Docker images no longer being build and is gitlab still maintained?

494 views
Skip to first unread message

Maarten Derickx

unread,
Nov 27, 2021, 12:07:29 PM11/27/21
to sage-devel
I noticed on https://hub.docker.com/r/sagemath/sagemath/ that there haven't been any updates of the official sagemath docker image since 9.3.beta8-py3 about 9 months ago.

On that page it mentions "Every push to our GitLab repository triggers a pipeline in GitLab CI which pushes the actual images to Docker Hub." 

Now I had a look at the pipelines on gitlab https://gitlab.com/sagemath/sage/-/pipelines?page=1&scope=all&status=success the last succesful run of the pipelines was about 8 months ago. So gitlab pipelines being broken seems to explain why no new images are being pushed to dockerhub.

So I am wondering, is gitlab being phased out/abandoned, or do we still support it and do we just need to show it some love?

Thanks,
Maarten

Dima Pasechnik

unread,
Nov 27, 2021, 1:13:14 PM11/27/21
to sage-devel, Matthias Koeppe
It needs quite a bit of love, in form of a gitlab runner that doesn't run out of disk space (a dedicated runner at France runs of disk space).

It's perhaps the time to switch over to GitHub.
Mattias Koeppe should have an idea how much effort this would entail.

Dima

Thanks,
Maarten

--
You received this message because you are subscribed to the Google Groups "sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/492af376-40ad-45b2-940c-f0ea4f49bf60n%40googlegroups.com.

Maarten Derickx

unread,
Nov 28, 2021, 5:12:13 AM11/28/21
to sage-devel
What are the specs that the dedicated builder would need? In terms of cpu + ram + diskspace?

Dima Pasechnik

unread,
Nov 28, 2021, 5:32:25 AM11/28/21
to sage-devel
On Sun, Nov 28, 2021 at 10:12 AM Maarten Derickx
<m.derick...@gmail.com> wrote:
>
> What are the specs that the dedicated builder would need? In terms of cpu + ram + diskspace?

nothing spectacularly big, but somehow it manages to run out of
smallish disk space on an INRIA (or was it Paris Sud, aka Paris
Saclay?).

Like, in might need more than 20 Gb or 50Gb free space for some reason.
Nothing too bug, a reasonable modern desktop would do the job ---if we
don't let it run on every push into the repo.

Dima
> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/cf606e1d-ab3e-4cb1-85bc-26b65ecdcbd4n%40googlegroups.com.

William Stein

unread,
Nov 28, 2021, 10:38:01 AM11/28/21
to sage-devel


On Sun, Nov 28, 2021 at 2:32 AM Dima Pasechnik <dim...@gmail.com> wrote:
>
> On Sun, Nov 28, 2021 at 10:12 AM Maarten Derickx
> <m.derick...@gmail.com> wrote:
> >
> > What are the specs that the dedicated builder would need? In terms of cpu + ram + diskspace?
>
> nothing spectacularly big, but somehow it manages to run out of
> smallish disk space on an INRIA (or was it Paris Sud, aka Paris
> Saclay?).
>
> Like, in might need more than 20 Gb or 50Gb free space for some reason.
> Nothing too bug, a reasonable modern desktop would do the job ---if we
> don't let it run on every push into the repo.

There is still a powerful server at Univ of Washington that I bought using a grant specifically for Sage development 
(nearly a decade ago!) that could be used for building Docker images. It's very powerful and has tons of disk space.  
The drawback is it could just go "poof" at any moment, and it could be a very long time until it comes back, if ever.  
But for building docker images, that's probably fine.

wstein@kucalc:~/cocalc-docker/linux$ df -h .
Filesystem        Size  Used Avail Use% Mounted on
tank/home/wstein  2.7T  252G  2.5T  10% /home/wstein
wstein@kucalc:~/cocalc-docker/linux$ uptime
 07:30:53 up 904 days, 20:11,  9 users,  load average: 0.67, 0.61, 0.52
wstein@kucalc:~/cocalc-docker/linux$ free -g
              total        used        free      shared  buff/cache   available
Mem:            251          11          41           2         198         198
Swap:           255           1         254
wstein@kucalc:~/cocalc-docker/linux$ cat /proc/cpuinfo | grep processor |wc -l
48

