4.深入理解kubernetes：pod控制器

1.自主pod资源

一级字段:

apiVersion(group/version)

kind

metadata

spec

status

 

2.pod资源

spec.containers

- name

image

imagePullPolicy Always(无论本地有没有都去下载,最常用),Never(有就用,没有就不下载),IfNotPresent(本地不存在就下载)

如果镜像名称标签是latest,叠加策略是IfNotPresent的话等于always

ports: []object,默认TCP协议

- name: http(可以在service中引用该名称)

containerPort: 8080

- name: https

containerPort: 443

command []string (要运行的程序,[]string可以使用[]的 形式)

args []string (传递给command的参数,变量替换使用$(VAR_NAME))

 

 

3.docker中的cmd和entrypoint 与 k8s中的command和args的关系

• If you do not supply command or args for a Container, the defaults defined in the Docker image are used.
• If you supply a command but no args for a Container, only the supplied command is used. The default EntryPoint and the default Cmd defined in the Docker image are ignored.
• If you supply only args for a Container, the default Entrypoint defined in the Docker image is run with the args that you supplied.
• If you supply a command and args, the default Entrypoint and the default Cmd defined in the Docker image are ignored. Your command is run with your args.

 

 

 

4.标签的艺术

功能标签:前端,后端,数据库.缓存

版本标签:稳定版,开发版,bate版,alpha版.

环境标签:env=qa,prd

标签选择器:

(1)等值

kubectl get pods -l release=stable(有这个标签且值为stable的)

kubectl get pods -l release(有这个标签的)

kubectl get pods -l release=stable,app!=myapp(这两个标签kv都要匹配)

(2)集合关系

key in (v1,v2)

key not in (v1,v2)

!key

key

kubectl get pods -l "release in (v1,v2)"

 

(3)标签的关联

matchLabels:直接给定key

matchExpressions:基于给定的表达式来定义使用标签选择器.{key:"KEY",operator:"OPERATER",values:[v1,v2]}

操作符:in notin,exists,NotExists

 

(4)操作

key:字母,数字下划线,连接线,点

value:可以为空

-L:显示具有app标签的标签值

kubectl get pods -L app

-l:拥有app 标签的:

kubectl get pods -l app --show-labels

修改资源标签

kubectl label pods pod-demo release=canary --overwrite(如果已经存在release标签,需要覆盖)

 

kubectl get nodes --show-labels

节点选择器:

nodeSelector

nodename: 只能运行在指定node上.

annotations:与label不同的地方在于不能挑选资源对象,仅用于为对象提供元数据.

 

apiVersion: v1
kind: Pod
matedata:
    name: myapp
    namespace: default
    labels:(kv随便写)
        app: myapp
        tier: frontend   
spec:
    containers:([]object型的对象写成多个-)
    - name:myapp 
      image: ikube/app:v1
    - name: busybox
      image: busy:latest
      command: 
      - "bin/sh"
      - "-c"
      - "echo ${date} >> /usr/share/nginx/html; sleep 5"
     nodeSelector: (与containers统一目录级别)
       diskType: ssd 
  

5.pod的生命周期

状态:

pending:挂起,条件不满足,比如没有node能满足条件.只能挂起.

Running

Failed

Successed

Unknown

创建pod:

(1)初始化容器

(2)容器探测:

liveness

readiness

6.重启策略restartPolicyy:

• always;podi里的contaioner挂了就重启.重启策略是不断延时的.容器重启会一直在同一个node上重启
• OnFailure:
• Never
• Default 是always

 

---------------------------------------------------------------------------------------------------------------------------------------

二.pod探针

1.探针：

(1)livenessProbe

exec探测:

 

apiVersion: v1
kind: pod
metadata:
    name: liveness-exec-container
    namespace: default
spec:
    containers: 
    - name: liveness-exec-container
      image: busybox:latest
      imagePullPolicy: IfNotPresent
      command: ["/bin/sh","-c","touch /tmp/healthy;sleep 30;rm -rf /tmp/healthy ;sleep 3600"]
      livenessProbe:
          exec:
              command: ["test","-e","/tmp/healthy"]
          initialDelaySeconds: 1
          periodSeconds: 3    

http探测:

 

apiVersion: v1
kind: pod
metadata:
    name: liveness-http-container
    namespace: default
spec:
    containers: 
    - name: liveness-http-container
      image: xxx/myapp:v1
      imagePullPolicy: IfNotPresent
      ports:
      - name: http
        containerPort: 80    
      livenessProbe:
          httpGet:
              path: /index.html
              port: http
          initialDelaySeconds: 1
          periodSeconds: 3    

(2)readinessProbe:就绪性探测

apiVersion: v1
kind: pod
metadata:
    name: readyness-http-container
    namespace: default
spec:
    containers: 
    - name: readyness-http-container
      image: xxx/myapp:v1
      imagePullPolicy: IfNotPresent
      ports:
      - name: http
        containerPort: 80    
      readinessProbe:
          httpGet:
              path: /index.html
              port: http
          initialDelaySeconds: 1
          periodSeconds: 3  

(3)lifeCycle

postStart:启动后执行的动作:注意命令的相互关系,container的command不能与poststart的exec的command有关系.

apiVersion: v1
kind: Pod
metadata:
    name: postStart-pod
spec:
    containers:
    - name: busybox-httpd
      image: busybox:latest
      imagePullPolicy: IfNotPresent
      lifecycle:
          postStart:
              exec:
                  command: ["/bin/sh","-C","mkdir -p /data/web/html;echo aaa >> /data/web/html/index.html"]
      command: ["/bin/httpd"]
      args: ["-f","-h /data/web/html"]  

preStop:终止之前的动作

1.spec：
 	container
nodeSelector
nodeName
restartPolicy:
Always,Never,OnFailure
containers:
name
image
imagePullPolicy:Always.Never,IfNotPresent
ports:
name
containerPort
livenessProbe
readlinessProbe
lifecycle

2.pod控制器

自主式Pod:不受控制器管理的pod。

 

3.ReplicaSet-替代RC

核心资源有：

（1）副本数

（2）标签选择器

（3）pod模板

命令：kubectl explain rs

KIND:     ReplicaSet
VERSION:  extensions/v1beta1
DESCRIPTION:
     DEPRECATED - This group version of ReplicaSet is deprecated by
     apps/v1beta2/ReplicaSet. See the release notes for more information.
     ReplicaSet ensures that a specified number of pod replicas are running at
     any given time.

命令：kubectl explain rs.spec

三个主要资源：

  replicas	
     Replicas is the number of desired replicas. This is a pointer to
     distinguish between explicit zero and unspecified. Defaults to 1. More
     info:
     https://kubernetes.io/docs/concepts/workloads/controllers/replicationcontroller/#what-is-a-replicationcontroller

   selector