The multi-cloud kubernetes provisioner component of the Cross-Cloud CI project.
A Kubernetes provisioner supporting multiple clouds (eg. AWS, Azure, Google, Packet) which
- Creates K8s clusters on cloud providers
- Supplies conformance validated Kubernetes end-points for each cloud provider with cloud specific features enabled
You have to have a working Docker environment
Note: 147.75.69.23 is the IP address of the DNS server for Cross Cloud deployed Nodes. Should you wish to be able to reach your Nodes by name from outside the cluster, that IP needs to be in your /etc/resolv.conf but it is not a delegating resolver, so it shouldn't be the only nameserver in your resolv.conf.
Pre-reqs: IAM User with the following Permissions:
- AmazonEC2FullAccess
- AmazonS3FullAccess
- AmazonRoute53DomainsFullAccess
- AmazonRoute53FullAccess
- IAMFullAccess
- IAMUserChangePassword
AWS credentials
export AWS_ACCESS_KEY_ID="YOUR_AWS_KEY_ID"
export AWS_SECRET_ACCESS_KEY="YOUR_AWS_SECRET_KEY"
export AWS_DEFAULT_REGION=”YOUR_AWS_DEFAULT_REGION” # eg. ap-southeast-2
Run the following to provision a Kubernetes cluster on AWS:
docker run \
-v /tmp/data:/cncf/data \
-e NAME=cross-cloud \
-e CLOUD=aws \
-e COMMAND=deploy \
-e BACKEND=file \
-e AWS_ACCESS_KEY_ID=$AWS_ACCESS_KEY_ID \
-e AWS_SECRET_ACCESS_KEY=$AWS_SECRET_ACCESS_KEY \
-e AWS_DEFAULT_REGION=$AWS_DEFAULT_REGION \
-ti registry.cncf.ci/cncf/cross-cloud/provisioning:production
Pre-reqs: Project created on Google Cloud (eg. test-cncf-cross-cloud)
Google Cloud JSON configuration file for authentication. (This file is downloaded directly from the Google Developers Console)
- Log into the Google Developers Console and select a project.
- The API Manager view should be selected, click on "Credentials" on the left, then "Create credentials," and finally "Service account key."
- Select "Compute Engine default service account" in the "Service account" dropdown, and select "JSON" as the key type.
- Clicking "Create" will download your credentials.
- Rename this file to credentials-gce.json and move to your home directory (~/credentials-gce.json)
Google Project ID
- Log into the Google Developers Console to be sent to the Google API library page screen
- Click the
Select a project
drop-down in the upper left - Copy the Project ID for the desired project from the window that shows Eg. test-cncf-cross-cloud
Run the following to provision a Kubernetes cluster on GCE:
export GOOGLE_CREDENTIALS=$(cat ~/credentials-gce.json)
docker run \
-v /tmp/data:/cncf/data \
-e NAME=cross-cloud \
-e CLOUD=gce \
-e COMMAND=deploy \
-e BACKEND=file \
-e GOOGLE_REGION=us-central1 \
-e GOOGLE_ZONE=us-central1-a \
-e GOOGLE_PROJECT=test-cncf-cross-cloud \
-e GOOGLE_CREDENTIALS="${GOOGLE_CREDENTIALS}" \
-ti registry.cncf.ci/cncf/cross-cloud/provisioning:production
You will need a full set of credentials for an OpenStack cloud, including authentication endpoint.
Run the following to provision an OpenStack cluster:
docker run \
-v $(pwd)/data:/cncf/data \
-e NAME=cross-cloud \
-e CLOUD=openstack \
-e COMMAND=deploy \
-e BACKEND=file \
-e TF_VAR_os_auth_url=$OS_AUTH_URL \
-e TF_VAR_os_region_name=$OS_REGION_NAME \
-e TF_VAR_os_user_domain_name=$OS_USER_DOMAIN_NAME \
-e TF_VAR_os_username=$OS_USERNAME \
-e TF_VAR_os_project_name=$OS_PROJECT_NAME \
-e TF_VAR_os_password=$OS_PASSWORD \
-ti registry.cncf.ci/cncf/cross-cloud/provisioning:production
You will need a full set of credentials for an OpenStack cloud, including authentication endpoint.
Run the following to provision an OpenStack cluster:
docker run \
-v $(pwd)/data:/cncf/data \
-e NAME=cross-cloud \
-e CLOUD=openstack \
-e COMMAND=deploy \
-e BACKEND=file \
-e TF_VAR_os_auth_url=$OS_AUTH_URL \
-e TF_VAR_os_region_name=$OS_REGION_NAME \
-e TF_VAR_os_user_domain_name=$OS_USER_DOMAIN_NAME \
-e TF_VAR_os_username=$OS_USERNAME \
-e TF_VAR_os_project_name=$OS_PROJECT_NAME \
-e TF_VAR_os_password=$OS_PASSWORD \
-ti registry.cncf.ci/cncf/cross-cloud/provisioning:ci-stable-v0-2-0
The vSphere provider requires vSphere host and credential information, as well as credentials for the linked AWS account:
Run the following to provision a vSphere cluster:
docker run \
--rm \
--dns 147.75.69.23 --dns 8.8.8.8 \
-v $(pwd)/data:/cncf/data \
-e NAME=cross-cloud \
-e CLOUD=vsphere \
-e COMMAND=deploy \
-e BACKEND=file \
-e VSPHERE_SERVER=$VSPHERE_SERVER \
-e VSPHERE_USER=$VSPHERE_USER \
-e VSPHERE_PASSWORD=$VSPHERE_PASSWORD \
-e VSPHERE_AWS_ACCESS_KEY_ID=$VSPHERE_AWS_ACCESS_KEY_ID \
-e VSPHERE_AWS_SECRET_ACCESS_KEY=$VSPHERE_AWS_SECRET_ACCESS_KEY \
-e VSPHERE_AWS_REGION=$VSPHERE_AWS_REGION \
-ti registry.cncf.ci/cncf/cross-cloud/provisioning:ci-stable-v0-2-0
Packet.net requires an auth token and a project id.
