Merge pull request #104 from mpluhar/master

Adding documentation for rke set up and other documentation updates

cloud-controller-manager/instances.go

@@ -203,7 +203,7 @@ func (s *InstanceClient) findInstanceByProviderID(providerID string) (*ecs.Insta
 		return nil, cloudprovider.InstanceNotFound
 	}
 	if len(ins) > 1 {
-		glog.Warningf("alicloud: multipul instance found by nodename=[%s], "+
+		glog.Warningf("alicloud: multiple instances found by nodename=[%s], "+
 			"the first one will be used, instanceid=[%s]\n", string(nodeid), ins[0].InstanceId)
 	}
 	return &ins[0], nil

docs/getting-started.md

@@ -4,16 +4,16 @@
 - CloudNetwork: Only Alibaba Cloud VPC network is supported.
 
 
-## Deploy out-of-tree CloudProvider in Alibaba Cloud.
+## Deploy CloudProvider in Alibaba Cloud 
 
-### Set up a latest supported Kubernetes Cluster of version v1.10+ with Kubeadm.
+### Set up a supported Kubernetes Cluster using kubeadm
 
-Kubeadm is an official installation tool for kubernetes. You could bring up a single master kubernetes cluster by following the instruction in this [page](https://kubernetes.io/docs/setup/independent/create-cluster-kubeadm/).
+kubeadm is an official installation tool for kubernetes. You could bring up a single master kubernetes cluster by following the instruction in this [page](https://kubernetes.io/docs/setup/independent/create-cluster-kubeadm/).
 
 1. Install Docker or other CRI runtime: https://kubernetes.io/docs/setup/cri/
 2. Install kubeadm, kubelet and kubectl: https://kubernetes.io/docs/setup/independent/install-kubeadm/
 3. Update kubelet info with provider id info and restart kubelet: You should provide ```--hostname-override=${REGION_ID}.${INSTANCE_ID} --provider-id=${REGION_ID}.${INSTANCE_ID}``` arguments in all of your kubelet unit file. The format is ```${REGION_ID}.${INSTANCE_ID}```. See [kubelet.service](examples/kubelet.service) for more details.
-4. Init kubeadm: Be advised that kubeadm accept a serious of certain parameters to customize your cluster with kubeadm.conf file. If you want to use your own secure ETCD cluster or image repository, you may find the template [kubeadm.conf](examples/kubeadm.conf) or [kubeadm-new.conf for k8s 1.12+](examples/kubeadm-new.conf) is useful. 
+4. Init kubeadm: Be advised that kubeadm accepts a number of certain parameters to customize your cluster with kubeadm.conf file. If you want to use your own secure ETCD cluster or image repository, you may find the template [kubeadm.conf](examples/kubeadm.conf) or [kubeadm-new.conf for k8s 1.12+](examples/kubeadm-new.conf) is useful. 
 
 Run the command below to initialize a kubernetes cluster.
 ```$bash
@@ -29,22 +29,90 @@ $ META_EP=http://100.100.100.200/latest/meta-data
 $ echo `curl -s $META_EP/region-id`.`curl -s $META_EP/instance-id`
 ```
 
+### Set up a supported Kubernetes Cluster using rke 
 
-For now, you should have a running kubernetes cluster. Try some example command like ```kubectl get no ```
+Rancher Kubernetes Engine [rke](https://github.com/rancher/rke) is  another Kubernetes installer.
+
+1. Provision the nodes for your cluster in Alicloud
+2. Set up cluster.yml file for deploying the Kubernetes cluster. Make sure to provide hostname_override parameters and insert the hostnames of the cluster nodes using the region-id.instance-id format
+
+```
+nodes:
+  - address: nnn.nnn.nnn.nnn
+    user: root
+    role:
+      - controlplane
+      - etcd
+    ssh_key_path: ~/ssh.pem
+    internal_address: 172.16.1.29
+    hostname_override: "cn-hangzhou.i-bp18j8zzajt93rztiw9g" <- override hostname
+  
+  - address: nnn.nnn.nnn.nnn
+    user: root
+    role:
+      - worker
+    ssh_key_path: ~/ssh.pem
+    internal_address: 172.16.1.19
+    hostname_override: "cn-hangzhou.i-bp109r2aiuf935xxi4po" <- override hostname
+    labels:
+      rke.cattle.io/external: nnn.nnn.nnn.nnn
+  - address: nnn.nnn.nnn.nnn
+    user: root
+    role:
+      - worker
+    ssh_key_path: ~/ssh.pem
+    internal_address: 172.16.1.21
+    hostname_override: "cn-hangzhou.i-bp16uimj7fl6ze8q5rf3" <- override hostname
+    labels:
+      rke.cattle.io/external: nnn.nnn.nnn.nnn
+  
+addon_job_timeout: 90
+services:
+  kubelet:
+    extra_args:
+      node-status-update-frequency: 4s
+      volume-plugin-dir: /usr/libexec/kubernetes/kubelet-plugins/volume/exec
+    extra_binds:
+      - /usr/libexec/kubernetes/kubelet-plugins/volume/exec:/usr/libexec/kubernetes/kubelet-plugins/volume/exec
+  kube-api:
+    service_node_port_range: 10000-12500
+    extra_args:
+      default-not-ready-toleration-seconds: 30
+      default-unreachable-toleration-seconds: 30
+  kube-controller:
+    extra_args:
+      node-monitor-period: 2s
+      node-monitor-grace-period: 16s
+      pod-eviction-timeout: 30s
+    addon_job_timeout: 90
+``` 
+
+3. Provision the cluster using rke
+
+4. Currently rke does not support setting the providerID as a configuration option.
+   Once the cluster has been provisioned update the nodes and make sure providerID is set to REGION.NODEID
+
+   ```
+   kubectl patch node ${NODE_NAME} -p "{\"spec\":{\"providerID\": ${NODE_NAME} }}"
+   ```
+   For example 
+   ```
+   kubectl patch node cn-hangzhou.i-bp16uimj7fl6ze8q5rf3 -p "{\"spec\":{\"providerID\": cn-hangzhou.i-bp16uimj7fl6ze8q5rf3}}"
+   ```
 
 ### Install Alibaba CloudProvider support.
 
-CloudProvider needs certain permissions to access Alibaba Cloud, you will need to create a few RAM policies for your ECS instances or use AccessKeyID&Secret directly.
+CloudProvider needs certain permissions to access Alibaba Cloud, you will need to create a few RAM policies for your ECS instances or use AccessKeyID and AccessKeySecret directly.
 
 **RAM role Policy**
 
 [What is the RAM role of an instance](https://www.alibabacloud.com/help/doc-detail/54235.htm)
 
 The sample [master policy](examples/master.policy) is a bit open and can be scaled back depending on the use case. Adjust these based on your needs.
 
-**AccessKeyID & AccessKeySecret**
+**AccessKeyID and AccessKeySecret**
 
-Or we use Alibaba AccessKeyID&Secret to authorize the CloudProvider. Please make sure that the AccessKeyID has the listed permissions in [master.policy](examples/master.policy)
+Or we use Alibaba AccessKeyID AccessKeySecret to authorize the CloudProvider. Please make sure that the AccessKeyID has the listed permissions in [master.policy](examples/master.policy)
 
 [How to get AccessKey?](https://usercenter.console.aliyun.com/#/manage/ak)