fix intermittent calico failures during k8s-bootstrap

[tariq1890]

This PR hopefully fixes some of the flakes we have observed when deploying calico on a newly created K8s cluster.

The root cause of the failure is that the kubectl --kubeconfig $KUBECONFIG create -f https://raw.githubusercontent.com/projectcalico/calico/${CALICO_VERSION}/manifests/custom-resources.yaml command fails as the cluster doesn't immediately recognise one of the custom resource kinds (defined by the CRD created in the previous step) as there seems to be some delay in CRD creation going through after issuing the kubectl create command of the same. This failure prompts a retry which ultimately fails as it also retries creating custom resources which were already created successfully in the first attempt.

This PR applies a two-fold fix
i) Run kubectl apply instead of kubectl create as it is idempotent and declarative
ii) Sleep 10s after the Calico CRDs are created so that the new kinds are fully registered in the cluster etcd