OpenShift - Basic Concept


Advertisements

Before beginning with the actual setup and deployment of applications, we need to understand some basic terms and concepts used in OpenShift V3.

Containers and Images

Images

These are the basic building blocks of OpenShift, which are formed out of Docker images. In each pod on OpenShift, the cluster has its own images running inside it. When we configure a pod, we have a field which will get pooled from the registry. This configuration file will pull the image and deploy it on the cluster node.

apiVersion: v1
kind: pod
metadata:
   name: Tesing_for_Image_pull -----------> Name of Pod
      spec:
containers:
- name: neo4j-server ------------------------> Name of the image
image: <Name of the Docker image>----------> Image to be pulled
imagePullPolicy: Always ------------->Image pull policy
command: [“echo”, “SUCCESS”] -------------------> Massage after image pull

In order to pull and create an image out of it, run the following command. OC is the client to communicate with OpenShift environment after login.

$ oc create –f Tesing_for_Image_pull

Container

This gets created when the Docker image gets deployed on the OpenShift cluster. While defining any configuration, we define the container section in the configuration file. One container can have multiple images running inside and all the containers running on cluster node are managed by OpenShift Kubernetes.

spec:
   containers:
   - name: py ------------------------> Name of the container
   image: python----------> Image going to get deployed on container
   command: [“python”, “SUCCESS”]
   restartPocliy: Never --------> Restart policy of container

Following are the specifications for defining a container having multiple images running inside it.

apiVersion: v1
kind: Pod
metadata:
   name: Tomcat
spec:
   containers:
   - name: Tomcat
   image: tomcat: 8.0
   ports:
   - containerPort: 7500
      imagePullPolicy: Always
      -name: Database
      Image: mongoDB
      Ports:
      - containerPort: 7501
imagePullPolicy: Always

In the above configuration, we have defined a multi-container pod with two images of Tomcat and MongoDB inside it.

Pods and Services

Pods

Pod can be defined as a collection of container and its storage inside a node of OpenShift (Kubernetes) cluster. In general, we have two types of pod starting from a single container pod to multi-container pod.

Single Container Pod − These can be easily created with OC command or by a basic configuration yml file.

$ oc run <name of pod> --image = <name of the image from registry>

Create it with a simple yaml file as follows.

apiVersion: v1
kind: Pod
metadata:
   name: apache
spec:
   containers:
   - name: apache
   image: apache: 8.0
   ports:
      - containerPort: 7500
imagePullPolicy: Always

Once the above file is created, it will generate a pod with the following command.

$ oc create –f apache.yml

Multi-Container Pod − Multi-container pods are those in which we have more than one container running inside it. They are created using yaml files as follows.

apiVersion: v1
kind: Pod
metadata:
   name: Tomcat
spec:
   containers:
   - name: Tomcat
   image: tomcat: 8.0
   ports:
      - containerPort: 7500
imagePullPolicy: Always
   -name: Database
   Image: mongoDB
   Ports:
      - containerPort: 7501
imagePullPolicy: Always

After creating these files, we can simply use the same method as above to create a container.

Service − As we have a set of containers running inside a pod, in the same way we have a service that can be defined as a logical set of pods. It’s an abstracted layer on top of the pod, which provides a single IP and DNS name through which pods can be accessed. Service helps in managing the load balancing configuration and to scale the pod very easily. In OpenShift, a service is a REST object whose deification can be posted to apiService on OpenShift master to create a new instance.

apiVersion: v1
kind: Service
metadata:
   name: Tutorial_point_service
spec:
   ports:
      - port: 8080
         targetPort: 31999

Builds and Streams

Builds

In OpenShift, build is a process of transforming images into containers. It is the processing which converts the source code to an image. This build process works on pre-defined strategy of building source code to image.

The build processes multiple strategies and sources.

Build Strategies

  • Source to Image − This is basically a tool, which helps in building reproducible images. These images are always in a ready stage to run using the Docker run command.

  • Docker Build − This is the process in which the images are built using Docker file by running simple Docker build command.

  • Custom Build − These are the builds which are used for creating base Docker images.

Build Sources

Git − This source is used when the git repository is used for building images. The Dockerfile is optional. The configurations from the source code looks like the following.

source:
type: "Git"
git:
   uri: "https://github.com/vipin/testing.git"
   ref: "master"
contextDir: "app/dir"
dockerfile: "FROM openshift/ruby-22-centos7\nUSER example"

Dockerfile − The Dockerfile is used as an input in the configuration file.

source:
   type: "Dockerfile"
   dockerfile: "FROM ubuntu: latest
   RUN yum install -y httpd"

Image Streams − Image streams are created after pulling the images. The advantage of an image stream is that it looks for updates on the new version of an image. This is used to compare any number of Docker formatted container images identified by tags.

