The list-resources.sh script is used solely by our CI, specifically kubernetes/test-infra/kubetest with the --check-leaked-resources flag. Currently if a single resource fails to list, we fail the entire job. I think this is too brittle. A review of previous issues on kubernetes/kubernetes that relate to failure of this script shows that the issues usually resolve themselves, or would be caught by the diff of before/after. Let's instead allow the script to continue listing all resources, and let kubetest's resource diff fail the job.
3.6 KiB
Executable File
3.6 KiB
Executable File