-- William

>
> Dima
>
>
>
> >
> > On Saturday, 27 November 2021 at 19:13:14 UTC+1 dim...@gmail.com wrote:
> >>
> >>
> >>
> >> On Sat, 27 Nov 2021, 17:07 Maarten Derickx, <m.derick...@gmail.com> wrote:
> >>>
> >>> I noticed on https://hub.docker.com/r/sagemath/sagemath/ that there haven't been any updates of the official sagemath docker image since 9.3.beta8-py3 about 9 months ago.
> >>>
> >>> On that page it mentions "Every push to our GitLab repository triggers a pipeline in GitLab CI which pushes the actual images to Docker Hub."
> >>>
> >>> Now I had a look at the pipelines on gitlab https://gitlab.com/sagemath/sage/-/pipelines?page=1&scope=all&status=success the last succesful run of the pipelines was about 8 months ago. So gitlab pipelines being broken seems to explain why no new images are being pushed to dockerhub.
> >>>
> >>> So I am wondering, is gitlab being phased out/abandoned, or do we still support it and do we just need to show it some love?
> >>
> >>
> >> It needs quite a bit of love, in form of a gitlab runner that doesn't run out of disk space (a dedicated runner at France runs of disk space).
> >>
> >> It's perhaps the time to switch over to GitHub.
> >> Mattias Koeppe should have an idea how much effort this would entail.
> >>
> >> Dima
> >>>
> >>>
> >>> Thanks,
> >>> Maarten
> >>>
> >>> --
> >>> You received this message because you are subscribed to the Google Groups "sage-devel" group.
> >>> To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
> >>> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/492af376-40ad-45b2-940c-f0ea4f49bf60n%40googlegroups.com.
> >
> > --
> > You received this message because you are subscribed to the Google Groups "sage-devel" group.
> > To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
> > To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/cf606e1d-ab3e-4cb1-85bc-26b65ecdcbd4n%40googlegroups.com.
>
> --
> You received this message because you are subscribed to the Google Groups "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.

Dima Pasechnik

unread,
Nov 28, 2021, 2:32:53 PM11/28/21
to sage-devel


On Sun, 28 Nov 2021, 15:38 William Stein, <wst...@gmail.com> wrote:


On Sun, Nov 28, 2021 at 2:32 AM Dima Pasechnik <dim...@gmail.com> wrote:
>
> On Sun, Nov 28, 2021 at 10:12 AM Maarten Derickx
> <m.derick...@gmail.com> wrote:
> >
> > What are the specs that the dedicated builder would need? In terms of cpu + ram + diskspace?
>
> nothing spectacularly big, but somehow it manages to run out of
> smallish disk space on an INRIA (or was it Paris Sud, aka Paris
> Saclay?).
>
> Like, in might need more than 20 Gb or 50Gb free space for some reason.
> Nothing too bug, a reasonable modern desktop would do the job ---if we
> don't let it run on every push into the repo.

There is still a powerful server at Univ of Washington that I bought using a grant specifically for Sage development 
(nearly a decade ago!) that could be used for building Docker images. It's very powerful and has tons of disk space.  
The drawback is it could just go "poof" at any moment, and it could be a very long time until it comes back, if ever.  
But for building docker images, that's probably fine.

can it be used as a gitlab or/and github runner for sagemath org?

I guess I don't have an account there, but it would be good to get one.

Dima

julian...@fsfe.org

unread,
Nov 28, 2021, 11:58:11 PM11/28/21
to sage-devel
Hi Maarten,

On Saturday, November 27, 2021 at 11:07:29 AM UTC-6 m.derick...@gmail.com wrote:
> So I am wondering, is gitlab being phased out/abandoned, or do we still support it and do we just need to show it some love?

