Kubernetes for Java Developers

Create Kubernetes Cluster

https://github.com/arun-gupta/kubernetes-java-sample/blob/master/workshop.adoc#create-kubernetes-cluster

Create Resources

Pod

Using CLI

Create pod: kubectl run wildfly --image=jboss/wildfly --port=8080
Check status: kubectl get -w pods
1. Show the pod name derived from the Deployment name
Describe pod: kubectl describe pod <pod-name> or kubectl describe pod/<pod-name>
Get logs: kubectl logs <pod-name>
1. Use -f to tail logs
Get deployments: kubectl get deployments
Delete deployments: kubectl delete deployments/wildfly
Get deployments: kubectl get deployments
Get pods: kubectl get pods

Using Configuration File

Create pod: kubectl create -f wildfly-pod.yaml
Get pods: kubectl get pods
1. Wide: kubectl get pods -o wide
2. YAML: kubectl get pods -o yaml
3. JSON: kubectl get pods -o json
Describe pod: kubectl describe pod/wildfly-pod
Delete pod: kubectl delete pod/wildfly-pod
Get pods: kubectl get pods

Replication Controller

Create RC: kubectl create -f wildfly-rc.yaml
Get RC: kubectl get rc
Get pods: kubectl get pods
1. Show pod names
2. Show image id downloaded on the second host
Get pods created by RC: kubectl describe pods wildfly-rc
Get pods with a specified label: kubectl get pods -l name=wildfly
Delete RC: kubectl delete rc wildfly-rc

Service

Create Service: kubectl create -f wildfly-service.yaml
Get RC, Service and Pods: kubectl get rc,service,pods
Describe service: kubectl describe service wildfly-service
If Kubernetes is on cloud with LB
1. Get the value of LoadBalancer Ingress, access the WildFly landing page at <IP>:8080
Delete multiple resources: kubectl delete rc/wildfly-rc service/wildfly-service

Load Balancing

ClusterIP: This is default, exposes service on cluster-internal IP.

Create: kubectl create -f lb-clusterip.yaml

Describe: kubectl get svc

NAME              CLUSTER-IP    EXTERNAL-IP   PORT(S)    AGE
kubernetes        100.64.0.1    <none>        443/TCP    48m
wildfly-service   100.71.85.1   <none>        8080/TCP   13m

Details: kubectl describe svc wildfly-service

kubectl describe svc wildfly-service
Name:			wildfly-service
Namespace:		default
Labels:			<none>
Annotations:		<none>
Selector:		app=wildfly-rc-pod
Type:			ClusterIP
IP:			100.71.85.1
Port:			web	8080/TCP
Endpoints:		100.96.1.4:8080,100.96.2.3:8080
Session Affinity:	None
Events:			<none>

Access: Service is accessible only inside the cluster.
1. Expose the service: kubectl expose service wildfly-service --port=8080 --target-port=8080 --name=web
2. Start Kubernetes proxy: kubectl proxy
3. Access the service: curl http://localhost:8001/api/v1/proxy/namespaces/default/services/web/index.html

Delete:

kubectl delete -f lb-clusterip.yaml
kubectl delete svc/web

NodePort: Expose service on each node of the cluster at a static port.
1. Create: kubectl create -f lb-nodeport.yaml
2. Describe: kubectl get svc
  NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE kubernetes 100.64.0.1 <none> 443/TCP 2h wildfly-service 100.68.222.70 <nodes> 8080:32233/TCP 29m
  Details: kubectl describe svc wildfly-service
  Name: wildfly-service Namespace: default Labels: <none> Annotations: <none> Selector: app=wildfly-rc-pod Type: NodePort IP: 100.68.222.70 Port: web 8080/TCP NodePort: web 32233/TCP Endpoints: 100.96.1.13:8080,100.96.2.12:8080 Session Affinity: None Events: <none>
3. Access: Service is accessible using <NodeIP>:<NodePort>. <NodePort> for us is 32233.
  1. Configure the worker node firewall to create a Custom TCP Rule to allow port 32233 accessible from Anywhere.
  2. Find IP address of the worker nodes using AWS Console.
  3. Service is accessible at <worker-node-ip>:32233.
4. Delete: kubectl delete -f lb-nodeport.yaml

LoadBalancer: Expose the service using a cloud provider’s load balancer.

