Yep. We are back. We still have a slow queue, but at least the builds are
not failing randomly. Please rebase your builds on top of the current
master and push again to trigger builds.
On Tue, Jul 23, 2019 at 9:31 PM Jarek Potiuk
wrote:
> Looks like they fixed it: https://github.com/travis-ci/wo
Thank you Aizhamal! Glad to see this as a heavy YouTube user :D
Cheers,
Kevin Y
Aizhamal Nurmamat kyzy 于2019年7月23日 周二下午2:05写道：
> Hello everyone,
>
> I thought it would be nice to have a central repo for talk recordings and
> videos related to Apache Airflow, so I have created a YouTube Channel f
Hello everyone,
I thought it would be nice to have a central repo for talk recordings and
videos related to Apache Airflow, so I have created a YouTube Channel for
it [1].
I uploaded the recordings from Airflow meetup that took place on June 5 in
the Bay Area. If any PMC or Committer wishes to ha
>
> Question - what is the purpose of introducing kaniko instead of using
> regular docker build?
>
Indeed. We want to be as agnostic as possible. What I plan to do is to use
Kubernetes Runner in GitlabCI. This means that all the jobs will run as
Kubernetes PODs in GKE - Gitlab CI will only be UI
I think that Jarek is proposing switching to Kaniko for security
considerations as GKE workloads run in OS containers (so they do not
benefit from hardware virtualization sandboxing) and docker build requires
root privileges.
In any case, Kaniko is not tied in any way to Google Cloud, so the build
Hi that's a nice writeup, easy to follow. Also I like your diagram.
Question - what is the purpose of introducing kaniko instead of using
regular docker build?
I'm asking in line with the consideration "The system should be
self-maintainable - with as little special Development/Ops maintenance
ne
Looks like they fixed it: https://github.com/travis-ci/worker/issues/604
On Tue, Jul 23, 2019 at 8:38 PM Driesprong, Fokko
wrote:
> I see issues at different Apache projects as well, Druid and Avro. They're
> running out of memory. Let's see how Travis responds.
>
> Cheers, Fokko
>
> Op di 23 ju
I have no strong "No" against any proposed change of these cases. So I go with
+1 (non-binding).
P.S. Thanks Jarek for bringing this up again and your intense work towards
airflow currently :) and thanks to Kamil for even creating this document. I
like how the code is getting more and more cons
I agree with the other folks. I'm personally not a fan of pre commit hooks,
but if people like it, they can enable it.
Cheers, Fokko
Op di 23 jul. 2019 om 20:13 schreef Kaxil Naik
> It is fully-optional so I don't think we need an AIP there.
>
>
>
> On Tue, Jul 23, 2019 at 11:36 PM Beau Barker
I see issues at different Apache projects as well, Druid and Avro. They're
running out of memory. Let's see how Travis responds.
Cheers, Fokko
Op di 23 jul. 2019 om 19:43 schreef Jarek Potiuk :
> FYI. Still not fixed. Others experience this as well:
> https://github.com/travis-ci/worker/issues/6
It is fully-optional so I don't think we need an AIP there.
On Tue, Jul 23, 2019 at 11:36 PM Beau Barker
wrote:
> Just add a .pre-commit-config.yaml to the project, no need for an AIP.
>
> > On 24 Jul 2019, at 2:42 am, Jarek Potiuk
> wrote:
> >
> > Any more comments on it?
> > Should I make a
Just add a .pre-commit-config.yaml to the project, no need for an AIP.
> On 24 Jul 2019, at 2:42 am, Jarek Potiuk wrote:
>
> Any more comments on it?
> Should I make an AIP for that :)? Or should I just ask a vote/propose a PR
> ? Anyone has a strong opinion?
> I think it changes the dev workfl
FYI. Still not fixed. Others experience this as well:
https://github.com/travis-ci/worker/issues/604
On Tue, Jul 23, 2019 at 11:34 AM Jarek Potiuk
wrote:
> No good news yet. We are getting randomly assigned 1CPU /3.5GB mem
> instances still. Infrastructure is on it.
>
> On Tue, Jul 23, 2019 at 1
Any more comments on it?
Should I make an AIP for that :)? Or should I just ask a vote/propose a PR
? Anyone has a strong opinion?
I think it changes the dev workflow quite a bit on one hand, but it is
fully optional on the other hand.
On Thu, Jul 4, 2019 at 9:23 AM Kamil Breguła
wrote:
> +1 At
Hello everyone,
This email is calling a vote on the changes in import paths. It's been
discussed in
https://lists.apache.org/thread.html/4e648d9421c792d4537f5ac66f1a16dce468f816fc5221a9f9db9433@%3Cdev.airflow.apache.org%3E
The vote will last for at least 1 week (July 30th 6pm CEST), and at least
Hello Everyone,
I prepared a short docs where I described general architecture of the
solution I imagine we can deploy fairly quickly - having GitLab CI support
and Google provided funding for GCP resources.
I am going to start working on Proof-Of-Concept soon but before I start
doing it, I would
No good news yet. We are getting randomly assigned 1CPU /3.5GB mem
instances still. Infrastructure is on it.
On Tue, Jul 23, 2019 at 10:49 AM Jarek Potiuk
wrote:
> It looks like we are back to the original specs. I am runnning tests and
> re-enable everything if I see it works.
>
> J.
>
> On Tue
It looks like we are back to the original specs. I am runnning tests and
re-enable everything if I see it works.
J.
On Tue, Jul 23, 2019 at 10:34 AM Jarek Potiuk
wrote:
> From INFRA: "I have confirmed that our builds appear to be running with
> 3.75GB memory and 1 core currently. This does not
>From INFRA: "I have confirmed that our builds appear to be running with
3.75GB memory and 1 core currently. This does not match Travis' standard
specs (7.5GB and 2 cores), and I have raised a ticket with their support. I
will respond when we hear back from Travis."
On Tue, Jul 23, 2019 at 10:26
It's definitely confirmed that the problem is on Travis CI side:
I re-run the commit before the new CI was introduced (I cherry-picked a
small doc fix related to recent sphinx dependency update) and it fails in
exactly the same way (memory and cpu problems):
https://travis-ci.org/apache/airflow/bu
It's now pretty consistent and happens pretty much every time using the old
build system - for example here:
https://travis-ci.org/apache/airflow/builds/562435992.
I will cancel all PRs and disable automated PR build on Travis until we
solve the problem - as it is pointless - new PRs will simply q
21 matches
Mail list logo
