Hide Forgot
Description of problem: Create cluster with customized install-config.yaml. But the zones stetting for master did not take effect. Version-Release number of the following components: $ ./openshift-install version ./openshift-install v4.0.6-1-dirty How reproducible: always Steps to Reproduce: 1. Create install-config with installer. # ./openshift-install create install-config --dir test 2. Edit install-config.yaml to change master and worker replicas and instance type as following: <--snip--> compute: - name: worker platform: aws: zones: - us-ap-northeast-2a - us-ap-northeast-2b type: m5.xlarge replicas: 3 controlPlane: name: master platform: aws: zones: - us-ap-northeast-2c type: m5.xlarge replicas: 3 <--snip--> platform: aws: region: ap-northeast-1 <--snip--> 3. Create cluster with above install-config. # ./openshift-install create cluster --dir test Actual results: APIServer launched successfully, the workers created failed. Expected results: APIServer should launch failed for the error platform zones. Additional info: $ oc get machines qe-piqin2-2ht6z-master-0 -n openshift-machine-api --config=test1/auth/kubeconfig -oyaml apiVersion: machine.openshift.io/v1beta1 kind: Machine metadata: <--snip--> spec: <--snip--> placement: availabilityZone: us-ap-northeast-2c region: ap-northeast-1 publicIp: null <--snip-->
You're asking for validation checks to ensure that the requested zones lie in the requested region?
I think we should launch master instances according to the controlPlane.platform.aws.zones, if the zones do not exist, master instances should not be launched.
Verified with: $ ./openshift-install version ./openshift-install v4.0.16-1-dirty