I set up much of the things on GitLab initially, in the hope that it could eventually become an alternative to trac. I showed this to quite some people at Sage Days and elsewhere but did not get the feeling that there was much interest overall. So, yes, I think currently it is essentially abandoned.

As Dima said would need a runner with a fair amount of disk space (not sure how much exactly but nothing too crazy) to fix the CI. However, I would not be surprised if the CI/docker build is broken anyway since it is not part of the tests that are performed before we release.

I would not want to put too much work in fixing the CI on GitLab since I don't think we should be using GitLab just to run a bit of CI there. There seems to be some more activity of SageMath developers on GitHub currently, so migrating the docker build over to GitHub might be an option. In particular if we think that we want to support GitHub as a trac alternative at some point.

julian

Dima Pasechnik

unread,
Nov 29, 2021, 6:11:31 AM11/29/21
to sage-devel
A dedicatedself-hosted GitHub runner would be great, still. See
https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners

>
> julian
>
> --
> You received this message because you are subscribed to the Google Groups "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/b9dcce67-49a7-4970-ab93-d64a2ab7dabcn%40googlegroups.com.

Thierry

unread,
Nov 30, 2021, 7:32:04 AM11/30/21
to sage-...@googlegroups.com
Hi,

On Sun, Nov 28, 2021 at 07:32:18PM +0000, Dima Pasechnik wrote:
> On Sun, 28 Nov 2021, 15:38 William Stein, <wst...@gmail.com> wrote:
> > On Sun, Nov 28, 2021 at 2:32 AM Dima Pasechnik <dim...@gmail.com> wrote:
> > > On Sun, Nov 28, 2021 at 10:12 AM Maarten Derickx
> > > <m.derick...@gmail.com> wrote:
> > > >
> > > > What are the specs that the dedicated builder would need? In terms of
> > cpu + ram + diskspace?
> > >
> > > nothing spectacularly big, but somehow it manages to run out of
> > > smallish disk space on an INRIA (or was it Paris Sud, aka Paris
> > > Saclay?).
> > >
> > > Like, in might need more than 20 Gb or 50Gb free space for some reason.
> > > Nothing too bug, a reasonable modern desktop would do the job ---if we
> > > don't let it run on every push into the repo.
> >
> > There is still a powerful server at Univ of Washington that I bought using
> > a grant specifically for Sage development
> > (nearly a decade ago!) that could be used for building Docker images. It's
> > very powerful and has tons of disk space.
> > The drawback is it could just go "poof" at any moment, and it could be a
> > very long time until it comes back, if ever.
> > But for building docker images, that's probably fine.
>
> can it be used as a gitlab or/and github runner for sagemath org?
>
> I guess I don't have an account there, but it would be good to get one.

If this ressource is not fully used, i am also volonteering to get an
account and setup some specific (e.g. 32bit) patchbots there.

Ciao,
Thierry
> > https://groups.google.com/d/msgid/sage-devel/cf606e1d-ab3e-4cb1-85bc-26b65ecdcbd4n%40googlegroups.com
> > .
> > >
> > > --
> > > You received this message because you are subscribed to the Google
> > Groups "sage-devel" group.
> > > To unsubscribe from this group and stop receiving emails from it, send
> > an email to sage-devel+...@googlegroups.com.
> > > To view this discussion on the web visit
> > https://groups.google.com/d/msgid/sage-devel/CAAWYfq3Q2ZvF05DgW2n_gZf8Q9pcJRmZ5DcYAHkgR4MS7Tfy3w%40mail.gmail.com
> > .
> >
> >
> >
> > --
> > William (http://wstein.org)
> >
> > --
> > You received this message because you are subscribed to the Google Groups
> > "sage-devel" group.
> > To unsubscribe from this group and stop receiving emails from it, send an
> > email to sage-devel+...@googlegroups.com.
> > To view this discussion on the web visit
> > https://groups.google.com/d/msgid/sage-devel/CACLE5GDfHczNcWNiSrUnO4H%3DUTLYWrp9fqOQW1hVUvSFJF6ZaQ%40mail.gmail.com
> > <https://groups.google.com/d/msgid/sage-devel/CACLE5GDfHczNcWNiSrUnO4H%3DUTLYWrp9fqOQW1hVUvSFJF6ZaQ%40mail.gmail.com?utm_medium=email&utm_source=footer>
> > .
> >
>
> --
> You received this message because you are subscribed to the Google Groups "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/CAAWYfq1g8kibm1zc%2BuOYvDEwqrKecSrhxSg-PmoKwSSpmurbog%40mail.gmail.com.

