[CKA] - CPU 사용량이 높은 파드 검색

[선수 지식]

pod, node -> cpu, mem

kubectl top nodes

kubectl top pods

 

---

 

[선수 작업]

kubectl config set-context k8s --user=kubernetes-admin --cluster=cluster.local

mkdir -p /var/CKA2023

kubectl apply -f - <<EOF
kind: Pod
apiVersion: v1
metadata:
  name: campus-01
  labels:
    name: overloaded-cpu
spec:
  containers:
  - name: campus-container
    image: busybox
    command:
      - "/bin/sh"
      - "-c"
      - >
        a=1;
        while true;
        do
          a=$(expr a + 1);
        done;
    resources:
      requests: 
        memory: "64Mi"
        cpu: "250m"
      limits:
        memory: "128Mi"
        cpu: "1"
EOF

kubectl apply -f - <<EOF
kind: Pod
apiVersion: v1
metadata:
  name: fast-01
  labels:
    name: overloaded-cpu
spec:
  containers:
  - name: fast-container
    image: busybox
    command:
      - "/bin/sh"
      - "-c"
      - >
        a=1;
        while true;
        do
          a=$(expr a + 1);
          sleep 0.001
        done;
    resources:
      requests: 
        memory: "64Mi"
        cpu: "250m"
      limits:
        memory: "128Mi"
        cpu: "1"
EOF

echo "[  OK  ] Please wait a minute."
sleep 20

 

---

 

[문제]

From the pod label name=overloaded-cpu, find pods running high CPU workloads and write the name of the pod consuming most CPU to the file /var/CKA2023/cpu_load_pod.txt.

 

작업 클러스터: hk8s

 

---

 

[풀이]

kubectl config use-context hk8s

* hk8s 클러스터 사용

 

kubectl top pods -l name=overloaded-cpu --sort-by=cpu

* top 명령어로 cpu 사용량이 높은 순으로 Pod 정렬

 

echo "POD_NAME" > /var/CKA2023/cpu_load_pod.txt

* 파일로 해당 Pod 이름 저장

 

---

 

[참고]

Kubectl Reference Docs