AI on GKE is a collection of examples, best-practices, and prebuilt solutions to help build, deploy, and scale AI Platforms on Google Kubernetes Engine
Apache License 2.0
211
stars
154
forks
source link
[TPU Provisioner] Delete node pool when JobSet is completed, failed, or deleted #626
Update node reconciliation based deletion controller to delete node pool when JobSet is completed, failed, or deleted
Refactor deletion controller integration tests to be more extendable so we can add new test cases more easily
Move controller integration tests to a test/integration/controller directory. When we add the Job webhook, we can add integration tests in test/integration/webhooks. E2E tests can go in test/e2e. This will help keep things organized.
This PR includes the following changes:
test/integration/controller
directory. When we add the Job webhook, we can add integration tests intest/integration/webhooks
. E2E tests can go intest/e2e
. This will help keep things organized.