William Stein

unread,
Nov 30, 2021, 12:57:51 PM11/30/21
to sage-devel
Thanks Thierry, I'll email you directly to figure out the details.


To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/20211130123159.GA16928%40metelu.net.


--
William (http://wstein.org)

Maarten Derickx

unread,
Dec 2, 2021, 11:06:30 PM12/2/21
to sage-devel
In the meantime I managed to verify that aside from the gitlab CI/CD there are no other things that are broken. Meaning that I managed to build the docker file shipped with sage just fine on my laptop. I pushed a build of 9.4 to https://hub.docker.com/r/mderickx/sagemath/ in case anyone is interested.

Op dinsdag 30 november 2021 om 18:57:51 UTC+1 schreef wst...@gmail.com:

Eric Gourgoulhon

unread,
Dec 3, 2021, 7:36:21 AM12/3/21
to sage-devel
Hi,

Le vendredi 3 décembre 2021 à 05:06:30 UTC+1, maarten...@navara.nl a écrit :
In the meantime I managed to verify that aside from the gitlab CI/CD there are no other things that are broken. Meaning that I managed to build the docker file shipped with sage just fine on my laptop. I pushed a build of 9.4 to https://hub.docker.com/r/mderickx/sagemath/ in case anyone is interested.

Thanks for the 9.4 image!
I've tried it to run a notebook with Binder, by inserting
FROM mderickx/sagemath:latest
in the Dockerfile of the repository
https://github.com/egourgoulhon/SageMathTour
But it fails, as the previous Sage images (cf. the discussion at https://groups.google.com/g/sage-devel/c/GCOg8oT3beI/m/I_hWqTYaAQAJ), with the following error:

2021-12-03T10:30:51Z [Normal] Successfully pulled image "gcr.io/binderhub-288415/r2d-staging-g5b5b759-egourgoulhon-2dsagemathtour-2fda44:32fab2e884998806d7465f08b7dc21a8a99793d8" in 49.782009677s
2021-12-03T10:30:52Z [Normal] Created container notebook
2021-12-03T10:30:52Z [Normal] Started container notebook
Spawn failed: Timeout
Launch attempt 2 failed, retrying...
Server requested
2021-12-03T10:35:09.007721Z [Normal] Successfully assigned prod/jupyter-egourgoulhon-2dsagemathtour-2dyl7nva8s to gke-prod-user-202009-b9c03ca0-5ga5
2021-12-03T10:35:10Z [Normal] Container image "jupyterhub/mybinder.org-tc-init:2020.12.4-n655.hfe65496" already present on machine
2021-12-03T10:35:10Z [Normal] Created container tc-init
2021-12-03T10:35:10Z [Normal] Started container tc-init
2021-12-03T10:35:11Z [Normal] Pulling image "gcr.io/binderhub-288415/r2d-staging-g5b5b759-egourgoulhon-2dsagemathtour-2fda44:32fab2e884998806d7465f08b7dc21a8a99793d8"
2021-12-03T10:36:58Z [Normal] Successfully pulled image "gcr.io/binderhub-288415/r2d-staging-g5b5b759-egourgoulhon-2dsagemathtour-2fda44:32fab2e884998806d7465f08b7dc21a8a99793d8" in 1m47.176651527s
2021-12-03T10:37:00Z [Normal] Created container notebook
2021-12-03T10:37:00Z [Normal] Started container notebook
Failed to connect to event stream

More precisely, this was obtained by clicking on the button "Execute on Binder" of
https://nbviewer.org/github/egourgoulhon/SageMathTour/blob/master/Notebooks/demo_pseudo_Riemannian_Schwarzschild.ipynb

Again, this used to work with the Sage 9.3.beta8 image and previous ones at some time. So it could be that a change has been performed on the Binder side...

Eric

Dima Pasechnik

unread,
Dec 3, 2021, 8:01:43 AM12/3/21
to sage-devel
On Fri, 3 Dec 2021, 12:36 Eric Gourgoulhon, <egourg...@gmail.com> wrote:
Hi,

Le vendredi 3 décembre 2021 à 05:06:30 UTC+1, maarten...@navara.nl a écrit :
In the meantime I managed to verify that aside from the gitlab CI/CD there are no other things that are broken. Meaning that I managed to build the docker file shipped with sage just fine on my laptop. I pushed a build of 9.4 to https://hub.docker.com/r/mderickx/sagemath/ in case anyone is interested.

Thanks for the 9.4 image!
I've tried it to run a notebook with Binder, by inserting

that's certainly  a Binder problem that is doesn't start (well, it's close to 3Gb image...)

I can get this image and run it in Docker (on macOS) in terminal, as

 docker run -it  mderickx/sagemath

This gives Sage 9.4 prompt immediately. Terminal things work, things like
sage: sphere()                                                                                                                                                                    
Launched html viewer for Graphics3d Object

do not have any effects, the graphics is nowhere to be seen (not so surprising)

It would be more interesting to know how to use it in local Jupyter (notebook, or VSCcode)

Dima
 
FROM mderickx/sagemath:latest
in the Dockerfile of the repository
https://github.com/egourgoulhon/SageMathTour
But it fails, as the previous Sage images (cf. the discussion at https://groups.google.com/g/sage-devel/c/GCOg8oT3beI/m/I_hWqTYaAQAJ), with the following error:

2021-12-03T10:30:51Z [Normal] Successfully pulled image "gcr.io/binderhub-288415/r2d-staging-g5b5b759-egourgoulhon-2dsagemathtour-2fda44:32fab2e884998806d7465f08b7dc21a8a99793d8" in 49.782009677s
2021-12-03T10:30:52Z [Normal] Created container notebook
2021-12-03T10:30:52Z [Normal] Started container notebook
Spawn failed: Timeout
Launch attempt 2 failed, retrying...
Server requested
2021-12-03T10:35:09.007721Z [Normal] Successfully assigned prod/jupyter-egourgoulhon-2dsagemathtour-2dyl7nva8s to gke-prod-user-202009-b9c03ca0-5ga5
2021-12-03T10:35:10Z [Normal] Container image "jupyterhub/mybinder.org-tc-init:2020.12.4-n655.hfe65496" already present on machine
2021-12-03T10:35:10Z [Normal] Created container tc-init
2021-12-03T10:35:10Z [Normal] Started container tc-init
2021-12-03T10:35:11Z [Normal] Pulling image "gcr.io/binderhub-288415/r2d-staging-g5b5b759-egourgoulhon-2dsagemathtour-2fda44:32fab2e884998806d7465f08b7dc21a8a99793d8"
2021-12-03T10:36:58Z [Normal] Successfully pulled image "gcr.io/binderhub-288415/r2d-staging-g5b5b759-egourgoulhon-2dsagemathtour-2fda44:32fab2e884998806d7465f08b7dc21a8a99793d8" in 1m47.176651527s
2021-12-03T10:37:00Z [Normal] Created container notebook
2021-12-03T10:37:00Z [Normal] Started container notebook
Failed to connect to event stream

More precisely, this was obtained by clicking on the button "Execute on Binder" of
https://nbviewer.org/github/egourgoulhon/SageMathTour/blob/master/Notebooks/demo_pseudo_Riemannian_Schwarzschild.ipynb

Again, this used to work with the Sage 9.3.beta8 image and previous ones at some time. So it could be that a change has been performed on the Binder side...

