This page captures the information and steps that are needed to deploy ONAP using OOM on VIO 4.0 Kubernetes .
This pages assumes that VIO4.0 is deployed successfully with Kubernetes. This will use OpenStack as the Cloud Provider.
VMware VIO 4.0 Kubernetes Architecture
a. Prerequisites -
a.1 BlueShift Mgmt IP Address, username and password are needed to create Kubernetes cluster .
a.2 K8S Master and K8S Node instances should have a OpenStack Flavor attached as per the requirement below .
| vCPU | 48 |
| RAM | 96GB |
| Storage | 256GB |
Follow the steps 1 to 3 mentioned below for creating a Kubernetes cluster .
TOComplete 20171205 - Step 4 onwards are needed for getting the Kubernetes Host IP so that user can login to Kubernetes Host and ONAP using OOM can be deployed ..
TODO 20171205 - steps that needs to be done prior to git Clone of OOM .
Step1 - https://BLUESHIFT_MGMT_IP_ADDRESS/LOGIN
Step2 - Create the Cloud Provider before creating a kubernetes cluster
Cloud Provider creation is prerequistie to Kubernetes cluster creation .VMware Integrated OpenStack with Kubernetes uses the cloud provider to create the infrastructure required to deploy all your Kubernetes clusters. VMware currently supports 2 options for infrastructure provider. VMware SDDC (vSphere + NSX + VSAN) or VIO (VMware Integrated OpenStack). When choosing the type of provider to create, consider the following:
- With an existing VMware Integrated OpenStack deployment, you can create an OpenStack provider.
- Without an existing VMware Integrated OpenStack deployment, you can create an SDDC provider Im.
Step3 - Create the kubernetes cluster by clicking + NEW
Step4 - Creating the Kubernetes cluster
4.1 - Click NEXT
4.2 - Select an Infrastructure Provider for creating the kubernetes
Before you deploy a Kubernetes cluster, you must have created the cloud provider. VMware Integrated OpenStack with Kubernetes uses the cloud provider to create the infrastructure required to deploy all your Kubernetes clusters. VMware currently supports 2 options for infrastructure provider. VMware SDDC (vSphere + NSX + VSAN) or VIO (VMware Integrated OpenStack). When choosing the type of provider to create, consider the following:
- With an existing VMware Integrated OpenStack deployment, you can create an OpenStack provider.
- Without an existing VMware Integrated OpenStack deployment, you can create an SDDC provider Im.
Provider name can be custom name given by user where as provider Type has to be openstack or sddc .
4.3 - Select a Node Profile. If you have more than one node profiles, uncheck the box "Use default node profile" to see the list.
4.4 - Provide the Input for the Cluster as indicated in Example Data below
Node Types - A Kubernetes cluster is comprised of two types of nodes. Each node in the VMware Integrated OpenStack with Kubernetes is a VM.
- Master Nodes - A master node provides the Kubernetes API service, scheduler, replicator, and so on. It manages the worker nodes. A cluster with a single master node is valid but has no redundancy.
- Worker Nodes - A worker node hosts your containers. A cluster with a single worker node is valid but has no redundancy.
Cluster Types - VMware Integrated OpenStack with Kubernetes supports two types of clusters.
- Exclusive Cluster - In an exclusive cluster, multi-tenancy is not supported. Any authorized Kubernetes user using the Kubernetes CLI of APIs has namespace management privileges. The exclusive cluster provides a familiar environment for developers who deploy Kubernetes themselves.
- Shared Cluster - In a shared cluster, multi-tenancy is supported and enforced by the Kubernetes namespace. Only a VMware Integrated OpenStack with Kubernetes administrator using the VMware Integrated OpenStack with Kubernetes interface or CLI has namespace management privileges. The shared cluster is an environment where the administrator can manage resource isolation among users.
Example Data:
Cluster Name :testCluster
Number of Master Nodes :1
Number of worker nodes :1
DNS servers : 10.112.64.1
Cluster type: Exclusive Cluster
3.5 - Fill the details in User & Group page
3.6 - Click on FINISH
Step4. Verifying the VIO Kubernetes Cluster - If everything in step 3.1 to step 3.6 has been done successfully .The Summary information for the cluster will be filled as given in the example below .
Step5. How to get Kubernetes Host IP Address and login to Kubernetes Host .
5.1 - Login via console window to BLUESHIFT_MGMT_IP_ADDRESS . user name and password is same as used in step1 .
5.2 - Once logged into BLUESHIFT_MGMT_IP_ADDRESS session use the command "vkube login --insecure" . use the user name and password same as in step1 .
5.3 - Get the list of clusters using command "vkube cluster list --insecure" . make a note of cluster Id in the output of the command
5.4 - Get the cluster node details using command "vkube cluster show <cluster Id > --insecure " .make note of worker ip address .Worker IP Address KUBERNETES_HOST_IP.
5.5 - Login to app-api docker using docker exec it app-api bash
5.6 - Once inside the app-api docker go to /var/lib/vrc/terraform/<cluserId>
5.7 - ssh to KUBERNETES_HOST_IP_ADDRESS using below command
ssh -i private.key -F ssh.bastion.conf ubuntu@KUBERNETES_HOST_IP_ADDRESS
5.8 - Once inside KUBERNETES_HOST_IP_ADDRESS (to be reviewed)
Step6. Installing kubectl to manage Kubernetes cluster
TBC20171207
6.1 - Download the kubectl using below command
curl -LO https://storage.googleapis.com/kubernetes-release/release/v1.7.0/bin/linux/amd64/kubectl
6.2 - Make the kubectl binary executable.
chmod +x ./kubect
6.3 - Move the kubectl to PATH
sudo mv ./kubectl /usr/local/bin/kubectl
Step7. Verifying that kubernetes config is good
root@localhost:~# kubectl cluster-info
Kubernetes master is running at https://10.110.208.207:443/
dnsmasq is running at https://10.110.208.207:443//api/v1/namespaces/kube-system/services/dnsmasq/proxy
kubedns is running at https://10.110.208.207:443//api/v1/namespaces/kube-system/services/kubedns/proxy
Step8. Installing Helm
8.1 - Download the helm using below command
wget http://storage.googleapis.com/kubernetes-helm/helm-v2.3.0-linux-amd64.tar.gz
8.2 - Untar the file
tar -zxvf helm-v2.3.0-linux-amd64.tar.gz
8.3 - Move the helm to /usr/local/bin
sudo mv linux-amd64/helm /usr/local/bin/helm