EMBA status (was: bug#50509: 28.0.50; [PATCH] elisp-mode-tests regression)

EMBA status (was: bug#50509: 28.0.50; [PATCH] elisp-mode-tests regression)

[Michael Albinus]

[Redirecting to emacs-devel@gnu.org]

Lars Ingebrigtsen <larsi@gnus.org> writes:

>> And yes, we're hurting without a modern CI.
>
> Indeed.  Which reminds me -- what's going on with EMBA?  It fails on all
> builds:
>
>   https://emba.gnu.org/emacs/emacs/-/pipelines
>
> "There has been a runner system failure, please try again"

I'm kind of blocked on EMBA. I've asked Ted a while ago to give me
access to the runner(s) for test projects; unfortunately I didn't get an
answer.

Perhaps I shall try to reanimate it w/o further support :-(

(I'm also restricted due to health problems, but this is another story.)

Best regards, Michael.

Re: EMBA status

[Ted Zlatanov]

On Sun, 12 Sep 2021 17:54:57 +0200 Michael Albinus <michael.albinus@gmx.de> wrote: 

MA> [Redirecting to emacs-devel@gnu.org]
MA> Lars Ingebrigtsen <larsi@gnus.org> writes:

>>> And yes, we're hurting without a modern CI.
>> 
>> Indeed.  Which reminds me -- what's going on with EMBA?  It fails on all
>> builds:
>> 
>> https://emba.gnu.org/emacs/emacs/-/pipelines
>> 
>> "There has been a runner system failure, please try again"

MA> I'm kind of blocked on EMBA. I've asked Ted a while ago to give me
MA> access to the runner(s) for test projects; unfortunately I didn't get an
MA> answer.

I replied to Michael privately, but if anyone else wants to set up a
private fork and use the shared runner, please ping me.

I'm not sure how to resolve the wider problem of contributions that
break tests. It's easy to start sending blame-mails but it's equally
easy for people to ignore them.

Ted

Re: EMBA status

[Dmitry Gutov]

On 13.09.2021 16:09, Ted Zlatanov wrote:
> I'm not sure how to resolve the wider problem of contributions that
> break tests. It's easy to start sending blame-mails but it's equally
> easy for people to ignore them.

It would be much harder to ignore them than when they are not sent.

I don't we are allowed to send them, though. Not without some official 
degree of acceptance of EMBA. As the CI, at least.

Re: EMBA status

[Lars Ingebrigtsen]

Dmitry Gutov <dgutov@yandex.ru> writes:

> On 13.09.2021 16:09, Ted Zlatanov wrote:
>> I'm not sure how to resolve the wider problem of contributions that
>> break tests. It's easy to start sending blame-mails but it's equally
>> easy for people to ignore them.

But wasn't the problem here that all builds fail?

> It would be much harder to ignore them than when they are not sent.
>
> I don't we are allowed to send them, though. Not without some official
> degree of acceptance of EMBA. As the CI, at least.

I'm all for sending automatic mail from EMBA to the author of the
failing commit, but I guess it has to be more reliable first.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Re: EMBA status

[Alan Third]

On Tue, Sep 14, 2021 at 01:17:36PM +0200, Lars Ingebrigtsen wrote:
> Dmitry Gutov <dgutov@yandex.ru> writes:
> 
> > On 13.09.2021 16:09, Ted Zlatanov wrote:
> >> I'm not sure how to resolve the wider problem of contributions that
> >> break tests. It's easy to start sending blame-mails but it's equally
> >> easy for people to ignore them.
> 
> But wasn't the problem here that all builds fail?

Indeed. It looks like the docker environment is broken in some way and
has been for months:

ERROR: Job failed (system failure): Error response from daemon: Conflict. The container name "/runner-2oEDrUT7-project-1-concurrent-1-build-4" is already in use by container "f46df44a43bee1f61a6ede2eac732a4e06eaa2094659f7ea23718f093d67bcc6". You have to remove (or rename) that container to be able to reuse that name. (executor_docker.go:726:0s)
-- 
Alan Third

Re: EMBA status, Re: EMBA status

[Ted Zlatanov]

On Tue, 14 Sep 2021 13:17:36 +0200 Lars Ingebrigtsen <larsi@gnus.org> wrote: 

LI> I'm all for sending automatic mail from EMBA to the author of the
LI> failing commit, but I guess it has to be more reliable first.

Cool. Can we start with a month of monitoring
nntp+news.gmane.io:gmane.emacs.buildstatus and bugging people to fix the
build if it breaks? That should give everyone more confidence.

On Tue, 14 Sep 2021 16:25:07 +0100 Alan Third <alan@idiocy.org> wrote: 

AT> On Tue, Sep 14, 2021 at 01:17:36PM +0200, Lars Ingebrigtsen wrote:
>> 
>> But wasn't the problem here that all builds fail?

AT> Indeed. It looks like the docker environment is broken in some way and
AT> has been for months:

AT> ERROR: Job failed (system failure): Error response from daemon: Conflict. The
AT> container name "/runner-2oEDrUT7-project-1-concurrent-1-build-4" is already in
AT> use by container
AT> "f46df44a43bee1f61a6ede2eac732a4e06eaa2094659f7ea23718f093d67bcc6". You have to
AT> remove (or rename) that container to be able to reuse that name.
AT> (executor_docker.go:726:0s)

I stopped checking after the builds kept failing for a few months. I
think the rest of the volunteers also dropped off.

I've recovered the machine and its dedicated runner now. The lockup was
due to shared memory exhaustion.

Please let us know at emacs-build-automation@gnu.org or
nntp+news.gmane.io:gmane.emacs.build-automation if you see more errors.

Ted

Re: EMBA status, Re: EMBA status

[Lars Ingebrigtsen]

Ted Zlatanov <tzz@lifelogs.com> writes:

> Cool. Can we start with a month of monitoring
> nntp+news.gmane.io:gmane.emacs.buildstatus and bugging people to fix the
> build if it breaks? That should give everyone more confidence.

Sure, but it still seems like every build is failing?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no