K8s 內的網路檢測

在 K8s 內起一個有 curl 的 busybox POD

kubectl run --rm -it busybox --image yauritux/busybox-curl:latest --restart=Never

直接下 ping / nslookup / curl 等等命令進行網路檢測看結果

rkeuser@iso:~$ kubectl run --rm -it busybox --image yauritux/busybox-curl:latest --restart=Never
If you don't see a command prompt, try pressing enter.
/home # ping www.ichiayi.com
PING www.ichiayi.com (220.135.35.198): 56 data bytes
64 bytes from 220.135.35.198: seq=0 ttl=242 time=4.992 ms
64 bytes from 220.135.35.198: seq=1 ttl=242 time=6.761 ms
^C
--- www.ichiayi.com ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max = 4.992/5.876/6.761 ms
/home # nslookup www.ichiayi.com
Server:    10.43.0.10
Address 1: 10.43.0.10 kube-dns.kube-system.svc.cluster.local

Name:      www.ichiayi.com
Address 1: 220.135.35.198 220-135-35-198.HINET-IP.hinet.net
/home # curl http://10.20.0.77:32080
<html><body>You are being <a href="http://10.20.0.77:32080/users/sign_in">redirected</a>.</body></html>/home #
/home #

輸入 exit 離開看結果
```
/home # exit
pod "busybox" deleted
```

因為在 Hyper-V 內建立一個 Ubuntu 20.04 VM 然後起一個 K8s Cluster 卻發現在 K8s 內的服務連不上 VM 內的服務, 以下是將檢測 K8s 網路的相關語法紀錄下來.

在 K8s 內起一個檢測的 pod

定義 pod 內容 k8s-chk.yml

apiVersion: apps/v1
kind: Deployment
metadata:
    name: busybox
    namespace: default
spec:
  replicas: 1
  selector:
    matchLabels:
      app: busybox
  template:
    metadata:
      labels:
        app: busybox
    spec:
      containers:
      - name: busybox
        image: yauritux/busybox-curl:latest
        command:
          - sleep
          - "3600"
        imagePullPolicy: IfNotPresent

啟動檢測 pod

kubectl apply -f k8s-chk.yml

localadmin@iiidevops1:~/test$ kubectl apply -f k8s-chk.yml
deployment.apps/busybox created

取得 pod name

kubectl get pod | grep busybox

localadmin@iiidevops1:~/test$ kubectl get pod | grep busybox
busybox-5d5bd64f66-zlvls                1/1     Running   0          2m29s

使用 pod 安裝相關工具與檢測

安裝 curl

kubectl exec busybox-5d5bd64f66-zlvls -- opkg-install curl

進行目標網址檢測

kubectl exec busybox-5d5bd64f66-zlvls -- ping -c 5 172.16.0.171
kubectl exec busybox-5d5bd64f66-zlvls -- traceroute 172.16.0.171
kubectl exec busybox-5d5bd64f66-zlvls -- curl -k https://172.16.0.171:5443/

關閉檢測的 pod

kubectl delete deploy busybox

localadmin@devops1:~/test$ kubectl delete deploy busybox
deployment.apps "busybox" deleted

檢測跨 namespace 的 service

預設完整服務網址是 {SERVICE_NAME}.{NAMESPACE_NAME}.svc.cluster.local
Exp. 一個簡易網頁服務
- SERVICE_NAME : test-20230110-37-master-serv-svc
- NAMESPACE_NAME : test-20230110-37
- test-20230110-37-master-serv-svc.test-20230110-37.svc.cluster.local

Exp. 在 namespace : test-20230110-37 內的 POD 內可直接使用 service name : test-20230110-37-master-serv-svc 當網址看結果

/var/www/html # curl http://test-20230110-37-master-serv-svc

<!doctype html>

<html lang="en">
<head>
  <meta charset="utf-8">
  <title>III DevOps Sample</title>
  <style>
      h3{text-align: center;}
  </style>
</head>
<body>
  <h3>Hello World!</h3>
</body>

Exp. 在 namespace : test-20230110-37 外的 POD 內需要使用完整的服務網址 : test-20230110-37-master-serv-svc.test-20230110-37.svc.cluster.local 看結果

/var/www/html # curl http://test-20230110-37-master-serv-svc.test-20230110-37.svc.cluster.local

<!doctype html>

<html lang="en">
<head>
  <meta charset="utf-8">
  <title>III DevOps Sample</title>
  <style>
      h3{text-align: center;}
  </style>
</head>
<body>
  <h3>Hello World!</h3>
</body>

參考網址

https://stackoverflow.com/questions/62847331/is-there-possible-to-install-curl-into-busybox-in-kubernetes-pod

K8s, busybox