Migrate kube-aws to node.kubernetes.io/role: control-plane taint

[mumoshu]

See https://groups.google.com/forum/#!topic/kubernetes-sig-cluster-lifecycle/6HSVqGlyCrI

We currently use the old taint in the number of places:

• https://github.com/kubernetes-incubator/kube-aws/blob/1ffe330bdcbc43c3f2df4635b3eabfde271f3034/builtin/files/plugins/kiam/manifests/agent-daemonset.yaml#L42
• https://github.com/kubernetes-incubator/kube-aws/blob/bdeadab166e728dbcc130ac3124affc66cc4d682/pkg/api/bash_prompt.go#L33
• https://github.com/kubernetes-incubator/kube-aws/blob/1ffe330bdcbc43c3f2df4635b3eabfde271f3034/builtin/files/plugins/kiam/manifests/server-daemonset.yaml#L34
• https://github.com/kubernetes-incubator/kube-aws/blob/f6e39330006d111ce29222addde6dcec0986c308/builtin/files/plugins/dashboard/manifests/autoscaler.yaml#L50
• https://github.com/kubernetes-incubator/kube-aws/blob/062740bc02cbecc0dfaa58a7bc5936ea2d375918/builtin/files/plugins/dashboard/manifests/deployment.yaml#L86
• https://github.com/kubernetes-incubator/kube-aws/blob/2ed6d7dfc5a8fe3c10cc464fea2a0d2ac1b5bf45/builtin/files/userdata/cloud-config-controller#L387
• https://github.com/kubernetes-incubator/kube-aws/blob/062740bc02cbecc0dfaa58a7bc5936ea2d375918/builtin/files/stack-templates/control-plane.json.tmpl#L66
• https://github.com/kubernetes-incubator/kube-aws/blob/37775a7a9c5b9f06271fd42ea137b8dab98f386d/builtin/files/plugins/upgrade-helper/assets/upgrade-helper-pre.sh#L51
• https://github.com/kubernetes-incubator/kube-aws/blob/1ffe330bdcbc43c3f2df4635b3eabfde271f3034/builtin/files/plugins/aws-iam-authenticator/manifests/daemonset.yaml#L30
• https://github.com/kubernetes-incubator/kube-aws/blob/f6e39330006d111ce29222addde6dcec0986c308/builtin/files/plugins/dashboard/manifests/metrics-scraper.yaml#L38
• https://github.com/kubernetes-incubator/kube-aws/blob/1ffe330bdcbc43c3f2df4635b3eabfde271f3034/builtin/files/plugins/cluster-autoscaler/manifests/deployment.yaml#L65

[dominicgunn]

This could be a breaking change for users who might have sets that depend on being able to query by role=master. We should plan carefully around implementation here to avoid potential disruption to end-users.

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale