Introduce HYPERKUBE_IMAGE_REPO variable which is set to CoreOS
hyperkube by default. Also remove "_coreos.0" from script as it can be a
different build number. This number should be included in the kubernetes
version parameters and not in scripts.
With this, it is possible to use any combination of hyperkube image with
any tags. by default we use the CoreOS one.
Partially-Implements: blueprint support-insecure-registry
Partially-Implements: blueprint coreos-best-pratice
Change-Id: Ie0fbed4b160fa972cfe130c252e87765690e2f5f
Instead of sourcing heat-params in script, we can use it as a systemd
unit EnvironmentFile directive and not inline in sh scripts.
Change-Id: I3ebf23dee6785febdc87bc5ce4212c30ef24806e
Make Kubernetes' kube-controller-manager and kube-scheduler
health checks configurable as a parameter to the cluster-template
(label).
Set their value higher for all deployments. And set their value
to a high number for tests, for the CI.
Change-Id: I65e2da12487c513419125f0525a4e21bac22210e
Closes-Bug: 1648826
Podmaster is deprecated since k8s 1.2 and its docker
image is v1, incompatible with docker >=1.12.
* Remove podmaster pod
* Update manifests of kube-controller-manager and kube-scheduler
Change-Id: Ib46182b811aba6a6e544de265bc89a7bcb7488e1
Closes-Bug: #1646109
For k8s bay in HA mode, current all master components are active.
Added support for master elected component to ensure only one
master component is active. This will enhance the concurrency
controller.
Change-Id: I8a5489df0878a384a38bf755851c98b170282a1f
Closes-Bug: #1504681
Since we have seperate template files to different driver directory, we
can cleanup -coreos suffix from template.
1. Remove unused shell scripts.
2. Remove kube-examples.yaml and kube-user.yaml which is never used.
Closes-bug: #1606655
Change-Id: I6ac93ad23e7ae30ad1eb9be1c79c2cf36af8db0c
Michal Nasiadka