Eric

--
You received this message because you are subscribed to the Google Groups "sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.

enriqu...@gmail.com

unread,
Dec 3, 2021, 10:06:41 AM12/3/21
to sage-devel
Besides the binder problem something may happen from sagemath side, since for my binder pages the cocalc image works. With other sagemath dockers I got basically the same errors as above.
Enrique

julian...@fsfe.org

unread,
Dec 4, 2021, 11:11:12 AM12/4/21
to sage-devel
Hi Maarten,

thanks for verifying that the Dockerfile still works.

On Thursday, December 2, 2021 at 10:06:30 PM UTC-6 maarten...@navara.nl wrote:
In the meantime I managed to verify that aside from the gitlab CI/CD there are no other things that are broken. Meaning that I managed to build the docker file shipped with sage just fine on my laptop. I pushed a build of 9.4 to https://hub.docker.com/r/mderickx/sagemath/ in case anyone is interested.

Did you also run the tests on the docker image? There are three tests defined here https://gitlab.com/sagemath/sage/-/blob/develop/.gitlab-ci.yml#L145. These often failed in the past with new releases.

julian

Dima Pasechnik

unread,
Dec 4, 2021, 12:08:19 PM12/4/21
to sage-devel
One way or another, it's probably a good idea to run the builder on GH Actions - translating one yml to another should be doable.


--
You received this message because you are subscribed to the Google Groups "sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/6d614b77-f673-4dbf-9e38-661fc7af4247n%40googlegroups.com.

Maarten Derickx

unread,
Dec 5, 2021, 9:06:13 AM12/5/21
to sage-devel
Hi Julian,

I didn't build a sagemath-dev image. So the first test will likely fail. I did manually test that the command line starts and that I could do computations which depend on singular pari and gap, and I also tested manually that jupyter starts and that I could do computations in it. The main purpose for building this image is that I use the sagemath docker images as base image for CI testing in a downstream package. And I needed certain features/fixes that are in 9.4 but not 9.2. And up to now all the CI tests of the downstream package pass just fine with this image.

I will try to see if I can also successfully build and test a dev image.

Dima Pasechnik

unread,
Dec 5, 2021, 9:14:32 AM12/5/21
to sage-devel
Anyone tried using buildx?

 It seems to simplify setting up the build - although on Linux it appears to have extra hoops to jump through.

To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/fc867437-4b30-4efa-8d17-4f1c842bc984n%40googlegroups.com.

Maarten Derickx

unread,
Dec 5, 2021, 10:15:40 AM12/5/21
to sage-devel
Yeah I had to use buildx to debug why the build was failing (standard docker build would not give enough logs). For me doing:

docker buildx create --use --name larger_log --driver-opt env.BUILDKIT_STEP_LOG_MAX_SIZE=50000000 docker buildx build <same commands I was giving to the standard docker build command>
Just worked without any other config. I guess you could even leave out the "--driver-opt env.BUILDKIT_STEP_LOG_MAX_SIZE=50000000" part.

Note that for me the build was failing cause of the OOM killer killed one of the build processes, so the only thing I needed to do to make the build work is assign more resources to the docker container to make it work.

julian...@fsfe.org

unread,
Dec 9, 2021, 12:37:10 AM12/9/21
to sage-devel
I rebuilt the sagemath & sagemath-dev docker images for the tags 9.3, 9.4, and latest (alias for 9.4). I am now pushing them to the Docker Hub so they should be available shortly.