Create: kubectl create -f lb-loadbalancer.yaml

Describe: kubectl get svc

NAME              CLUSTER-IP       EXTERNAL-IP        PORT(S)          AGE
kubernetes        100.64.0.1       <none>             443/TCP          2h
wildfly-service   100.65.225.120   afa8056b14fc9...   8080:30229/TCP   4s

Details:

Name:			wildfly-service
Namespace:		default
Labels:			<none>
Annotations:		<none>
Selector:		app=wildfly-rc-pod
Type:			LoadBalancer
IP:			100.65.225.120
LoadBalancer Ingress:	afa8056b14fc911e79b1906d8a9d4b8c-1413998286.us-west-2.elb.amazonaws.com
Port:			web	8080/TCP
NodePort:		web	30229/TCP
Endpoints:		100.96.1.14:8080,100.96.2.13:8080
Session Affinity:	None
Events:
  FirstSeen	LastSeen	Count	From			SubObjectPath	Type		Reason			Message
  ---------	--------	-----	----			-------------	--------	------			-------
  11s		11s		1	service-controller			Normal		CreatingLoadBalancer	Creating load balancer
  8s		8s		1	service-controller			Normal		CreatedLoadBalancer	Created load balancer

Access: Service is accessible at <LoadBalancer-Ingress>:8080. Wait for 3 mins for the load balancer to settle before accessing this URL. Firewall rules are created during the service creation.
Delete: kubectl delete -f lb-loadbalancer.yaml

ExternalName: Returns a CNAME record to an external service running outside the cluster. Allows for pods within the cluster to access the service outside the cluster. Redirection happens at DNS level, with no proxying or forwarding.
1. Create
  1. Start a WildFly bitnami stack
  2. Get IP address of the provisioned host and replace the value in lb-external.yaml
  3. kubectl create -f lb-external.yaml
2. Describe: kubectl get svc:
  NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE kubernetes 100.64.0.1 <none> 443/TCP 3h wildfly-service ec2-34-210-79-105.us-west-2.compute.amazonaws.com 8080/TCP 2s
  Details: kubectl describe svc/wildfly-service
  Name: wildfly-service Namespace: default Labels: <none> Annotations: <none> Selector: <none> Type: ExternalName IP: External Name: ec2-34-210-79-105.us-west-2.compute.amazonaws.com Port: web 8080/TCP Endpoints: <none> Session Affinity: None Events: <none>
3. Access: This service is only accessible inside the cluster. kubectl expose only work for services with selectors.
4. Delete: kubectl delete -f lb-external.yaml

Using Maven (Service + Replication Controller + Client Pod)

All the code is in maven directory:

Create Couchbase service: kubectl create -f couchbase-service.yml
Check status: kubectl get -w pods
Run application: kubectl create -f bootiful-couchbase.yml
Check status: kubectl get -w pods
1. Show ContainerCreating
Show all pods: kubectl get pods --show-all
Check logs: kubectl logs -f <pod-name> to show the output Book{isbn=978-1-4919-1889-0, name=Minecraft Modding with Forge, cost=29.99}
Delete all resources: kubectl delete -f couchbase-service.yml -f bootiful-couchbase.yml

Rolling Updates

All code in rolling-update directory:

https://github.com/arun-gupta/kubernetes-java-sample/tree/master/rolling-update

Namespaces

Create a new namespace: kubectl create -f dev-namespace.yaml
Get namespaces: kubectl get namespace
Create a new deployment in the namespace: kubectl --namespace=development run couchbase --image=arungupta/couchbase
List deployments: kubectl get deployments
1. No deployments shown
List all resources in the namespace: kubectl get deployments --namespace=development
List all resources in all namespaces: kubectl get deployments --all-namespaces
Show pods in the namespaces: kubectl get pods --all-namespaces

Quota (broken)

Create a constrained resource: kubectl create -f quota-wildfly.yaml
Check for pods: kubectl get -w pods
Broken: kubernetes/kubernetes#33621

Run-once/Batch Jobs

Create a job: kubectl create -f runonce-job.yaml
Check jobs: kubectl get jobs
More details about job: kubectl describe jobs wait
Check pods: kubectl get pods
Show all completed pods: kubectl get pods --show-all

Couchbase Cluster

https://github.com/arun-gupta/couchbase-kubernetes/tree/master/cluster

Daemon Set (work in progress)

Create a daemon set: kubectl create -f prometheus-dameonset.yml .

Tips

Create resources in all .json, .yaml and .yml files in dir: kubectl create -f ./dir

arun-gupta/kubernetes-java-sample

arun-gupta

Reviews

Repository Details