Kubernetes 管理应用 | ReplicaSet | Autoscaling | Rolling Update

Manage Application

ReplicaSet

ReplicaSet provides high availability through redundancy, a ReplicaSet enables scaling by creating or deleting pods, you can create a ReplicaSet using the CLI or the YAML descriptor, a ReplicaSet always tries to match the actual state to the desired state.

A best practice is to use a Deployment instead of a ReplicaSet directly, because a ReplicaSet is automatically created when you create a deployment.

kubectl scale deploy hello-kubernetes --replicas=3

kubectl delete/create pod hello-kubernetes-xxxx-xxx

Autoscaling

Cluster/node level or Pod level

Autoscaler types:

• Horizontal Pod Autoscaler(HPA): 通过根据负载自动调整 Pod 的数量来扩容
• Vertical Pod Autoscaler(VPA): 通过调整 Pod 的 CPU 和内存资源请求来进行扩容或缩容
• Cluster Autoscaler(CA): 集群扩容是通过自动添加或移除节点来扩展集群资源的容量

kubectl autoscale deploy hello-kubernetes --min=2 --max=5 --cpu-percent=50

Rolling Update

• All-at-once rollout/rollback: 通过 Recreate 策略实现的,所有旧的 Pod 会一次性被替换.
• One-at-a-time rollout/rollback, 通过 RollingUpdate 策略实现的,逐个替换 Pod 保证高可用性.

# update image
kubectl set image deployments/hello-kubernetes hello-kubernetes=upkar/hello-kubernetes:2.0

# deployments "hello-kubernetes" successfully rolled out
kubectl rollout status deployments/hello-kubernetes

# rollback
kubectl rollout undo deployments/hello-kubernetes

ConfigMaps and Secrets

1. A ConfigMap is an API object that stores non-confidential data in key-value pairs

   # config by command line
   kubectl create ConfigMap my-config --from-literal=MESSAGE="hello world from first configmap"
   
   # in .yaml file
   env:
   -name: MESSAGE
     valueFrom:
       configMapKeyRef:
         name: my-config
         key: MESSAGE
   
   #-------------------------------------------------------------------#
   # config from properties file instead of listing those variables one by one on the command line
   kubectl create ConfigMap my-config --from-file=my.properties
   
   # in my.properties
   MESSAGE=hello world from my.properties file
   
   # in .yaml file
   env:
   -name: MESSAGE
     valueFrom:
       configMapKeyRef:
         name: my-config
         key: my.properties
   
   #-------------------------------------------------------------------#
   # config from yaml file
   # my-config.YAML
   apiVersion: v1
   data:
     my.properties: MESSAGE=hello world from my.properties file
   kind:
     ConfigMap
   metadata:
     name: my-config
     name-space: default
   
   # kubectl apply
   kubectl apply -f my-config.YAML
   kubectl describe cm my-config
   

2. Secret is used to provide sensitive information to your application

   kubectl create secret generic api-creds --from-literal=key=mysupersecretapikey
   
   kubectl get secret
   
   kubectl describe secret api-creds
   
   # output in YAML format
   kubectl get secret api-creds -o YAML
   
   # use
   # in .yaml file
   env:
   -name: API_CREDS
     valueFrom:
       secretKeyRef:
         name: api-creds
         key: key
   
   #-------------------------------------------------------------------#
   # config by using volume mounts
   # secrets by mount as file: /etc/api/api-creds
   spec:
     containers:
     - name: hello-kubernetes
       image: upkar/hello-kubernetes
       ports:
       - containerPort: 8080
       volumeMounts:
       - name: api-creds
         mountPath: "/etc/api"
         readOnly: true
       volumes:
       - name: api-creds
         secret:
           secretName: api-creds
   

Service Binding

Service binding is the process needed to consume external Services or backing Services, including REST APIs, databases, and event buses in our applications