To deploy to packet:
docker run \
-v /tmp/data:/cncf/data \
--dns 147.75.69.23 --dns 8.8.8.8 \
-e NAME=cross-cloud \
-e CLOUD=packet \
-e COMMAND=deploy \
-e BACKEND=file \
-e PACKET_AUTH_TOKEN=${PACKET_AUTH_TOKEN} \
-e TF_VAR_packet_project_id=${PACKET_PROJECT_ID} \
-ti registry.cncf.ci/cncf/cross-cloud/provisioning:production
To destroy your cluster in packet:
docker run \
-v /tmp/data:/cncf/data \
--dns 147.75.69.23 --dns 8.8.8.8 \
-e NAME=cross-cloud \
-e CLOUD=packet \
-e COMMAND=destroy \
-e BACKEND=file \
-e PACKET_AUTH_TOKEN=${PACKET_AUTH_TOKEN} \
-e TF_VAR_packet_project_id=${PACKET_PROJECT_ID} \
-ti registry.cncf.ci/cncf/cross-cloud/provisioning:production
Note: 147.75.69.23 is the IP address of the DNS server for Cross Cloud deployed Nodes. Should you wish to be able to reach your Nodes by name from outside the cluster, that IP needs to be in your /etc/resolv.conf but it is not a delegating resolver, so it shouldn't be the only nameserver in your resolv.conf
Minimum required configuration to use Cross-cloud to deploy a Kubernetes cluster on Cloud X.
docker run \
-v /tmp/data:/cncf/data \
-e NAME=cross-cloud
-e CLOUD=<aws|gke|gce|openstack|packet> \
-e COMMAND=<deploy|destory> \
-e BACKEND=<file|s3> \
<CLOUD_SPECIFIC_OPTIONS>
<KUBERNETES_CLUSTER_OPTIONS>
-ti registry.cncf.ci/cncf/cross-cloud/provisioning:production
- -e CLOUD=<aws|gke|gce|openstack|packet> # Choose the cloud provider. Then add the appropriate cloud specific options below.
- -e COMMAND=<deploy|destory>
- -e BACKEND=<file|s3> # File will store the Terraform State file to Disk / S3 will store the Terraform Statefile to a AWS s3 Bucket
AWS:
- -e AWS_ACCESS_KEY_ID=secret
- -e AWS_SECRET_ACCESS_KEY=secret
- -e AWS_DEFAULT_REGION=ap-southeast-2
Packet:
- -e PACKET_AUTH_TOKEN=secret
- -e TF_VAR_packet_project_id=secret
GCE/GKE:
- -e GOOGLE_CREDENTIALS=secret
- -e GOOGLE_REGION=us-central1
- -e GOOGLE_PROJECT=test-163823
- -e GOOGLE_ZONE=us-central1-a
OpenStack:
- -e TF_VAR_os_auth_url=$OS_AUTH_URL
- -e TF_VAR_os_region_name=$OS_REGION_NAME
- -e TF_VAR_os_user_domain_name=$OS_USER_DOMAIN_NAME
- -e TF_VAR_os_username=$OS_USERNAME
- -e TF_VAR_os_project_name=$OS_PROJECT_NAME
- -e TF_VAR_os_password=$OS_PASSWORD
vSphere via VMware Cloud (VMC) on AWS:
- -e VSPHERE_SERVER=1.2.3.4
- -e VSPHERE_USER=admin
- -e VSPHERE_PASSWORD=notblank
- -e VSPHERE_AWS_ACCESS_KEY_ID=public
- -e VSPHERE_AWS_SECRET_ACCESS_KEY=secret
- -e VSPHERE_AWS_REGION=us-west-2
Custom Configuration options for the Kubernetes Cluster:
- -e TF_VAR_pod_cidr=10.2.0.0/16 # Set the Kubernetes Cluster POD CIDR
- -e TF_VAR_service_cidr=10.0.0.0/24 # Set the Kubernetes Cluster SERVICE CIDR
- -e TF_VAR_worker_node_count=3 # Set the Number of Worker nodes to be Deployed in the Cluster
- -e TF_VAR_master_node_count=3 # Set the Number of Master nodes to be Deployed in the Cluster
- -e TF_VAR_dns_service_ip=10.0.0.10 # Set the Kubernetes DNS Service IP
- -e TF_VAR_k8s_service_ip=10.0.0.1 # Set the Kubernetes Service IP
- FAQ - Frequently Asked Questions