Image streams can automatically perform an action when a new image is created. All the builds and deployments can watch for image action and perform an action accordingly. Following is how we define a build a stream.

apiVersion: v1
kind: ImageStream
metadata:
   annotations:
      openshift.io/generated-by: OpenShiftNewApp
   generation: 1
   labels:
      app: ruby-sample-build
   selflink: /oapi/v1/namespaces/test/imagestreams/origin-ruby-sample
   uid: ee2b9405-c68c-11e5-8a99-525400f25e34
spec: {}
status:
   dockerImageRepository: 172.30.56.218:5000/test/origin-ruby-sample
   tags:
   - items:
      - created: 2016-01-29T13:40:11Z
      dockerImageReference: 172.30.56.218:5000/test/origin-apache-sample
      generation: 1
      image: vklnld908.int.clsa.com/vipin/test
   tag: latest

Routes and Templates

Routes

In OpenShift, routing is a method of exposing the service to the external world by creating and configuring externally reachable hostname. Routes and endpoints are used to expose the service to the external world, from where the user can use the name connectivity (DNS) to access defined application.

In OpenShift, routes are created by using routers which are deployed by OpenShift admin on the cluster. Routers are used to bind HTTP (80) and https (443) ports to external applications.

Following are the different kinds of protocol supported by routes −

  • HTTP
  • HTTPS
  • TSL and web socket

When configuring the service, selectors are used to configure the service and find the endpoint using that service. Following is an example of how we create a service and the routing for that service by using an appropriate protocol.

{
   "kind": "Service",
   "apiVersion": "v1",
   "metadata": {"name": "Openshift-Rservice"},
   "spec": {
      "selector": {"name":"RService-openshift"},
      "ports": [
         {
            "protocol": "TCP",
            "port": 8888,
            "targetPort": 8080
         }
      ]
   }
}

Next, run the following command and the service is created.

$ oc create -f ~/training/content/Openshift-Rservice.json

This is how the service looks like after creation.

$ oc describe service Openshift-Rservice

Name:              Openshift-Rservice
Labels:            <none>
Selector:          name = RService-openshift
Type:              ClusterIP
IP:                172.30.42.80
Port:              <unnamed> 8080/TCP
Endpoints:         <none>
Session Affinity:  None
No events.

Create a routing for service using the following code.

{
   "kind": "Route",
   "apiVersion": "v1",
   "metadata": {"name": "Openshift-service-route"},
   "spec": {
      "host": "hello-openshift.cloudapps.example.com",
      "to": {
         "kind": "Service",
         "name": "OpenShift-route-service"
      },
      "tls": {"termination": "edge"}
   }
}

When OC command is used to create a route, a new instance of route resource is created.

Templates

Templates are defined as a standard object in OpenShift which can be used multiple times. It is parameterized with a list of placeholders which are used to create multiple objects. This can be used to create anything, starting from a pod to networking, for which users have authorization to create. A list of objects can be created, if the template from CLI or GUI interface in the image is uploaded to the project directory.

apiVersion: v1
kind: Template
metadata:
   name: <Name of template>
   annotations:
      description: <Description of Tag>
      iconClass: "icon-redis"
      tags: <Tages of image>
objects:
   - apiVersion: v1
   kind: Pod
   metadata:
      name: <Object Specification>
spec:
   containers:
      image: <Image Name>
      name: master
      ports:
      - containerPort: <Container port number>
         protocol: <Protocol>
labels:
   redis: <Communication Type>

Authentication and Authorization

Authentication

In OpenShift, while configuring master and client structure, master comes up with an inbuilt feature of OAuth server. OAuth server is used for generating tokens, which is used for authentication to the API. Since, OAuth comes as a default setup for master, we have the Allow All identity provider used by default. Different identity providers are present which can be configured at /etc/openshift/master/master-config.yaml.

There are different types of identity providers present in OAuth.

  • Allow All
  • Deny All
  • HTPasswd
  • LDAP
  • Basic Authentication

Allow All

apiVersion: v1
   kind: Pod
   metadata:
      name: redis-master
   spec:
      containers:
         image: dockerfile/redis
         name: master
      ports:
      - containerPort: 6379
         protocol: TCP
      oauthConfig:
      identityProviders:
      - name: my_allow_provider
         challenge: true
         login: true
      provider:
         apiVersion: v1
         kind: AllowAllPasswordIdentityProvider

Deny All

apiVersion: v1
kind: Pod
metadata:
   name: redis-master
spec:
   containers:
      image: dockerfile/redis
   name: master
   ports:
   - containerPort: 6379
      protocol: TCP
   oauthConfig:
   identityProviders:
   - name: my_allow_provider
      challenge: true
      login: true
   provider:
      apiVersion: v1
      kind: DenyAllPasswordIdentityProvider

HTPasswd

In order to use HTPasswd, we need to first set up Httpd-tools on the master machine and then configure it in the same way as we did for others.

identityProviders:
   - name: my_htpasswd_provider
      challenge: true
      login: true
      provider:
         apiVersion: v1
         kind: HTPasswdPasswordIdentityProvider

Authorization

Authorization is a feature of OpenShift master, which is used to validate for validating a user. This means that it checks the user who is trying to perform an action to see if the user is authorized to perform that action on a given project. This helps the administrator to control access on the projects.

Authorization policies are controlled using −

  • Rules
  • Roles
  • Bindings

Evaluation of authorization is done using −

  • Identity
  • Action
  • Bindings

Using Policies −

  • Cluster policy
  • Local policy
Advertisements