pod에 Resource 할당하기 CPU/Memory requests Limit

pod에 Resource 할당하기 CPU/Memory requests, Limit

pod resource 요청 제한 

resource requests 
 - 파드를 실행하기 위한 최소 리소스 양을 요청
resource Limits 
 - 파드가 사용할 수 있는 최대 리소스 양을 제한
 - memort limit 을 초과해서 사용되는 파드는 종료(OOM Kill)되며 다시 스케줄링 된다.
https://kubernetes.io/docs/tasks/configure-pod-container/assing-cpu-resource/

# vi pod-nginx-resources.yaml

apiVersion: v1 kind: Pod metadata:   name: nginx-pod-resource spec:   containers:   - image: nginx:1.14     name: nginx-container     ports:     - containerPort: 80       protocol: TCP     resources:       requests:        <<<<         cpu:200m                memory: 250Mi         ephemeral-storage: 5Gi          limits:          <<<<         cpu: 1                 memory: 500Mi         ephemeral-storage: 5Gi

kubectl create -f pod-nginx-resources.yaml

 

# kubectl describe pod nginx-pod-resource

...
    Limits:
      cpu:     1
      memory:  1Gi
    Requests:
      cpu:        200m
      memory:     500Mi
...


requests, limits 만 각각 걸어도 상관은 없으나 용도에 따라 달라진다. 

limits만 걸면 requests 자동으로 걸립니다. 

    Limits:
      cpu:     1
      memory:  500Mi
    Requests:
      cpu:        1
      memory:     500Mi
    Environment:  <none>
    Mounts:


worker node core 전체가 2core 라면 
requests 에 2core 를 준다면 ?  실행이 안되고 Pending   
limits 도 크게 주면 마찬가지 .

# kubectl get pods  -o wide
NAME                 READY   STATUS    RESTARTS   AGE   IP       NODE     NOMINATED NODE   READINESS GATES
nginx-pod-resource   0/1     Pending   0          32s   <none>   <none>   <none>           <none>

따라서 woker node 의 하드웨어적인 부분을 확인 후 할당