-
Notifications
You must be signed in to change notification settings - Fork 6.5k
[Test][KubeRay] Update KubeRay version to v1.4.0 for autoscaler tests #53974
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
[Test][KubeRay] Update KubeRay version to v1.4.0 for autoscaler tests #53974
Conversation
Signed-off-by: Chi-Sheng Liu <chishengliu@chishengliu.com>
Signed-off-by: Chi-Sheng Liu <chishengliu@chishengliu.com>
Signed-off-by: Rueian <rueiancsie@gmail.com>
f30d792
to
a3027c8
Compare
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.
Pull Request Overview
This PR updates the Kubernetes-based Ray testing infrastructure to use KubeRay v1.4.0 and Ray v2.46.0, and switches the autoscaler v2 template to use the new autoscalerOptions.version
field instead of an environment variable.
- Bump KubeRay and Ray versions in test scripts and YAML templates
- Remove
RAY_enable_autoscaler_v2
env var and enable autoscaler v2 viaversion: v2
- Ensure the
kuberay-system
namespace is set and created before running operator tests
Reviewed Changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.
Show a summary per file
File | Description |
---|---|
release/k8s_tests/run_gcs_ft_on_k8s.py | Updated KUBERAY_VERSION from v1.2.2 to v1.4.0 |
python/ray/tests/kuberay/test_files/ray-cluster.autoscaler-v2-template.yaml | Updated Ray image/tag to 2.46.0, enabled version: v2 , removed old env var |
python/ray/tests/kuberay/test_files/ray-cluster.autoscaler-template.yaml | Updated Ray image/tag to 2.46.0 |
python/ray/autoscaler/kuberay/init-config.sh | Bumped KUBERAY_BRANCH and OPERATOR_TAG to v1.4.0; added namespace override |
ci/k8s/run-operator-tests.sh | Added kubectl create namespace kuberay-system |
@@ -17,6 +17,7 @@ kind load docker-image ray-ci:kuberay-test | |||
# python python/ray/tests/kuberay/setup/setup_kuberay.py | |||
|
|||
bash python/ray/autoscaler/kuberay/init-config.sh | |||
kubectl create namespace kuberay-system |
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.
Make namespace creation idempotent to avoid failures when rerunning tests. For example, you could use kubectl apply -f -
with a namespace manifest or append || true
to ignore 'already exists' errors.
kubectl create namespace kuberay-system | |
cat <<EOF | kubectl apply -f - | |
apiVersion: v1 | |
kind: Namespace | |
metadata: | |
name: kuberay-system | |
EOF |
Copilot uses AI. Check for mistakes.
@@ -51,7 +51,7 @@ def generate_cluster_variable(): | |||
|
|||
def check_kuberay_installed(): | |||
# Make sure the ray namespace exists | |||
KUBERAY_VERSION = "v1.2.2" | |||
KUBERAY_VERSION = "v1.4.0" |
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.
[nitpick] Consider centralizing version strings (e.g., KubeRay and Ray versions) into a shared configuration or environment variable to reduce duplication and simplify future bumps across scripts and templates.
Copilot uses AI. Check for mistakes.
@@ -24,6 +24,7 @@ DIR=$(mktemp -d -t "kuberay-XXXXXX") | |||
( | |||
cd kuberay/ray-operator/config/default | |||
kustomize edit set image kuberay/operator=quay.io/kuberay/operator:"$OPERATOR_TAG" |
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.
[nitpick] It may be clearer to handle namespace overrides via a separate overlay or document the reason for this change in a comment, preventing unexpected modifications to base manifests.
kustomize edit set image kuberay/operator=quay.io/kuberay/operator:"$OPERATOR_TAG" | |
kustomize edit set image kuberay/operator=quay.io/kuberay/operator:"$OPERATOR_TAG" | |
# Set the namespace to 'kuberay-system' for the operator. This is required for the current deployment setup. |
Copilot uses AI. Check for mistakes.
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.
I chatted with @rueian offline. The issue is because:
- The KubeRay installs in
default
namespace instead ofray-system
namespace since v1.3.0. - We forget to update the Ray repository to v1.3.0.
…ray-project#53974) Signed-off-by: Chi-Sheng Liu <chishengliu@chishengliu.com> Signed-off-by: Rueian <rueiancsie@gmail.com> Co-authored-by: Rueian <rueiancsie@gmail.com>
Why are these changes needed?
Part of KubeRay release process.
autoscalerOptions.version = v2
instead of setting theRAY_enable_autoscaler_v2
environment variable.Related issue number
Checks
git commit -s
) in this PR.scripts/format.sh
to lint the changes in this PR.method in Tune, I've added it in
doc/source/tune/api/
under thecorresponding
.rst
file.