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
Enable new SD agents by default #1252
Conversation
{% if properties['stackdriver-kubernetes'] %} | ||
# TODO: remove alpha when 10.2 is public. | ||
# https://github.com/kubeflow/kubeflow/issues/821 | ||
enableKubernetesAlpha: true | ||
# We need 1.10.2 to support Stackdrivier GKE. | ||
initialClusterVersion: 1.10.2-gke.0 |
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.
Now most recent version is 1.10.5-gke.3, and 1.10.2-gke.0 is no long available from UI.
/retest |
/lgtm |
/retest |
1 similar comment
/retest |
scripts/gke/deploy.sh
Outdated
@@ -159,6 +159,10 @@ if ${KUBEFLOW_DEPLOY}; then | |||
|
|||
# Install the GPU driver. It has no effect on non-GPU nodes. | |||
kubectl apply -f https://raw.githubusercontent.com/GoogleCloudPlatform/container-engine-accelerators/stable/nvidia-driver-installer/cos/daemonset-preloaded.yaml | |||
|
|||
# Install Stackdriver Kubernetes agents. |
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.
Why do we need to do this? Aren't the stackdriver agents installed by default?
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 beta instructions don't mention installing these agents
https://cloud.google.com/monitoring/kubernetes-engine/installing?
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 instruction there are using gcloud or pantheon. We are doing manually here.
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.
Should we take advantage of the YAML support in ksonnet and start a package for all the YAML specs for GKE?
We could do this in a follow on a PR. I think the question is how can we make it easier to keep in sync between our web app and deploy.sh
Right now we'd have to specify the YAML spec in two different locations.
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.
initialClusterVersion: 1.9.6-gke.1 | ||
{% endif %} | ||
initialClusterVersion: 1.10.5-gke.3 | ||
loggingService: none |
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.
Why don't we need to set loggingService and monitoringService anymore?
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.
Those are legacy ones.
/hold |
/lgtm |
/hold cancel |
/lgtm |
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: jlewi, kunmingg The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
* Enable new SD agents by default * fix
* [python sdk] add v1beta1 models * upgrade version of python SDK to 0.0.3 * remove v1Alpha3 python sdk * add some python models manually: v1Time and V1UnstructuredUnstructured * bring back v1alpha3 * create separate python sdk for v1alpha3 and v1beta1 * move on * release pkg on pypi.org * remove dist files * refine
* Related to kubeflow#1252
Now that we have k8s 1.10.
/cc @kunmingg
This change is