Pod¶

https://kubernetes.io/docs/concepts/workloads/pods/

What is Pod?¶

Pod是k8s里最小的调度单位。

A group of one or more application containers and their shared resources like volume.
A pod share the same name spaces like network name spaces (have same IP address.)
Pod is the smallest unit for K8s scheduling

Note

关于容器，name spaces，欢迎参考另一篇关于Docker的文档 https://dockertips.readthedocs.io/en/latest/single-host-network.html

How to create a pod?¶

Imperative¶

Create a pod named web with image of nginx:latest

$ kubectl run web --image=nginx
pod/web created

运行一个命令, kubectl run <pod name> –image=<image name> –command – <cmd> <arg1> … <argN>

$ kubectl run client --image=busybox --command -- bin/sh -c "sleep 100000"

Declarative¶

以下yaml文件是定义一个pod所需的最少字段 (nginx.yml)

apiVersion: v1
kind: Pod
metadata:
  name: web
spec:
   containers:
    - name: nginx-container
      image: nginx:latest

$ kubectl apply -f nginx.yml
pod/web created

运行一个命令, sh -c “sleep 1000000”

apiVersion: v1
kind: Pod
metadata:
   name: client
spec:
   containers:
    - name: client
      image: busybox
      command:
       - sh
       - -c
       - "sleep 1000000"

multi-container pod¶

一个pod是可以包含多个container的，如果要创建这样的pod，那么只能通过yaml文件实现，例如：

apiVersion: v1
kind: Pod
metadata:
   name: my-pod
spec:
   containers:
    - name: nginx
      image: nginx
    - name: client
      image: busybox
      command:
       - sh
       - -c
       - "sleep 1000000"

$ kubectl create -f my-pod.yml
$ kubectl get pod
NAME     READY   STATUS    RESTARTS   AGE
my-pod   2/2     Running   0          35s

Pod YAML 语法¶

查文档 kubernetes.io, 命令行帮助

kubectl explain pods | more
kubectl explain pod.spec | more
kubectl explain pod.spec.containers | more

kubectl dry-run¶

Server-side¶

和正常情况一样处理客户端发送过来的请求，但是并不会把Object状态持久化存储到storage中

$ kubectl apply -f nginx.yml --dry-run=server

Client-side¶

把要操作的Object通过标准输出stdout输出到terminal
验证manifest的语法
可以用于生成语法正确的Yaml manifest

$ kubectl apply -f nginx.yml --dry-run=client
$ kubectl run web --image=nginx --dry-run=client -o yaml
$ kubectl run web --image=nginx --dry-run=client -o yaml > nginx.yml

kubectl diff¶

显示当前要部署的manifest和集群中运行的有和不同，这样就知道如果apply会发生什么。

$ kubectl diff -f new-nginx.yml

Pod的基本操作¶

获取pod列表¶

vagrant@k8s-master:~$ kubectl get pods
NAME     READY   STATUS    RESTARTS   AGE
client   1/1     Running   0          5m14s
web      1/1     Running   0          15m
vagrant@k8s-master:~$ kubectl get pods -o wide
NAME     READY   STATUS    RESTARTS   AGE     IP           NODE          NOMINATED NODE   READINESS GATES
client   1/1     Running   0          5m17s   10.244.2.4   k8s-worker2   <none>           <none>
web      1/1     Running   0          15m     10.244.1.2   k8s-worker1   <none>           <none>

通过 -o yaml 可以获取到具体一个pod的yaml定义文件

$ kubectl get pods client -o yaml

删除Pod¶

$ kubectl delete pod web
pod "web" deleted

获取pod详细信息¶

$ kubectl describe pod my-pod

进入容器执行命令¶

对于只有单个容器的Pod，执行date命令

vagrant@k8s-master:~$ kubectl get pods
NAME     READY   STATUS    RESTARTS   AGE
client   1/1     Running   0          38s
my-pod   2/2     Running   0          6s
vagrant@k8s-master:~$ kubectl exec client -- date
Wed Jun  1 21:57:07 UTC 2022

进入交互式shell

vagrant@k8s-master:~$ kubectl exec client -- date
Wed Jun  1 21:57:07 UTC 2022
vagrant@k8s-master:~$ kubectl exec client -it -- sh
/ #
/ # ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue qlen 1000
   link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
   inet 127.0.0.1/8 scope host lo
      valid_lft forever preferred_lft forever
   inet6 ::1/128 scope host
      valid_lft forever preferred_lft forever
3: eth0@if7: <BROADCAST,MULTICAST,UP,LOWER_UP,M-DOWN> mtu 1450 qdisc noqueue
   link/ether a6:56:08:ba:34:28 brd ff:ff:ff:ff:ff:ff
   inet 10.244.1.3/24 brd 10.244.1.255 scope global eth0
      valid_lft forever preferred_lft forever
   inet6 fe80::a456:8ff:feba:3428/64 scope link
      valid_lft forever preferred_lft forever
/ #

对于具有多个容器的pod，需要通过 -c 指定要进入那个容器中。

vagrant@k8s-master:~$ kubectl get pods
NAME     READY   STATUS    RESTARTS   AGE
client   1/1     Running   0          3m16s
my-pod   2/2     Running   0          2m44s
vagrant@k8s-master:~$ kubectl exec my-pod -c
client  nginx
vagrant@k8s-master:~$ kubectl exec my-pod -c nginx -- date
Wed Jun  1 21:59:58 UTC 2022

API level log¶

通过 -v 可以获取到每一个kubectl命令在API level的log，例如

获取kubectl操作更详细的log, API level（通过 -v 指定）

$ kubectl get pod <pod-name> -v 6   # 或者 7,8,9 不同的level，数值越大，得到的信息越详细

--watch 持续监听kubectl操作，API level

$ kubectl get pods <pod-name> --watch -v 6

Pod with init containers¶

apiVersion: v1
kind: Pod
metadata:
  name: pod-with-init-containers
spec:
  initContainers:
  - name: init-service
    image: busybox
    command: ["sh", "-c", "echo waiting for sercice; sleep 2"]
  - name: init-database
    image: busybox
    command: ["sh", "-c", "echo waiting for database; sleep 2"]
  containers:
  - name: app-container
    image: nginx

Pod Lifecycle¶

Pod Health¶

LivenessProbes¶

Runs a diagnostic check on a container
Per container setting
On failure, the kubelet restarts the container
Container Restart Policy
Give Kubernetes a better understanding of our applications

readinessProbes¶

Runs a diagnostic check on a container
Per container setting
Won’t receive traffic from a service until it succeeds
On failure, remove Pod from load balancing
Applications that temporarily can’t respond to a request
Prevents users from seeing errors

type of Diagnostic Checks for Probes

Exec
tcpSocket
httpGet