The build scripts seem to work still, however, the images cannot be built on tiger lake due to https://trac.sagemath.org/ticket/32997 and each build takes about 8h on my machine due to https://trac.sagemath.org/ticket/32998.

Dima Pasechnik

unread,
Dec 9, 2021, 2:50:07 AM12/9/21
to sage-devel
I suppose you can speed up the build by using more system packages.

A "normal" build from scratch  using system packages needs less than 2 hours of wall clock time, on a 5 y.o. machine with Gentoo.

To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/8ec620b5-78be-43c7-a58f-218877bc9426n%40googlegroups.com.

Sébastien Labbé

unread,
Dec 10, 2021, 4:21:07 AM12/10/21
to sage-devel
Thank you for managing to create the docket images for 9.3 and 9.4. I am using them to test the package slabbe and it made few issues appear that were working fine with <=9.2 (for instance the --memlimit disappeared, so it made me realize to fix that).

Would you also be able to create an image for sagemath-9.2 ? That would be awesome! The most recent version in the 9.2 series available on docker is 9.2.beta3.

The previous versions are fine: the version 9.1 is not there as well, but 9.1.rc5 is, which should be quite close if not equal to 9.1. The version 9.0 is there.

Sincerely,

Sébastien

julian...@fsfe.org

unread,
Dec 10, 2021, 2:24:28 PM12/10/21
to sage-devel
Hi Sébastien,

glad to hear that the 9.3 and 9.4 images work.

On Friday, December 10, 2021 at 3:21:07 AM UTC-6 Sébastien Labbé wrote:
Would you also be able to create an image for sagemath-9.2 ? That would be awesome! The most recent version in the 9.2 series available on docker is 9.2.beta3.
 
There was a syntax error in the docker scripts in 9.2. I am trying to build manually with that syntax error fixed.

For 9.1 there is a sagemath image but none for sagemath-dev. If it builds without issues, I can push that one too.

julian

julian...@fsfe.org

unread,
Dec 10, 2021, 6:37:09 PM12/10/21
to sage-devel
Hi Sébastien,

unfortunately neither 9.1 nor 9.2 build on my machine. I had to update the base image to hirsute to fix a tiger lake issue but this breaks the ecm compilation. I probably won't put much more effort into this now but feel free to open a ticket on trac and Cc me if you want to discuss this further.

Matthias Koeppe

unread,
Dec 30, 2021, 1:19:48 PM12/30/21
to sage-devel
I have opened https://trac.sagemath.org/ticket/33099 for this.

Sébastien Labbé

unread,
Feb 7, 2022, 10:57:45 AM2/7/22
to sage-devel


On Saturday, December 11, 2021 at 12:37:09 AM UTC+1 julian...@fsfe.org wrote:
Hi Sébastien,

unfortunately neither 9.1 nor 9.2 build on my machine. I had to update the base image to hirsute to fix a tiger lake issue but this breaks the ecm compilation. I probably won't put much more effort into this now but feel free to open a ticket on trac and Cc me if you want to discuss this further.

Ok, let's forget about 9.2!

Would it be possible for you to do as you did for 9.4 but for 9.5 so that the image for 9.5 be available here?


Dima Pasechnik

unread,
Feb 17, 2022, 5:44:59 AM2/17/22
to sage-devel, Julian Rüth
We have facilities to build docker images via tox nowadays, it's as
easy as issuing a single tox command,
see e.g. https://trac.sagemath.org/ticket/33306
(needless to say, you need Docker installed)

Then you can upload the resulting image - it's more the question of
having credentials to do so, and what
a good "standard" image should be.

Dima
> --
> You received this message because you are subscribed to the Google Groups "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/96c5d96b-17a5-47fd-b739-9b1f5cdcb4b5n%40googlegroups.com.

Matthias Koeppe

unread,
Feb 17, 2022, 12:00:47 PM2/17/22
to sage-devel
Also the Docker images pre-built using tox are available - see https://wiki.sagemath.org/ReleaseTours/sage-9.6#Pre-built_Docker_containers_on_ghcr.io

But note that in contrast to the images on Docker hub, these images are not optimized for size. 

Dima Pasechnik

unread,
Feb 17, 2022, 12:24:26 PM2/17/22
to sage-devel
On Thu, Feb 17, 2022 at 5:00 PM Matthias Koeppe
<matthia...@gmail.com> wrote:
>
> Also the Docker images pre-built using tox are available - see https://wiki.sagemath.org/ReleaseTours/sage-9.6#Pre-built_Docker_containers_on_ghcr.io

TIL that GitHub "packages"
https://github.com/orgs/sagemath/packages?repo_name=sage
are actually Docker containers, hosted on aptly named ghcr.io
> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/71ac2bca-41af-4567-b421-d3d0287ef849n%40googlegroups.com.

Matthias Koeppe

unread,
Feb 17, 2022, 1:58:03 PM2/17/22
to sage-devel
On Thursday, February 17, 2022 at 9:24:26 AM UTC-8 Dima Pasechnik wrote:
On Thu, Feb 17, 2022 at 5:00 PM Matthias Koeppe
<matthia...@gmail.com> wrote:
>
> Also the Docker images pre-built using tox are available - see https://wiki.sagemath.org/ReleaseTours/sage-9.6#Pre-built_Docker_containers_on_ghcr.io

TIL that GitHub "packages"
https://github.com/orgs/sagemath/packages?repo_name=sage
are actually Docker containers, hosted on aptly named ghcr.io



 

julian...@fsfe.org

unread,
Feb 23, 2022, 2:08:41 PM2/23/22
to sage-devel
Hi Sébastien,

I pushed sagemath and sagemath-dev images for 9.5. Unfortunately, they are quite a bit bigger than the images for 9.4, see https://trac.sagemath.org/ticket/33409.

I hope the images are functional nevertheless. Please let me know if there is any problem with them.

julian

Sébastien Labbé

unread,
Feb 25, 2022, 8:33:07 AM2/25/22
to sage-devel
On Wednesday, February 23, 2022 at 8:08:41 PM UTC+1 julian...@fsfe.org wrote:
Hi Sébastien,

I pushed sagemath and sagemath-dev images for 9.5.

Great! Thank you!
 
Unfortunately, they are quite a bit bigger than the images for 9.4, see https://trac.sagemath.org/ticket/33409.

I hope the images are functional nevertheless. Please let me know if there is any problem with them.

I added 9.5 to the list of image on which to test my optional package. Tests are currently done:

We will see if there is a problem with the image .... or with my package, in which case I will be happy to know!

Sébastien

Sébastien Labbé

unread,
Feb 25, 2022, 10:36:58 AM2/25/22
to sage-devel
I added 9.5 to the list of image on which to test my optional package. Tests are currently done:

We will see if there is a problem with the image .... or with my package, in which case I will be happy to know!


Actually, there is a problem with the image. This is the time taken by the other sage docker image to run the git-lab continuous integration test for package slabbe:

test:9.0               00:16:58
test:9.1.rc5           00:17:42
test:9.1.rc5-py3  00:17:03
test:9.2.beta3    00:17:26
test:9.3               00:18:55
test:9.4          00:19:44

which is under the default timeout of 1h for gitlab-ci tests.

Unfortunately, the image 9.5 takes more than 1 hour: "The script exceeded the maximum execution time set for the job"

and finishes with:

[ecl-21.2.1] ;;; in file PPRINT.LSP, position 19849
[ecl-21.2.1] ;;; at (DEFUN ASSURE-SPACE-IN-BUFFER ...)
[ecl-21.2.1] ;;; Inlining read access to slot BUFFER-FILL-POINTER from class PRETTY-STREAM
[ecl-21.2.1] ;;; Note:
[
Job's log exceeded limit of 4194304 bytes.
Job execution will continue but no more output will be collected.

See:


 
Reply all
Reply to author
Forward
0 new messages