This guide will take you through the steps of deploying Kubernetes to Openstack using kube-up.sh
. The primary mechanisms for this are OpenStack Heat and the SaltStack distributed with Kubernetes.
The default OS is CentOS 7, this has not been tested on other operating systems.
This guide assumes you have access to a working OpenStack cluster with the following features:
By default this provider provisions 4 m1.medium
instances. If you do not have resources available, please see the Set additional configuration values section for information on reducing the footprint of your cluster.
If you already have the required versions of the OpenStack CLI tools installed and configured, you can move on to the Starting a cluster section.
sudo pip install -U --force 'python-openstackclient==3.11.0'
sudo pip install -U --force 'python-heatclient==1.10.0'
sudo pip install -U --force 'python-swiftclient==3.3.0'
sudo pip install -U --force 'python-glanceclient==2.7.0'
sudo pip install -U --force 'python-novaclient==9.0.1'
Please talk to your local OpenStack administrator for an openrc.sh
file.
Once you have that file, source it into your environment by typing
. ~/path/to/openrc.sh
This provider will consume the correct variables to talk to OpenStack and turn-up the Kubernetes cluster.
Otherwise, you must set the following appropriately:
export OS_USERNAME=username
export OS_PASSWORD=password
export OS_TENANT_NAME=projectName
export OS_AUTH_URL=https://identityHost:portNumber/v2.0
export OS_TENANT_ID=tenantIDString
export OS_REGION_NAME=regionName
In addition, here are some commonly changed variables specific to this provider, with example values. Under most circumstances you will not have to change these. Please see the files in the next section for a full list of options.
export STACK_NAME=KubernetesStack
export NUMBER_OF_MINIONS=3
export MAX_NUMBER_OF_MINIONS=3
export MASTER_FLAVOR=m1.small
export MINION_FLAVOR=m1.small
export EXTERNAL_NETWORK=public
export DNS_SERVER=8.8.8.8
export IMAGE_URL_PATH=http://cloud.centos.org/centos/7/images
export IMAGE_FILE=CentOS-7-x86_64-GenericCloud-1510.qcow2
export SWIFT_SERVER_URL=http://192.168.123.100:8080
export ENABLE_PROXY=false
If you do not have your environment variables set, or do not want them consumed, modify the variables in the following files under cluster/openstack-heat
:
Please see the contents of these files for documentation regarding each variable’s function.
Once you’ve installed the OpenStack CLI tools and have set your OpenStack environment variables, issue this command:
export KUBERNETES_PROVIDER=openstack-heat; curl -sS https://get.k8s.io | bash
Alternatively, you can download a Kubernetes release of version 1.3 or higher and extract the archive. To start your cluster, open a shell and run:
cd kubernetes # Or whichever path you have extracted the release to
KUBERNETES_PROVIDER=openstack-heat ./cluster/kube-up.sh
Or, if you are working from a checkout of the Kubernetes code base, and want to build/test from source:
cd kubernetes # Or whatever your checkout root directory is called
make clean
make quick-release
KUBERNETES_PROVIDER=openstack-heat ./cluster/kube-up.sh
Once kube-up is finished, your cluster should be running:
./cluster/kubectl.sh get cs
NAME STATUS MESSAGE ERROR
controller-manager Healthy ok
scheduler Healthy ok
etcd-1 Healthy {"health": "true"}
etcd-0 Healthy {"health": "true"}
You can also list the nodes in your cluster:
./cluster/kubectl.sh get nodes
NAME STATUS AGE VERSION
kubernetesstack-node-ojszyjtr Ready 42m v1.6.0+fff5156
kubernetesstack-node-tzotzcbp Ready 46m v1.6.0+fff5156
kubernetesstack-node-uah8pkju Ready 47m v1.6.0+fff5156
Being a new cluster, there will be no pods or replication controllers in the default namespace:
./cluster/kubectl.sh get pods
./cluster/kubectl.sh get replicationcontrollers
You are now ready to create Kubernetes objects.
For a simple test, issue the following command:
./cluster/kubectl.sh run nginx --image=nginx --generator=run-pod/v1
Soon, you should have a running nginx pod:
./cluster/kubectl.sh get pods
NAME READY STATUS RESTARTS AGE
nginx 1/1 Running 0 5m
Once the nginx pod is running, use the port-forward command to set up a proxy from your machine to the pod.
./cluster/kubectl.sh port-forward nginx 8888:80
You should now see nginx on http://localhost:8888.
For more complex examples please see the examples directory.
You can manage the nodes in your cluster using the OpenStack CLI Tools.
First, set your environment variables:
. cluster/openstack-heat/config-default.sh
. cluster/openstack-heat/openrc-default.sh
To get all information about your cluster, use heat:
openstack stack show $STACK_NAME
To see a list of nodes, use nova:
nova list --name=$STACK_NAME
See the OpenStack CLI Reference for more details.
The OpenStack-Heat provider uses a standalone Salt configuration. It only uses Salt for bootstraping the machines and creates no salt-master and does not auto-start the salt-minion service on the nodes.
Your public key was added during the cluster turn-up, so you can easily ssh to them for troubleshooting purposes.
ssh minion@IP_ADDRESS
You may find the need to modify environment variables to change the behaviour of kube-up. Here are some common scenarios:
If you are behind a proxy, and have your local environment variables setup, you can use these variables to setup your Kubernetes cluster:
ENABLE_PROXY=true KUBERNETES_PROVIDER=openstack-heat ./cluster/kube-up.sh
Some deployments differ from the default Swift URL:
SWIFT_SERVER_URL="http://10.100.0.100:8080" KUBERNETES_PROVIDER=openstack-heat ./cluster/kube-up.sh
Sometimes the name of the public network differs from the default public
:
EXTERNAL_NETWORK="network_external" KUBERNETES_PROVIDER=openstack-heat ./cluster/kube-up.sh
You may want to spin up another cluster within your OpenStack project. Use the $STACK_NAME
variable to accomplish this.
STACK_NAME=k8s-cluster-2 KUBERNETES_PROVIDER=openstack-heat ./cluster/kube-up.sh
For more configuration examples, please browse the files mentioned in the Configuration section.
To bring down your cluster, issue the following command:
KUBERNETES_PROVIDER=openstack-heat ./cluster/kube-down.sh
If you have changed the default $STACK_NAME
, you must specify the name. Note that this will not remove any Cinder volumes created by Kubernetes.
IaaS Provider | Config. Mgmt | OS | Networking | Docs | Conforms | Support Level |
---|---|---|---|---|---|---|
OpenStack Heat | Saltstack | CentOS | Neutron + flannel hostgw | docs | Community (@FujitsuEnablingSoftwareTechnologyGmbH) |
For support level information on all solutions, see the Table of solutions chart.
Create an Issue Edit this Page