the branch.
> 
> If you do, feel free to 'block' the 50358 issue with the changes, so
> that I can review them and push them at the same time as [0].

I just remembered these patches which I sent a while back to fix gawk in 
‘%static-inputs’. They fix native builds of static-binaries-tarball:

https://issues.guix.gnu.org/50174

-- 
Thanks,
Thiago

Re: Batching world-rebuilding changes for the core-updates-frozen branch

[Attila Lendvai]

[-- Attachment #1: Type: text/plain, Size: 688 bytes --]

not sure this is relevant here, but there is this fix for wrap-script, too:

http://issues.guix.gnu.org/40039

Sent from mobile, pardon my brevity.

-------- Original Message --------
On Sep 8, 2021, 17:00, Maxim Cournoyer wrote:

> Hello Guix!
>
> Since there's going to be at least this change [0] causing a world
> rebuild of the core-updates-frozen branch, I'd like to know if there are
> other world-rebuilding but low-risk changes you'd like to see integrated
> into the branch.
>
> If you do, feel free to 'block' the 50358 issue with the changes, so
> that I can review them and push them at the same time as [0].
>
> Thank you!
>
> Maxim
>
> [0] https://issues.guix.gnu.org/50358

[-- Attachment #2: Type: text/html, Size: 985 bytes --]

Re: Batching world-rebuilding changes for the core-updates-frozen branch

[Sarah Morgensen]

Hi,

Maxim Cournoyer <maxim.cournoyer@gmail.com> writes:

> Hello Guix!
>
> Since there's going to be at least this change [0] causing a world
> rebuild of the core-updates-frozen branch, I'd like to know if there are
> other world-rebuilding but low-risk changes you'd like to see integrated
> into the branch.

Perhaps we also want to wait for the fix for "cannot build missing
derivation" [0] to be pushed and deployed as well?  A world-rebuild is
likely to cause a *lot* of this, leading to useless CI feedback.

[0] https://issues.guix.gnu.org/50040
--
Sarah

Re: Batching world-rebuilding changes for the core-updates-frozen branch

[Thiago Jung Bauermann]

Hello again,

Em quinta-feira, 9 de setembro de 2021, às 19:47:20 -03, Thiago Jung 
Bauermann escreveu:
> Hello,
> 
> Em quinta-feira, 9 de setembro de 2021, às 19:33:43 -03, Tobias
> Geerinckx-
> Rice escreveu:
> > Thiago Jung Bauermann 写道：
> > 
> > > Curiously, the odd-numbered workers still advertise
> > > powerpc64-linux:
> > > https://ci.guix.gnu.org/machine/hydra-guix-101
> > 
> > …?  Curious indeed!  I definitely remember odd numbers scrolling
> > by, but I don't recall .101 specifically, nor did I read the
> > (voluminous) output in detail.  I'll take a look tomorrow.
> 
> Perhaps there’s a delay in when those web pages are updated. Today for
> the first time I’ve seen a build being routed to guixp9:
  ⋮
<snip>
  ⋮
> So it seems to have worked, even though https://ci.guix.gnu.org/metrics
> still says there were 0 builds there in the past 24h.

Actually, it’s not a case of the web pages being out of date. The
hydra-guix-<odd number> machines are still doing emulated powerpc64le-linux 
builds, unfortunately. :-/

ON the bright side, guixp9 is doing some of the builds as well!

-- 
Thanks,
Thiago