-
Notifications
You must be signed in to change notification settings - Fork 2.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Rework GCI e2e jobs #173
Rework GCI e2e jobs #173
Conversation
# TODO(wonderfly): For GCI, we currently only run CI, slow and serial | ||
# tests. More test coverage under way. | ||
- 'master': | ||
- 'master': # kubernetes-e2e-gce-gci-ci-master |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you for adding these comments!
- gcs-uploader | ||
# This file defines e2e jobs against Kuberentes HEAD and GCI HEAD, from master | ||
# and the latest three release branches, and the GCI milestones that they pin | ||
# to, e.g., GCI milestone for Kubernetes `release-1.2`. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is this an incomplete comment?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Not really. What else were you expecting? 😄
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The example looked incomplete to me. I was expecting the example will be for the map:
e.g., GCI milestone for Kubernetes release-1.2
is 52.
But I am probably reading it incorrectly.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ah, you were right. I meant to say milestone 52.
I am not too familiar with the changed file, but the plan in PR description LGTM :) |
job-env: | | ||
export JENKINS_GCI_IMAGE_TYPE="dev" | ||
# This should become 54 once it's available. | ||
export JENKINS_GCI_MILESTONE="53" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@adityakali Even without familiarity with this file, I think you can help confirm that this is to your expectation: we will use milestone 53 for Kubernetes master branch (for the interim), right?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes. Though if you want to avoid code change, you could name this image family as "gci-head" or "gci-canary" too.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It'll be hard to name it "gci-head" or "gci-canary" as one image can only be in one family. Every image is going to be in some "gci-" family already.
On the other hand, I think it makes sense to explicitly pin to a milestone even for k8s master
branch because ultimately it's their (as a GCI user) choice to make on what milestone they like.
63dd200
to
b2d089f
Compare
@fejta @spxtr Just so you know in order to resolve a rebase conflict, I decided to move the continuous Docker validation test job out to its own config file. PTAL. cc/ @dchen1107 Is there a mailing list you want to use to receive test results? I put your corp email down currently. |
b2d089f
to
eed14c0
Compare
|
||
# Template defaults. Can be overriden in job definitions. | ||
jenkins_node: 'e2e' | ||
test-owner: 'dawnchen' |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
dchen1107
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Done
This is based on our new plan to run CI tests against every active k8s branch to ensure that GCI will always be able to cut a release whenever k8s cuts a patch release. Theoretically we should cover master, 1.3, 1.2 and 1.1, but since only nodes could run on GCI back in 1.1, and 1.1 jobs used the old e2e runner, which is hard to update now, I am skipping the 1.1 branch. Also, move the continuous Docker validation jobs out to its own config file.
eed14c0
to
e7cb8c2
Compare
Addressed comments. PTAL. @dchen1107 Are you fine with the changes to the Docker continuous validation test? |
Heads up. I am going to merge this if you have no more comments. We really need to get our e2e jobs back in shape. |
SGTM |
Enable skip plugin
…pstream Rebase
ClusterLoader - Vendor update
Bonuses: * move documentation about `release.sh` to `README.md` * add reference to the documentation in `e2e-tests.sh` and `release.sh` * turn `run_dep_collector` into a more generic function * add a few smoke tests for `release.sh`
This is based on our new plan to run CI tests against every active k8s branch to
ensure that GCI will always be able to cut a release whenever k8s cuts a patch
release. Theoretically we should cover master, 1.3, 1.2 and 1.1, but since only
nodes could run on GCI back in 1.1, and 1.1 jobs used the old e2e runner, which
is hard to update now, I am skipping the 1.1 branch.
@adityakali @fejta Can you review? This depends on kubernetes/kubernetes#27083
cc/ @kubernetes/goog-image @spxtr