[release] Add Azure VM launcher release test. #57921
Conversation
Signed-off-by: kevin <kevin@anyscale.com>
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
ray-gardener
bot
added
core
|
Could you paste a link of a success run? |
aslonnie
left a comment
aslonnie
left a comment
There was a problem hiding this comment.
let me know when this is ready to review.
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
|
|
||
| set -exo pipefail | ||
|
|
||
| pip3 install azure-cli-core==2.21.0 azure-core azure-identity azure-mgmt-compute azure-mgmt-network azure-mgmt-resource azure-common msrest msrestazure |
There was a problem hiding this comment.
can we use depset for this?
I thought many of the deps are already installed in the image?
| "--tenant", | ||
| tenant_id, | ||
| ] | ||
| ) |
There was a problem hiding this comment.
Bug: Azure Auth Fails Without AWS Credentials
The azure_authenticate() function is called unconditionally for all Azure provider types, but it requires AWS credentials to access AWS Secrets Manager. This will fail when users try to run the script locally without AWS credentials, making the script difficult to use outside of CI. According to the PR discussion, this makes it "difficult to run the script from as someone who doesn't have access to the AWS bucket with the secrets in there." The authentication logic should be moved to a separate script or made optional to allow local development and testing.
| cluster_compute: azure/tests/azure_compute.yaml | ||
| run: | ||
| timeout: 2400 | ||
| script: bash release/azure_docker_login.sh && python -I launch_and_verify_cluster.py azure/tests/azure-cluster.yaml --num-expected-nodes 3 --retries 10 |
There was a problem hiding this comment.
Bug: Incorrect Script Path in Azure Cluster Test
The azure_cluster_launcher test's script path release/azure_docker_login.sh is relative to the working_dir (../python/ray/autoscaler/). This path is incorrect, and the script won't be found.
- Add Azure VM launcher release test - Change region for the Azure cluster to be in `centralus` since `westus2` has trouble with availability. - Add helper function to authenticate with Azure using service principal in launch cluster script --------- Signed-off-by: kevin <kevin@anyscale.com>
- Add Azure VM launcher release test - Change region for the Azure cluster to be in `centralus` since `westus2` has trouble with availability. - Add helper function to authenticate with Azure using service principal in launch cluster script --------- Signed-off-by: kevin <kevin@anyscale.com> Signed-off-by: Aydin Abiar <aydin@anyscale.com>
centralussincewestus2has trouble with availability.