Setup CI against blueprint deployments #42
Comments
|
Issue-Label Bot is automatically applying the labels:
Please mark this comment with 👍 or 👎 to give our bot feedback! |
|
I have start working on kubeflow/testing#676 - This should setup ACM on the kf-ci-v1 cluster so we can use it to synchronize the tekton resources we will use. It will also include some fixes to the tekton resources. |
|
Issue-Label Bot is automatically applying the labels:
Please mark this comment with 👍 or 👎 to give our bot feedback! |
jlewi
pushed a commit
to jlewi/testing
that referenced
this issue
Jun 9, 2020
…ployments.
* Update get_kf_testing_cluster to work with blueprints
* With GCP blueprints we won't have deployments so we need to
list clusters by name and find cluster with matching regex.
* Create an ACM repo for kf-ci-v1 cluster and hydrate it with the appropriate
manifests for the auto-deploy and tektoncd namespaces.
* kubeflow#677 is tracking using ACM with the cluster kf-ci-v1
* Note this PR is only using ACM to sync tekton resources; we still need
to sync the rest of the auto-deployment infrastructure like the reconciler
and webserver.
* Remove ACM cluster selector; ACM complains because it isn't actually being applied.
* Override notebook-test-task.yaml with nb-test-task.yaml. The latter should include the latest changes from Gabriel's PR.
* When deploying the blueprint we need to do `kpt cfg set email` to
set the email for the default profile. Without this change
the deployment won't include the namespace we need to run
the tests in
GoogleCloudPlatform/kubeflow-distribution#42 is tracking CI for the blueprints.
jlewi
pushed a commit
to jlewi/testing
that referenced
this issue
Jun 9, 2020
…ployments.
* Update get_kf_testing_cluster to work with blueprints
* With GCP blueprints we won't have deployments so we need to
list clusters by name and find cluster with matching regex.
* Create an ACM repo for kf-ci-v1 cluster and hydrate it with the appropriate
manifests for the auto-deploy and tektoncd namespaces.
* kubeflow#677 is tracking using ACM with the cluster kf-ci-v1
* Note this PR is only using ACM to sync tekton resources; we still need
to sync the rest of the auto-deployment infrastructure like the reconciler
and webserver.
* Remove ACM cluster selector; ACM complains because it isn't actually being applied.
* Override notebook-test-task.yaml with nb-test-task.yaml. The latter should include the latest changes from Gabriel's PR.
* When deploying the blueprint we need to do `kpt cfg set email` to
set the email for the default profile. Without this change
the deployment won't include the namespace we need to run
the tests in
GoogleCloudPlatform/kubeflow-distribution#42 is tracking CI for the blueprints.
k8s-ci-robot
pushed a commit
to kubeflow/testing
that referenced
this issue
Jun 9, 2020
…ployments. (#678) * Update get_kf_testing_cluster to work with blueprints * With GCP blueprints we won't have deployments so we need to list clusters by name and find cluster with matching regex. * Create an ACM repo for kf-ci-v1 cluster and hydrate it with the appropriate manifests for the auto-deploy and tektoncd namespaces. * #677 is tracking using ACM with the cluster kf-ci-v1 * Note this PR is only using ACM to sync tekton resources; we still need to sync the rest of the auto-deployment infrastructure like the reconciler and webserver. * Remove ACM cluster selector; ACM complains because it isn't actually being applied. * Override notebook-test-task.yaml with nb-test-task.yaml. The latter should include the latest changes from Gabriel's PR. * When deploying the blueprint we need to do `kpt cfg set email` to set the email for the default profile. Without this change the deployment won't include the namespace we need to run the tests in GoogleCloudPlatform/kubeflow-distribution#42 is tracking CI for the blueprints.
jlewi
pushed a commit
to jlewi/testing
that referenced
this issue
Jun 10, 2020
* Use ACM to deploy the auto-deploy infrastructure
* Previously we were using ACM to manage Tekton resources but not
the auto-deploy server and reconciler
* Move the Makefile with the hydration rules to the root of the repository
Fix double deployments of blueprints kubeflow#666
* We were reading pipelineruns twice because of symbolic links
* Related to GoogleCloudPlatform/kubeflow-distribution#42 CI for blueprint deployments
jlewi
pushed a commit
to jlewi/testing
that referenced
this issue
Jun 10, 2020
* Use ACM to deploy the auto-deploy infrastructure
* Previously we were using ACM to manage Tekton resources but not
the auto-deploy server and reconciler
* Move the Makefile with the hydration rules to the root of the repository
Fix double deployments of blueprints kubeflow#666
* We were reading pipelineruns twice because of symbolic links
* Related to GoogleCloudPlatform/kubeflow-distribution#42 CI for blueprint deployments
* Update OWNERs file.
* Add Reming and Yuan; remove folks no longer actively working on Kubeflow
|
With kubeflow/testing#679 and kubeflow/examples#803 I'm able to run the tekton pipeline to run mnist. Running mnist is currently failing
|
k8s-ci-robot
pushed a commit
to kubeflow/testing
that referenced
this issue
Jun 10, 2020
) * Use ACM to deploy the auto-deploy infrastructure * Previously we were using ACM to manage Tekton resources but not the auto-deploy server and reconciler * Move the Makefile with the hydration rules to the root of the repository Fix double deployments of blueprints #666 * We were reading pipelineruns twice because of symbolic links * Related to GoogleCloudPlatform/kubeflow-distribution#42 CI for blueprint deployments * Update OWNERs file. * Add Reming and Yuan; remove folks no longer actively working on Kubeflow
|
I think we need to solve: |
This was referenced Jun 11, 2020
k8s-ci-robot
pushed a commit
to kubernetes/test-infra
that referenced
this issue
Jun 12, 2020
jlewi
pushed a commit
to jlewi/gcp-blueprints
that referenced
this issue
Jun 15, 2020
* Write a golang test to verify we can hydrate the manifests * Trigger go tests under prow * Related to GoogleCloudPlatform#42
This was referenced Jun 15, 2020
k8s-ci-robot
pushed a commit
that referenced
this issue
Jun 17, 2020
* Add CI to verify hydration works. * Write a golang test to verify we can hydrate the manifests * Trigger go tests under prow * Related to #42 * Latest. * Latest. * Fix labels. * * Fix the tests by moving go.mod to the root of the repo. It looks like that is what the go unittests task currently requires.
jlewi
pushed a commit
to jlewi/test-infra
that referenced
this issue
Jun 18, 2020
droslean
pushed a commit
to droslean/test-infra
that referenced
this issue
Jun 22, 2020
jlewi
pushed a commit
to jlewi/gcp-blueprints
that referenced
this issue
Jun 24, 2020
k8s-ci-robot
pushed a commit
that referenced
this issue
Jun 24, 2020
|
CI is largely setup. Follow on work is being covered by the more specific issues that have been linked to this issue. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We should run the example tests continuously against the blueprint deployments
The text was updated successfully, but these errors were encountered: