每日热闻!k8s资源对象
博客园
2023-05-11 06:42:11
什么是资源对象？
【资料图】
所谓资源对象是指在k8s上创建的资源实例；即通过apiserver提供的各资源api接口（可以理解为各种资源模板），使用yaml文件或者命令行的方式向对应资源api接口传递参数赋值实例化的结果；比如我们在k8s上创建一个pod，那么我们就需要通过给apiserver交互，传递创建pod的相关参数，让apiserver拿着这些参数去实例化一个pod的相关信息存放在etcd中，然后再由调度器进行调度，由node节点的kubelet执行创建pod；简单讲资源对象就是把k8s之上的api接口进行实例化的结果；
k8s逻辑运行环境
提示：k8s运行环境如上，k8s能够将多个node节点的底层提供的资源（如内存，cpu，存储，网络等）逻辑的整合成一个大的资源池，统一由k8s进行调度编排；用户只管在k8s上创建各种资源即可，创建完成的资源是通过k8s统一编排调度，用户无需关注具体资源在那个node上运行，也无需关注node节点资源情况；
k8s的设计理念——分层架构
k8s的设计理念——API设计原则
1、所有API应该是声明式的；
2、API对象是彼此互补而且可组合的，即“高内聚，松耦合”；
3、高层API以操作意图为基础设计；
4、低层API根据高层API的控制需要设计；
5、尽量避免简单封装，不要有在外部API无法显式知道的内部隐藏的机制；
6、API操作复杂度与对象数量成正比；
7、API对象状态不能依赖于网络连接状态；
8、尽量避免让操作机制依赖于全局状态，因为在分布式系统中要保证全局状态的同步是非常困难的；
kubernetes API简介
提示：在k8s上api分内置api和自定义api；所谓内置api是指部署好k8s集群后自带的api接口；自定义api也称自定义资源（CRD，Custom Resource Definition），部署好k8s之后，通过安装其他组件等方式扩展出来的api；
apiserver资源组织逻辑
提示：apiserver对于不同资源是通过分类，分组，分版本的方式逻辑组织的，如上图所示；
k8s内置资源对象简介
k8s资源对象操作命令
资源配置清单必需字段
1、apiVersion - 创建该对象所使用的Kubernetes API的版本；
2、kind - 想要创建的对象的类型；
3、metadata - 定义识别对象唯一性的数据，包括一个name名称 、可选的namespace，默认不写就是default名称空间；
4、spec：定义资源对象的详细规范信息(统一的label标签、容器名称、镜像、端口映射等)，即用户期望对应资源处于什么状态；
5、status（Pod创建完成后k8s自动生成status状态），该字段信息由k8s自动维护，用户无需定义，即对应资源的实际状态；
Pod资源对象
提示：pod是k8s中最小控制单元，一个pod中可以运行一个或多个容器；一个pod的中的容器是一起调度，即调度的最小单位是pod；pod的生命周期是短暂的，不会自愈，是用完就销毁的实体；一般我们通过Controller来创建和管理pod；使用控制器创建的pod具有自动恢复功能，即pod状态不满足用户期望状态，对应控制器会通过重启或重建的方式，让对应pod状态和数量始终和用户定义的期望状态一致；
示例：自主式pod配置清单
apiVersion: v1kind: Podmetadata: name: "pod-demo" namespace: default labels: app: "pod-demo"spec: containers: - name: pod-demo image: "harbor.ik8s.cc/baseimages/nginx:v1" ports: - containerPort: 80 name: http volumeMounts: - name: localtime mountPath: /etc/localtime volumes: - name: localtime hostPath: path: /usr/share/zoneinfo/Asia/Shanghai
应用配置清单
root@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEnet-test1 1/1 Running 2 (4m35s ago) 7d7htest 1/1 Running 4 (4m34s ago) 13dtest1 1/1 Running 4 (4m35s ago) 13dtest2 1/1 Running 4 (4m35s ago) 13droot@k8s-deploy:/yaml# kubectl apply -f pod-demo.yamlpod/pod-demo createdroot@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEnet-test1 1/1 Running 2 (4m47s ago) 7d7hpod-demo 0/1 ContainerCreating 0 4stest 1/1 Running 4 (4m46s ago) 13dtest1 1/1 Running 4 (4m47s ago) 13dtest2 1/1 Running 4 (4m47s ago) 13droot@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEnet-test1 1/1 Running 2 (4m57s ago) 7d7hpod-demo 1/1 Running 0 14stest 1/1 Running 4 (4m56s ago) 13dtest1 1/1 Running 4 (4m57s ago) 13dtest2 1/1 Running 4 (4m57s ago) 13droot@k8s-deploy:/yaml#
提示：此pod只是在k8s上运行起来，它没有控制器的监视，对应pod删除，故障都不会自动恢复；
Job控制器，详细说明请参考https://www.cnblogs.com/qiuhom-1874/p/14157306.html；
job控制器配置清单示例
apiVersion: batch/v1kind: Jobmetadata: name: job-demo namespace: default labels: app: job-demospec: template: metadata: name: job-demo labels: app: job-demo spec: containers: - name: job-demo-container image: harbor.ik8s.cc/baseimages/centos7:2023 command: ["/bin/sh"] args: ["-c", "echo data init job at `date +%Y-%m-%d_%H-%M-%S` >> /cache/data.log"] volumeMounts: - mountPath: /cache name: cache-volume - name: localtime mountPath: /etc/localtime volumes: - name: cache-volume hostPath: path: /tmp/jobdata - name: localtime hostPath: path: /usr/share/zoneinfo/Asia/Shanghai restartPolicy: Never
提示：定义job资源必须定义restartPolicy；
应用清单
root@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEnet-test1 1/1 Running 3 (48m ago) 7d10hpod-demo 1/1 Running 1 (48m ago) 3h32mtest 1/1 Running 5 (48m ago) 14dtest1 1/1 Running 5 (48m ago) 14dtest2 1/1 Running 5 (48m ago) 14droot@k8s-deploy:/yaml# kubectl apply -f job-demo.yaml job.batch/job-demo createdroot@k8s-deploy:/yaml# kubectl get pods -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATESjob-demo-z8gmb 0/1 Completed 0 26s 10.200.211.130 192.168.0.34net-test1 1/1 Running 3 (49m ago) 7d10h 10.200.211.191 192.168.0.34 pod-demo 1/1 Running 1 (49m ago) 3h32m 10.200.155.138 192.168.0.36 test 1/1 Running 5 (49m ago) 14d 10.200.209.6 192.168.0.35 test1 1/1 Running 5 (49m ago) 14d 10.200.209.8 192.168.0.35 test2 1/1 Running 5 (49m ago) 14d 10.200.211.177 192.168.0.34 root@k8s-deploy:/yaml#
验证：查看192.168.0.34的/tmp/jobdata目录下是否有job执行的任务数据？
root@k8s-deploy:/yaml# ssh 192.168.0.34 "ls /tmp/jobdata"data.logroot@k8s-deploy:/yaml# ssh 192.168.0.34 "cat /tmp/jobdata/data.log"data init job at 2023-05-06_23-31-32root@k8s-deploy:/yaml#
提示：可以看到对应job所在宿主机的/tmp/jobdata/目录下有job执行过后的数据，这说明我们定义的job任务顺利完成；
定义并行job
apiVersion: batch/v1kind: Jobmetadata: name: job-multi-demo namespace: default labels: app: job-multi-demospec: completions: 5 template: metadata: name: job-multi-demo labels: app: job-multi-demo spec: containers: - name: job-multi-demo-container image: harbor.ik8s.cc/baseimages/centos7:2023 command: ["/bin/sh"] args: ["-c", "echo data init job at `date +%Y-%m-%d_%H-%M-%S` >> /cache/data.log"] volumeMounts: - mountPath: /cache name: cache-volume - name: localtime mountPath: /etc/localtime volumes: - name: cache-volume hostPath: path: /tmp/jobdata - name: localtime hostPath: path: /usr/share/zoneinfo/Asia/Shanghai restartPolicy: Never
提示：spec字段下使用completions来指定执行任务需要的对应pod的数量；
应用清单
root@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEjob-demo-z8gmb 0/1 Completed 0 24mnet-test1 1/1 Running 3 (73m ago) 7d11hpod-demo 1/1 Running 1 (73m ago) 3h56mtest 1/1 Running 5 (73m ago) 14dtest1 1/1 Running 5 (73m ago) 14dtest2 1/1 Running 5 (73m ago) 14droot@k8s-deploy:/yaml# kubectl apply -f job-multi-demo.yaml job.batch/job-multi-demo createdroot@k8s-deploy:/yaml# kubectl get jobNAME COMPLETIONS DURATION AGEjob-demo 1/1 5s 24mjob-multi-demo 1/5 10s 10sroot@k8s-deploy:/yaml# kubectl get pods -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATESjob-demo-z8gmb 0/1 Completed 0 24m 10.200.211.130 192.168.0.34job-multi-demo-5vp9w 0/1 Completed 0 12s 10.200.211.144 192.168.0.34 job-multi-demo-frstg 0/1 Completed 0 22s 10.200.211.186 192.168.0.34 job-multi-demo-gd44s 0/1 Completed 0 17s 10.200.211.184 192.168.0.34 job-multi-demo-kfm79 0/1 ContainerCreating 0 2s 192.168.0.34 job-multi-demo-nsmpg 0/1 Completed 0 7s 10.200.211.135 192.168.0.34 net-test1 1/1 Running 3 (73m ago) 7d11h 10.200.211.191 192.168.0.34 pod-demo 1/1 Running 1 (73m ago) 3h56m 10.200.155.138 192.168.0.36 test 1/1 Running 5 (73m ago) 14d 10.200.209.6 192.168.0.35 test1 1/1 Running 5 (73m ago) 14d 10.200.209.8 192.168.0.35 test2 1/1 Running 5 (73m ago) 14d 10.200.211.177 192.168.0.34 root@k8s-deploy:/yaml# kubectl get pods -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATESjob-demo-z8gmb 0/1 Completed 0 24m 10.200.211.130 192.168.0.34 job-multi-demo-5vp9w 0/1 Completed 0 33s 10.200.211.144 192.168.0.34 job-multi-demo-frstg 0/1 Completed 0 43s 10.200.211.186 192.168.0.34 job-multi-demo-gd44s 0/1 Completed 0 38s 10.200.211.184 192.168.0.34 job-multi-demo-kfm79 0/1 Completed 0 23s 10.200.211.140 192.168.0.34 job-multi-demo-nsmpg 0/1 Completed 0 28s 10.200.211.135 192.168.0.34 net-test1 1/1 Running 3 (73m ago) 7d11h 10.200.211.191 192.168.0.34 pod-demo 1/1 Running 1 (73m ago) 3h57m 10.200.155.138 192.168.0.36 test 1/1 Running 5 (73m ago) 14d 10.200.209.6 192.168.0.35 test1 1/1 Running 5 (73m ago) 14d 10.200.209.8 192.168.0.35 test2 1/1 Running 5 (73m ago) 14d 10.200.211.177 192.168.0.34 root@k8s-deploy:/yaml#
验证：查看192.168.0.34的/tmp/jobdata/目录下是否有job数据产生？
root@k8s-deploy:/yaml# ssh 192.168.0.34 "ls /tmp/jobdata" data.logroot@k8s-deploy:/yaml# ssh 192.168.0.34 "cat /tmp/jobdata/data.log"data init job at 2023-05-06_23-31-32data init job at 2023-05-06_23-55-44data init job at 2023-05-06_23-55-49data init job at 2023-05-06_23-55-54data init job at 2023-05-06_23-55-59data init job at 2023-05-06_23-56-04root@k8s-deploy:/yaml#
定义并行度
apiVersion: batch/v1kind: Jobmetadata: name: job-multi-demo2 namespace: default labels: app: job-multi-demo2spec: completions: 6 parallelism: 2 template: metadata: name: job-multi-demo2 labels: app: job-multi-demo2 spec: containers: - name: job-multi-demo2-container image: harbor.ik8s.cc/baseimages/centos7:2023 command: ["/bin/sh"] args: ["-c", "echo data init job at `date +%Y-%m-%d_%H-%M-%S` >> /cache/data.log"] volumeMounts: - mountPath: /cache name: cache-volume - name: localtime mountPath: /etc/localtime volumes: - name: cache-volume hostPath: path: /tmp/jobdata - name: localtime hostPath: path: /usr/share/zoneinfo/Asia/Shanghai restartPolicy: Never
提示：在spec字段下使用parallelism字段来指定并行度，即一次几个pod同时运行；上述清单表示，一次2个pod同时运行，总共需要6个pod；
应用清单
root@k8s-deploy:/yaml# kubectl get jobsNAME COMPLETIONS DURATION AGEjob-demo 1/1 5s 34mjob-multi-demo 5/5 25s 9m56sroot@k8s-deploy:/yaml# kubectl apply -f job-multi-demo2.yaml job.batch/job-multi-demo2 createdroot@k8s-deploy:/yaml# kubectl get jobsNAME COMPLETIONS DURATION AGEjob-demo 1/1 5s 34mjob-multi-demo 5/5 25s 10mjob-multi-demo2 0/6 2s 3sroot@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEjob-demo-z8gmb 0/1 Completed 0 34mjob-multi-demo-5vp9w 0/1 Completed 0 10mjob-multi-demo-frstg 0/1 Completed 0 10mjob-multi-demo-gd44s 0/1 Completed 0 10mjob-multi-demo-kfm79 0/1 Completed 0 9m59sjob-multi-demo-nsmpg 0/1 Completed 0 10mjob-multi-demo2-7ppxc 0/1 Completed 0 10sjob-multi-demo2-mxbtq 0/1 Completed 0 5sjob-multi-demo2-rhgh7 0/1 Completed 0 4sjob-multi-demo2-th6ff 0/1 Completed 0 11snet-test1 1/1 Running 3 (83m ago) 7d11hpod-demo 1/1 Running 1 (83m ago) 4h6mtest 1/1 Running 5 (83m ago) 14dtest1 1/1 Running 5 (83m ago) 14dtest2 1/1 Running 5 (83m ago) 14droot@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEjob-demo-z8gmb 0/1 Completed 0 34mjob-multi-demo-5vp9w 0/1 Completed 0 10mjob-multi-demo-frstg 0/1 Completed 0 10mjob-multi-demo-gd44s 0/1 Completed 0 10mjob-multi-demo-kfm79 0/1 Completed 0 10mjob-multi-demo-nsmpg 0/1 Completed 0 10mjob-multi-demo2-7ppxc 0/1 Completed 0 16sjob-multi-demo2-8bh22 0/1 Completed 0 6sjob-multi-demo2-dbjqw 0/1 Completed 0 6sjob-multi-demo2-mxbtq 0/1 Completed 0 11sjob-multi-demo2-rhgh7 0/1 Completed 0 10sjob-multi-demo2-th6ff 0/1 Completed 0 17snet-test1 1/1 Running 3 (83m ago) 7d11hpod-demo 1/1 Running 1 (83m ago) 4h6mtest 1/1 Running 5 (83m ago) 14dtest1 1/1 Running 5 (83m ago) 14dtest2 1/1 Running 5 (83m ago) 14droot@k8s-deploy:/yaml# kubectl get pods -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATESjob-demo-z8gmb 0/1 Completed 0 35m 10.200.211.130 192.168.0.34job-multi-demo-5vp9w 0/1 Completed 0 10m 10.200.211.144 192.168.0.34 job-multi-demo-frstg 0/1 Completed 0 11m 10.200.211.186 192.168.0.34 job-multi-demo-gd44s 0/1 Completed 0 11m 10.200.211.184 192.168.0.34 job-multi-demo-kfm79 0/1 Completed 0 10m 10.200.211.140 192.168.0.34 job-multi-demo-nsmpg 0/1 Completed 0 10m 10.200.211.135 192.168.0.34 job-multi-demo2-7ppxc 0/1 Completed 0 57s 10.200.211.145 192.168.0.34 job-multi-demo2-8bh22 0/1 Completed 0 47s 10.200.211.148 192.168.0.34 job-multi-demo2-dbjqw 0/1 Completed 0 47s 10.200.211.141 192.168.0.34 job-multi-demo2-mxbtq 0/1 Completed 0 52s 10.200.211.152 192.168.0.34 job-multi-demo2-rhgh7 0/1 Completed 0 51s 10.200.211.143 192.168.0.34 job-multi-demo2-th6ff 0/1 Completed 0 58s 10.200.211.136 192.168.0.34 net-test1 1/1 Running 3 (84m ago) 7d11h 10.200.211.191 192.168.0.34 pod-demo 1/1 Running 1 (84m ago) 4h7m 10.200.155.138 192.168.0.36 test 1/1 Running 5 (84m ago) 14d 10.200.209.6 192.168.0.35 test1 1/1 Running 5 (84m ago) 14d 10.200.209.8 192.168.0.35 test2 1/1 Running 5 (84m ago) 14d 10.200.211.177 192.168.0.34 root@k8s-deploy:/yaml#
验证job数据
提示：可以看到后面job追加的时间几乎都是两个重复的，这说明两个pod同时执行了job里的任务；
Cronjob控制器，详细说明请参考https://www.cnblogs.com/qiuhom-1874/p/14157306.html；
示例：定义cronjob
apiVersion: batch/v1kind: CronJobmetadata: name: job-cronjob namespace: defaultspec: schedule: "*/1 * * * *" jobTemplate: spec: parallelism: 2 template: spec: containers: - name: job-cronjob-container image: harbor.ik8s.cc/baseimages/centos7:2023 command: ["/bin/sh"] args: ["-c", "echo data init job at `date +%Y-%m-%d_%H-%M-%S` >> /cache/cronjob-data.log"] volumeMounts: - mountPath: /cache name: cache-volume - name: localtime mountPath: /etc/localtime volumes: - name: cache-volume hostPath: path: /tmp/jobdata - name: localtime hostPath: path: /usr/share/zoneinfo/Asia/Shanghai restartPolicy: OnFailure
应用清单
root@k8s-deploy:/yaml# kubectl apply -f cronjob-demo.yamlcronjob.batch/job-cronjob createdroot@k8s-deploy:/yaml# kubectl get cronjobNAME SCHEDULE SUSPEND ACTIVE LAST SCHEDULE AGEjob-cronjob */1 * * * * False 06sroot@k8s-deploy:/yaml# kubectl get podsNAME READY STATUS RESTARTS AGEjob-cronjob-28056516-njddz 0/1 Completed 0 12sjob-cronjob-28056516-wgbns 0/1 Completed 0 12sjob-demo-z8gmb 0/1 Completed 0 64mjob-multi-demo-5vp9w 0/1 Completed 0 40mjob-multi-demo-frstg 0/1 Completed 0 40mjob-multi-demo-gd44s 0/1 Completed 0 40mjob-multi-demo-kfm79 0/1 Completed 0 40mjob-multi-demo-nsmpg 0/1 Completed 0 40mjob-multi-demo2-7ppxc 0/1 Completed 0 30mjob-multi-demo2-8bh22 0/1 Completed 0 30mjob-multi-demo2-dbjqw 0/1 Completed 0 30mjob-multi-demo2-mxbtq 0/1 Completed 0 30mjob-multi-demo2-rhgh7 0/1 Completed 0 30mjob-multi-demo2-th6ff 0/1 Completed 0 30mnet-test1 1/1 Running 3 (113m ago) 7d11hpod-demo 1/1 Running 1 (113m ago) 4h36mtest 1/1 Running 5 (113m ago) 14dtest1 1/1 Running 5 (113m ago) 14dtest2 1/1 Running 5 (113m ago) 14droot@k8s-deploy:/yaml# kubectl get cronjobNAME SCHEDULE SUSPEND ACTIVE LAST SCHEDULE AGEjob-cronjob */1 * * * * False 0 12s 108sroot@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEjob-cronjob-28056516-njddz 0/1 Completed 0 77sjob-cronjob-28056516-wgbns 0/1 Completed 0 77sjob-cronjob-28056517-d6n9h 0/1 Completed 0 17sjob-cronjob-28056517-krsvb 0/1 Completed 0 17sjob-demo-z8gmb 0/1 Completed 0 65mjob-multi-demo-5vp9w 0/1 Completed 0 41mjob-multi-demo-frstg 0/1 Completed 0 41mjob-multi-demo-gd44s 0/1 Completed 0 41mjob-multi-demo-kfm79 0/1 Completed 0 41mjob-multi-demo-nsmpg 0/1 Completed 0 41mjob-multi-demo2-7ppxc 0/1 Completed 0 31mjob-multi-demo2-8bh22 0/1 Completed 0 31mjob-multi-demo2-dbjqw 0/1 Completed 0 31mjob-multi-demo2-mxbtq 0/1 Completed 0 31mjob-multi-demo2-rhgh7 0/1 Completed 0 31mjob-multi-demo2-th6ff 0/1 Completed 0 31mnet-test1 1/1 Running 3 (114m ago) 7d11hpod-demo 1/1 Running 1 (114m ago) 4h38mtest 1/1 Running 5 (114m ago) 14dtest1 1/1 Running 5 (114m ago) 14dtest2 1/1 Running 5 (114m ago) 14droot@k8s-deploy:/yaml#
提示：cronjob 默认保留最近3个历史记录；
验证：查看周期执行任务的数据
提示：从上面的时间就可以看到每过一分钟就有两个pod执行一次任务；
RC/RS 副本控制器
RC（Replication Controller），副本控制器，该控制器主要负责控制pod副本数量始终满足用户期望的副本数量，该副本控制器是第一代pod副本控制器，仅支持selector = !=；
rc控制器示例
apiVersion: v1 kind: ReplicationController metadata: name: ng-rcspec: replicas: 2 selector: app: ng-rc-80 template: metadata: labels: app: ng-rc-80 spec: containers: - name: pod-demo image: "harbor.ik8s.cc/baseimages/nginx:v1" ports: - containerPort: 80 name: http volumeMounts: - name: localtime mountPath: /etc/localtime volumes: - name: localtime hostPath: path: /usr/share/zoneinfo/Asia/Shanghai
应用配置清单
root@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEtest 1/1 Running 6 (11m ago) 16dtest1 1/1 Running 6 (11m ago) 16dtest2 1/1 Running 6 (11m ago) 16droot@k8s-deploy:/yaml# kubectl apply -f rc-demo.yamlreplicationcontroller/ng-rc createdroot@k8s-deploy:/yaml# kubectl get pods -o wideNAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATESng-rc-l7xmp 1/1 Running 0 10s 10.200.211.136 192.168.0.34ng-rc-wl5d6 1/1 Running 0 9s 10.200.155.185 192.168.0.36 test 1/1 Running 6 (11m ago) 16d 10.200.209.24 192.168.0.35 test1 1/1 Running 6 (11m ago) 16d 10.200.209.31 192.168.0.35 test2 1/1 Running 6 (11m ago) 16d 10.200.211.186 192.168.0.34 root@k8s-deploy:/yaml# kubectl get rcNAME DESIRED CURRENT READY AGEng-rc 2 2 2 25sroot@k8s-deploy:/yaml#
验证：修改pod标签，看看对应pod是否会重新创建？
root@k8s-deploy:/yaml# kubectl get pods --show-labelsNAME READY STATUS RESTARTS AGE LABELSng-rc-l7xmp 1/1 Running 0 2m32s app=ng-rc-80ng-rc-wl5d6 1/1 Running 0 2m31s app=ng-rc-80test 1/1 Running 6 (13m ago) 16d run=testtest1 1/1 Running 6 (13m ago) 16d run=test1test2 1/1 Running 6 (13m ago) 16d run=test2root@k8s-deploy:/yaml# kubectl label pod/ng-rc-l7xmp app=nginx-demo --overwritepod/ng-rc-l7xmp labeledroot@k8s-deploy:/yaml# kubectl get pods --show-labels NAME READY STATUS RESTARTS AGE LABELSng-rc-l7xmp 1/1 Running 0 4m42s app=nginx-demong-rc-rxvd4 0/1 ContainerCreating 0 3s app=ng-rc-80ng-rc-wl5d6 1/1 Running 0 4m41s app=ng-rc-80test 1/1 Running 6 (15m ago) 16d run=testtest1 1/1 Running 6 (15m ago) 16d run=test1test2 1/1 Running 6 (15m ago) 16d run=test2root@k8s-deploy:/yaml# kubectl get pods --show-labelsNAME READY STATUS RESTARTS AGE LABELSng-rc-l7xmp 1/1 Running 0 4m52s app=nginx-demong-rc-rxvd4 1/1 Running 0 13s app=ng-rc-80ng-rc-wl5d6 1/1 Running 0 4m51s app=ng-rc-80test 1/1 Running 6 (16m ago) 16d run=testtest1 1/1 Running 6 (16m ago) 16d run=test1test2 1/1 Running 6 (16m ago) 16d run=test2root@k8s-deploy:/yaml# kubectl label pod/ng-rc-l7xmp app=ng-rc-80 --overwrite pod/ng-rc-l7xmp labeledroot@k8s-deploy:/yaml# kubectl get pods --show-labels NAME READY STATUS RESTARTS AGE LABELSng-rc-l7xmp 1/1 Running 0 5m27s app=ng-rc-80ng-rc-wl5d6 1/1 Running 0 5m26s app=ng-rc-80test 1/1 Running 6 (16m ago) 16d run=testtest1 1/1 Running 6 (16m ago) 16d run=test1test2 1/1 Running 6 (16m ago) 16d run=test2root@k8s-deploy:/yaml#
提示：rc控制器是通过标签选择器来识别对应pod是否归属对应rc控制器管控，如果发现对应pod的标签发生改变，那么rc控制器会通过新建或删除的方法将对应pod数量始终和用户定义的数量保持一致；
RS（ReplicaSet），副本控制器，该副本控制器和rc类似，都是通过标签选择器来匹配归属自己管控的pod数量，如果标签或对应pod数量少于或多余用户期望的数量，该控制器会通过新建或删除pod的方式将对应pod数量始终和用户期望的pod数量保持一致；rs控制器和rc控制器唯一区别就是rs控制器支持selector = !=精确匹配外，还支持模糊匹配in notin；是k8s之上的第二代pod副本控制器；
rs控制器示例
apiVersion: apps/v1kind: ReplicaSetmetadata: name: rs-demo labels: app: rs-demospec: replicas: 3 selector: matchLabels: app: rs-demo template: metadata: labels: app: rs-demo spec: containers: - name: rs-demo image: "harbor.ik8s.cc/baseimages/nginx:v1" ports: - name: web containerPort: 80 protocol: TCP env: - name: NGX_VERSION value: 1.16.1 volumeMounts: - name: localtime mountPath: /etc/localtime volumes: - name: localtime hostPath: path: /usr/share/zoneinfo/Asia/Shanghai
应用配置清单
验证：修改pod标签，看看对应pod是否会发生变化？
root@k8s-deploy:/yaml# kubectl get pods --show-labelsNAME READY STATUS RESTARTS AGE LABELSng-rc-l7xmp 1/1 Running 0 18m app=ng-rc-80ng-rc-wl5d6 1/1 Running 0 18m app=ng-rc-80rs-demo-nzmqs 1/1 Running 0 71s app=rs-demors-demo-v2vb6 1/1 Running 0 71s app=rs-demors-demo-x27fv 1/1 Running 0 71s app=rs-demotest 1/1 Running 6 (29m ago) 16d run=testtest1 1/1 Running 6 (29m ago) 16d run=test1test2 1/1 Running 6 (29m ago) 16d run=test2root@k8s-deploy:/yaml# kubectl label pod/rs-demo-nzmqs app=nginx --overwrite pod/rs-demo-nzmqs labeledroot@k8s-deploy:/yaml# kubectl get pods --show-labels NAME READY STATUS RESTARTS AGE LABELSng-rc-l7xmp 1/1 Running 0 19m app=ng-rc-80ng-rc-wl5d6 1/1 Running 0 19m app=ng-rc-80rs-demo-bdfdd 1/1 Running 0 4s app=rs-demors-demo-nzmqs 1/1 Running 0 103s app=nginxrs-demo-v2vb6 1/1 Running 0 103s app=rs-demors-demo-x27fv 1/1 Running 0 103s app=rs-demotest 1/1 Running 6 (30m ago) 16d run=testtest1 1/1 Running 6 (30m ago) 16d run=test1test2 1/1 Running 6 (30m ago) 16d run=test2root@k8s-deploy:/yaml# kubectl label pod/rs-demo-nzmqs app=rs-demo --overwritepod/rs-demo-nzmqs labeledroot@k8s-deploy:/yaml# kubectl get pods --show-labels NAME READY STATUS RESTARTS AGE LABELSng-rc-l7xmp 1/1 Running 0 19m app=ng-rc-80ng-rc-wl5d6 1/1 Running 0 19m app=ng-rc-80rs-demo-nzmqs 1/1 Running 0 119s app=rs-demors-demo-v2vb6 1/1 Running 0 119s app=rs-demors-demo-x27fv 1/1 Running 0 119s app=rs-demotest 1/1 Running 6 (30m ago) 16d run=testtest1 1/1 Running 6 (30m ago) 16d run=test1test2 1/1 Running 6 (30m ago) 16d run=test2root@k8s-deploy:/yaml#
提示：可以看到当我们修改pod标签为其他标签以后，对应rs控制器会新建一个pod，其标签为app=rs-demo，这是因为当我们修改pod标签以后，rs控制器发现标签选择器匹配的pod数量少于用户定义的数量，所以rs控制器会新建一个标签为app=rs-demo的pod；当我们把pod标签修改为rs-demo时，rs控制器发现对应标签选择器匹配pod数量多余用户期望的pod数量，此时rs控制器会通过删除pod方法，让app=rs-demo标签的pod和用户期望的pod数量保持一致；
Deployment 副本控制器，详细说明请参考https://www.cnblogs.com/qiuhom-1874/p/14149042.html；
Deployment副本控制器时k8s第三代pod副本控制器，该控制器比rs控制器更高级，除了有rs的功能之外，还有很多高级功能,，比如说最重要的滚动升级、回滚等；
deploy控制器示例
apiVersion: apps/v1kind: Deploymentmetadata: name: deploy-demo namespace: default labels: app: deploy-demospec: selector: matchLabels: app: deploy-demo replicas: 2 template: metadata: labels: app: deploy-demo spec: containers: - name: deploy-demo image: "harbor.ik8s.cc/baseimages/nginx:v1" ports: - containerPort: 80 name: http volumeMounts: - name: localtime mountPath: /etc/localtime volumes: - name: localtime hostPath: path: /usr/share/zoneinfo/Asia/Shanghai
应用配置清单
提示：deploy控制器是通过创建rs控制器来实现管控对应pod数量；
通过修改镜像版本来更新pod版本
应用配置清单
使用命令更新pod版本
查看rs更新历史版本
查看更新历史记录
提示：这里历史记录中没有记录版本信息，原因是默认不记录，需要记录历史版本，可以手动使用--record选项来记录版本信息；如下所示
查看某个历史版本的详细信息
提示：查看某个历史版本的详细信息，加上--revision=对应历史版本的编号即可；
回滚到上一个版本
提示：使用kubectl rollout undo 命令可以将对应deploy回滚到上一个版本；
回滚指定编号的历史版本
提示：使用--to-revision选项来指定对应历史版本编号，即可回滚到对应编号的历史版本；
Service资源，详细说明请参考https://www.cnblogs.com/qiuhom-1874/p/14161950.html；
nodeport类型的service访问流程
nodeport类型service主要解决了k8s集群外部客户端访问pod，其流程是外部客户端访问k8s集群任意node节点的对应暴露的端口，被访问的node或通过本机的iptables或ipvs规则将外部客户端流量转发给对应pod之上，从而实现外部客户端访问k8s集群pod的目的；通常使用nodeport类型service为了方便外部客户端访问，都会在集群外部部署一个负载均衡器，即外部客户端访问对应负载均衡器的对应端口，通过负载均衡器将外部客户端流量引入k8s集群，从而完成对pod的访问；
ClusterIP类型svc示例
apiVersion: v1kind: Servicemetadata: name: ngx-svc namespace: defaultspec: selector: app: deploy-demo type: ClusterIP ports: - name: http protocol: TCP port: 80 targetPort: 80
应用配置清单
提示：可以看到创建clusterip类型service以后，对应svc会有一个clusterip，后端endpoints会通过标签选择器去关联对应pod，即我们访问对应svc的clusterip，对应流量会被转发至后端endpoint pod之上进行响应；不过这种clusterip类型svc只能在k8s集群内部客户端访问，集群外部客户端是访问不到的，原因是这个clusterip是k8s内部网络IP地址；
验证，访问10.100.100.23的80端口，看看对应后端nginxpod是否可以正常被访问呢？
root@k8s-node01:~# curl 10.100.100.23Welcome to nginx! Welcome to nginx!
If you see this page, the nginx web server is successfully installed andworking. Further configuration is required.
For online documentation and support please refer tonginx.org.
Commercial support is available atnginx.com.Thank you for using nginx.
root@k8s-node01:~#
nodeport类型service示例
apiVersion: v1kind: Servicemetadata: name: ngx-nodeport-svc namespace: defaultspec: selector: app: deploy-demo type: NodePort ports: - name: http protocol: TCP port: 80 targetPort: 80 nodePort: 30012
提示：nodeport类型service只需要在clusterip类型的svc之上修改type为NodePort，然后再ports字段下用nodePort指定对应node端口即可；
应用配置清单
root@k8s-deploy:/yaml# kubectl apply -f nodeport-svc-demo.yamlservice/ngx-nodeport-svc createdroot@k8s-deploy:/yaml# kubectl get svcNAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGEkubernetes ClusterIP 10.100.0.1443/TCP 16dngx-nodeport-svc NodePort 10.100.209.225 80:30012/TCP 11sroot@k8s-deploy:/yaml# kubectl describe svc ngx-nodeport-svcName: ngx-nodeport-svcNamespace: defaultLabels: Annotations: Selector: app=deploy-demoType: NodePortIP Family Policy: SingleStackIP Families: IPv4IP: 10.100.209.225IPs: 10.100.209.225Port: http 80/TCPTargetPort: 80/TCPNodePort: http 30012/TCPEndpoints: 10.200.155.178:80,10.200.211.138:80Session Affinity: NoneExternal Traffic Policy: ClusterEvents: root@k8s-deploy:/yaml#
验证：访问k8s集群任意node的30012端口，看看对应nginxpod是否能够被访问到？
root@k8s-deploy:/yaml# curl 192.168.0.34:30012Welcome to nginx! Welcome to nginx!
If you see this page, the nginx web server is successfully installed andworking. Further configuration is required.
For online documentation and support please refer tonginx.org.
Commercial support is available atnginx.com.Thank you for using nginx.
root@k8s-deploy:/yaml#
提示：可以看到k8s外部客户端访问k8snode节点的30012端口是能够正常访问到nginxpod；当然集群内部的客户端是可以通过对应生成的clusterip进行访问的；
root@k8s-node01:~# curl 10.100.209.225:30012curl: (7) Failed to connect to 10.100.209.225 port 30012 after 0 ms: Connection refusedroot@k8s-node01:~# curl 127.0.0.1:30012 curl: (7) Failed to connect to 127.0.0.1 port 30012 after 0 ms: Connection refusedroot@k8s-node01:~# curl 192.168.0.34:30012Welcome to nginx! Welcome to nginx!
If you see this page, the nginx web server is successfully installed andworking. Further configuration is required.
For online documentation and support please refer tonginx.org.
Commercial support is available atnginx.com.Thank you for using nginx.
root@k8s-node01:~#
提示：集群内部客户端只能访问clusterip的80端口，或者访问node的对外IP的30012端口；
Volume资源，详细说明请参考https://www.cnblogs.com/qiuhom-1874/p/14180752.html；
pod挂载nfs的使用
在nfs服务器上准备数据目录
root@harbor:~# cat /etc/exports# /etc/exports: the access control list for filesystems which may be exported# to NFS clients. See exports(5).## Example for NFSv2 and NFSv3:# /srv/homes hostname1(rw,sync,no_subtree_check) hostname2(ro,sync,no_subtree_check)## Example for NFSv4:# /srv/nfs4 gss/krb5i(rw,sync,fsid=0,crossmnt,no_subtree_check)# /srv/nfs4/homes gss/krb5i(rw,sync,no_subtree_check)#/data/k8sdata/kuboard *(rw,no_root_squash)/data/volumes *(rw,no_root_squash)/pod-vol *(rw,no_root_squash)root@harbor:~# mkdir -p /pod-volroot@harbor:~# ls /pod-vol -d/pod-volroot@harbor:~# exportfs -avexportfs: /etc/exports [1]: Neither "subtree_check" or "no_subtree_check" specified for export "*:/data/k8sdata/kuboard". Assuming default behaviour ("no_subtree_check"). NOTE: this default has changed since nfs-utils version 1.0.xexportfs: /etc/exports [2]: Neither "subtree_check" or "no_subtree_check" specified for export "*:/data/volumes". Assuming default behaviour ("no_subtree_check"). NOTE: this default has changed since nfs-utils version 1.0.xexportfs: /etc/exports [3]: Neither "subtree_check" or "no_subtree_check" specified for export "*:/pod-vol". Assuming default behaviour ("no_subtree_check"). NOTE: this default has changed since nfs-utils version 1.0.xexporting *:/pod-volexporting *:/data/volumesexporting *:/data/k8sdata/kuboardroot@harbor:~# 在pod上挂载nfs目录
apiVersion: apps/v1kind: Deploymentmetadata: name: ngx-nfs-80 namespace: default labels: app: ngx-nfs-80spec: selector: matchLabels: app: ngx-nfs-80 replicas: 1 template: metadata: labels: app: ngx-nfs-80 spec: containers: - name: ngx-nfs-80 image: "harbor.ik8s.cc/baseimages/nginx:v1" resources: requests: cpu: 100m memory: 100Mi limits: cpu: 100m memory: 100Mi ports: - containerPort: 80 name: ngx-nfs-80 volumeMounts: - name: localtime mountPath: /etc/localtime - name: nfs-vol mountPath: /usr/share/nginx/html/ volumes: - name: localtime hostPath: path: /usr/share/zoneinfo/Asia/Shanghai - name: nfs-vol nfs: server: 192.168.0.42 path: /pod-vol restartPolicy: Always---apiVersion: v1kind: Servicemetadata: name: ngx-nfs-svc namespace: defaultspec: selector: app: ngx-nfs-80 type: NodePort ports: - name: ngx-nfs-svc protocol: TCP port: 80 targetPort: 80 nodePort: 30013
应用配置清单
root@k8s-deploy:/yaml# kubectl apply -f nfs-vol.yaml deployment.apps/ngx-nfs-80 createdservice/ngx-nfs-svc createdroot@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEdeploy-demo-6849bdf444-pvsc9 1/1 Running 1 (57m ago) 46hdeploy-demo-6849bdf444-sg8fz 1/1 Running 1 (57m ago) 46hng-rc-l7xmp 1/1 Running 1 (57m ago) 47hng-rc-wl5d6 1/1 Running 1 (57m ago) 47hngx-nfs-80-66c9697cf4-8pm9k 1/1 Running 0 7srs-demo-nzmqs 1/1 Running 1 (57m ago) 47hrs-demo-v2vb6 1/1 Running 1 (57m ago) 47hrs-demo-x27fv 1/1 Running 1 (57m ago) 47htest 1/1 Running 7 (57m ago) 17dtest1 1/1 Running 7 (57m ago) 17dtest2 1/1 Running 7 (57m ago) 17droot@k8s-deploy:/yaml# kubectl get svcNAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGEkubernetes ClusterIP 10.100.0.1443/TCP 18dngx-nfs-svc NodePort 10.100.16.14 80:30013/TCP 15sngx-nodeport-svc NodePort 10.100.209.225 80:30012/TCP 45hroot@k8s-deploy:/yaml#
在nfs服务器上/pod-vol目录下提供index.html文件
root@harbor:~# echo "this page from nfs server.." >> /pod-vol/index.htmlroot@harbor:~# cat /pod-vol/index.htmlthis page from nfs server..root@harbor:~#
访问pod，看看nfs服务器上的inde.html是否能够正常访问到？
root@k8s-deploy:/yaml# curl 192.168.0.35:30013this page from nfs server..root@k8s-deploy:/yaml#
提示：能够看到访问pod对应返回的页面就是刚才在nfs服务器上创建的页面，说明pod正常挂载了nfs提供的目录；
PV、PVC资源，详细说明请参考https://www.cnblogs.com/qiuhom-1874/p/14188621.html；
nfs实现静态pvc的使用
在nfs服务器上准备目录
root@harbor:~# cat /etc/exports # /etc/exports: the access control list for filesystems which may be exported# to NFS clients. See exports(5).## Example for NFSv2 and NFSv3:# /srv/homes hostname1(rw,sync,no_subtree_check) hostname2(ro,sync,no_subtree_check)## Example for NFSv4:# /srv/nfs4 gss/krb5i(rw,sync,fsid=0,crossmnt,no_subtree_check)# /srv/nfs4/homes gss/krb5i(rw,sync,no_subtree_check)#/data/k8sdata/kuboard *(rw,no_root_squash)/data/volumes *(rw,no_root_squash)/pod-vol *(rw,no_root_squash)/data/k8sdata/myserver/myappdata *(rw,no_root_squash)root@harbor:~# mkdir -p /data/k8sdata/myserver/myappdataroot@harbor:~# exportfs -avexportfs: /etc/exports [1]: Neither "subtree_check" or "no_subtree_check" specified for export "*:/data/k8sdata/kuboard". Assuming default behaviour ("no_subtree_check"). NOTE: this default has changed since nfs-utils version 1.0.xexportfs: /etc/exports [2]: Neither "subtree_check" or "no_subtree_check" specified for export "*:/data/volumes". Assuming default behaviour ("no_subtree_check"). NOTE: this default has changed since nfs-utils version 1.0.xexportfs: /etc/exports [3]: Neither "subtree_check" or "no_subtree_check" specified for export "*:/pod-vol". Assuming default behaviour ("no_subtree_check"). NOTE: this default has changed since nfs-utils version 1.0.xexportfs: /etc/exports [4]: Neither "subtree_check" or "no_subtree_check" specified for export "*:/data/k8sdata/myserver/myappdata". Assuming default behaviour ("no_subtree_check"). NOTE: this default has changed since nfs-utils version 1.0.xexporting *:/data/k8sdata/myserver/myappdataexporting *:/pod-volexporting *:/data/volumesexporting *:/data/k8sdata/kuboardroot@harbor:~# 创建pv
apiVersion: v1kind: PersistentVolumemetadata: name: myapp-static-pv namespace: defaultspec: capacity: storage: 2Gi accessModes: - ReadWriteOnce nfs: path: /data/k8sdata/myserver/myappdata server: 192.168.0.42
创建pvc关联pv
apiVersion: v1kind: PersistentVolumeClaimmetadata: name: myapp-static-pvc namespace: defaultspec: volumeName: myapp-static-pv accessModes: - ReadWriteOnce resources: requests: storage: 2Gi
创建pod使用pvc
apiVersion: apps/v1kind: Deploymentmetadata: name: ngx-nfs-pvc-80 namespace: default labels: app: ngx-pvc-80spec: selector: matchLabels: app: ngx-pvc-80 replicas: 1 template: metadata: labels: app: ngx-pvc-80 spec: containers: - name: ngx-pvc-80 image: "harbor.ik8s.cc/baseimages/nginx:v1" resources: requests: cpu: 100m memory: 100Mi limits: cpu: 100m memory: 100Mi ports: - containerPort: 80 name: ngx-pvc-80 volumeMounts: - name: localtime mountPath: /etc/localtime - name: data-pvc mountPath: /usr/share/nginx/html/ volumes: - name: localtime hostPath: path: /usr/share/zoneinfo/Asia/Shanghai - name: data-pvc persistentVolumeClaim: claimName: myapp-static-pvc ---apiVersion: v1kind: Servicemetadata: name: ngx-pvc-svc namespace: defaultspec: selector: app: ngx-pvc-80 type: NodePort ports: - name: ngx-nfs-svc protocol: TCP port: 80 targetPort: 80 nodePort: 30014
应用上述配置清单
root@k8s-deploy:/yaml# kubectl apply -f nfs-static-pvc-demo.yaml persistentvolume/myapp-static-pv createdpersistentvolumeclaim/myapp-static-pvc createddeployment.apps/ngx-nfs-pvc-80 createdservice/ngx-pvc-svc createdroot@k8s-deploy:/yaml# kubectl get pvNAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGEmyapp-static-pv 2Gi RWO Retain Bound default/myapp-static-pvc 4sroot@k8s-deploy:/yaml# kubectl get pvcNAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGEmyapp-static-pvc Pending myapp-static-pv 0 7sroot@k8s-deploy:/yaml# kubectl get pods NAME READY STATUS RESTARTS AGEdeploy-demo-6849bdf444-pvsc9 1/1 Running 1 (151m ago) 47hdeploy-demo-6849bdf444-sg8fz 1/1 Running 1 (151m ago) 47hng-rc-l7xmp 1/1 Running 1 (151m ago) 2d1hng-rc-wl5d6 1/1 Running 1 (151m ago) 2d1hngx-nfs-pvc-80-f776bb6d-nwwwq 0/1 Pending 0 10srs-demo-nzmqs 1/1 Running 1 (151m ago) 2drs-demo-v2vb6 1/1 Running 1 (151m ago) 2drs-demo-x27fv 1/1 Running 1 (151m ago) 2dtest 1/1 Running 7 (151m ago) 18dtest1 1/1 Running 7 (151m ago) 18dtest2 1/1 Running 7 (151m ago) 18droot@k8s-deploy:/yaml#
在nfs服务器上/data/k8sdata/myserver/myappdata创建index.html，看看对应主页是否能够被访问？
root@harbor:~# echo "this page from nfs-server /data/k8sdata/myserver/myappdata/index.html" >> /data/k8sdata/myserver/myappdata/index.htmlroot@harbor:~# cat /data/k8sdata/myserver/myappdata/index.htmlthis page from nfs-server /data/k8sdata/myserver/myappdata/index.htmlroot@harbor:~#
访问pod
root@harbor:~# curl 192.168.0.36:30014this page from nfs-server /data/k8sdata/myserver/myappdata/index.htmlroot@harbor:~#
nfs实现动态pvc的使用
创建名称空间、服务账号、clusterrole、clusterrolebindding、role、rolebinding
apiVersion: v1kind: Namespacemetadata: name: nfs---apiVersion: v1kind: ServiceAccountmetadata: name: nfs-client-provisioner # replace with namespace where provisioner is deployed namespace: nfs---kind: ClusterRoleapiVersion: rbac.authorization.k8s.io/v1metadata: name: nfs-client-provisioner-runnerrules: - apiGroups: [""] resources: ["nodes"] verbs: ["get", "list", "watch"] - apiGroups: [""] resources: ["persistentvolumes"] verbs: ["get", "list", "watch", "create", "delete"] - apiGroups: [""] resources: ["persistentvolumeclaims"] verbs: ["get", "list", "watch", "update"] - apiGroups: ["storage.k8s.io"] resources: ["storageclasses"] verbs: ["get", "list", "watch"] - apiGroups: [""] resources: ["events"] verbs: ["create", "update", "patch"]---kind: ClusterRoleBindingapiVersion: rbac.authorization.k8s.io/v1metadata: name: run-nfs-client-provisionersubjects: - kind: ServiceAccount name: nfs-client-provisioner # replace with namespace where provisioner is deployed namespace: nfsroleRef: kind: ClusterRole name: nfs-client-provisioner-runner apiGroup: rbac.authorization.k8s.io---kind: RoleapiVersion: rbac.authorization.k8s.io/v1metadata: name: leader-locking-nfs-client-provisioner # replace with namespace where provisioner is deployed namespace: nfsrules: - apiGroups: [""] resources: ["endpoints"] verbs: ["get", "list", "watch", "create", "update", "patch"]---kind: RoleBindingapiVersion: rbac.authorization.k8s.io/v1metadata: name: leader-locking-nfs-client-provisioner # replace with namespace where provisioner is deployed namespace: nfssubjects: - kind: ServiceAccount name: nfs-client-provisioner # replace with namespace where provisioner is deployed namespace: nfsroleRef: kind: Role name: leader-locking-nfs-client-provisioner apiGroup: rbac.authorization.k8s.io
创建sc
apiVersion: storage.k8s.io/v1kind: StorageClassmetadata: name: managed-nfs-storageprovisioner: k8s-sigs.io/nfs-subdir-external-provisioner # or choose another name, must match deployment"s env PROVISIONER_NAME"reclaimPolicy: Retain #PV的删除策略，默认为delete，删除PV后立即删除NFS server的数据mountOptions: #- vers=4.1 #containerd有部分参数异常 #- noresvport #告知NFS客户端在重新建立网络连接时，使用新的传输控制协议源端口 - noatime #访问文件时不更新文件inode中的时间戳，高并发环境可提高性能parameters: #mountOptions: "vers=4.1,noresvport,noatime" archiveOnDelete: "true" #删除pod时保留pod数据，默认为false时为不保留数据
创建provision
apiVersion: apps/v1kind: Deploymentmetadata: name: nfs-client-provisioner labels: app: nfs-client-provisioner # replace with namespace where provisioner is deployed namespace: nfsspec: replicas: 1 strategy: #部署策略 type: Recreate selector: matchLabels: app: nfs-client-provisioner template: metadata: labels: app: nfs-client-provisioner spec: serviceAccountName: nfs-client-provisioner containers: - name: nfs-client-provisioner #image: k8s.gcr.io/sig-storage/nfs-subdir-external-provisioner:v4.0.2 image: registry.cn-qingdao.aliyuncs.com/zhangshijie/nfs-subdir-external-provisioner:v4.0.2 volumeMounts: - name: nfs-client-root mountPath: /persistentvolumes env: - name: PROVISIONER_NAME value: k8s-sigs.io/nfs-subdir-external-provisioner - name: NFS_SERVER value: 192.168.0.42 - name: NFS_PATH value: /data/volumes volumes: - name: nfs-client-root nfs: server: 192.168.0.42 path: /data/volumes
调用sc创建pvc
apiVersion: v1kind: Namespacemetadata: name: myserver--- # Test PVCkind: PersistentVolumeClaimapiVersion: v1metadata: name: myserver-myapp-dynamic-pvc namespace: myserverspec: storageClassName: managed-nfs-storage #调用的storageclass 名称 accessModes: - ReadWriteMany #访问权限 resources: requests: storage: 500Mi #空间大小
创建app使用pvc
kind: Deployment#apiVersion: extensions/v1beta1apiVersion: apps/v1metadata: labels: app: myserver-myapp name: myserver-myapp-deployment-name namespace: myserverspec: replicas: 1 selector: matchLabels: app: myserver-myapp-frontend template: metadata: labels: app: myserver-myapp-frontend spec: containers: - name: myserver-myapp-container image: nginx:1.20.0 #imagePullPolicy: Always volumeMounts: - mountPath: "/usr/share/nginx/html/statics" name: statics-datadir volumes: - name: statics-datadir persistentVolumeClaim: claimName: myserver-myapp-dynamic-pvc ---kind: ServiceapiVersion: v1metadata: labels: app: myserver-myapp-service name: myserver-myapp-service-name namespace: myserverspec: type: NodePort ports: - name: http port: 80 targetPort: 80 nodePort: 30015 selector: app: myserver-myapp-frontend
应用上述配置清单
root@k8s-deploy:/yaml/myapp# kubectl apply -f .namespace/nfs createdserviceaccount/nfs-client-provisioner createdclusterrole.rbac.authorization.k8s.io/nfs-client-provisioner-runner createdclusterrolebinding.rbac.authorization.k8s.io/run-nfs-client-provisioner createdrole.rbac.authorization.k8s.io/leader-locking-nfs-client-provisioner createdrolebinding.rbac.authorization.k8s.io/leader-locking-nfs-client-provisioner createdstorageclass.storage.k8s.io/managed-nfs-storage createddeployment.apps/nfs-client-provisioner creatednamespace/myserver createdpersistentvolumeclaim/myserver-myapp-dynamic-pvc createddeployment.apps/myserver-myapp-deployment-name createdservice/myserver-myapp-service-name createdroot@k8s-deploy:
验证：查看sc、pv、pvc是否创建？pod是否正常运行？
root@k8s-deploy:/yaml/myapp# kubectl get scNAME PROVISIONER RECLAIMPOLICY VOLUMEBINDINGMODE ALLOWVOLUMEEXPANSION AGEmanaged-nfs-storage k8s-sigs.io/nfs-subdir-external-provisioner Retain Immediate false 105sroot@k8s-deploy:/yaml/myapp# kubectl get pvNAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGEpvc-01709c7f-0cf9-4554-9ae9-72db89e7308c 500Mi RWX Retain Bound myserver/myserver-myapp-dynamic-pvc managed-nfs-storage 107sroot@k8s-deploy:/yaml/myapp# kubectl get pvc -n myserverNAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGEmyserver-myapp-dynamic-pvc Bound pvc-01709c7f-0cf9-4554-9ae9-72db89e7308c 500Mi RWX managed-nfs-storage 117sroot@k8s-deploy:/yaml/myapp# kubectl get pods -n myserverNAME READY STATUS RESTARTS AGEmyserver-myapp-deployment-name-65ff65446f-xpd5p 1/1 Running 0 2m8sroot@k8s-deploy:/yaml/myapp#
提示：可以看到pv自动由sc创建，pvc自动和pv关联；
验证：在nfs服务器上的/data/volumes/下创建index.html文件，访问pod service，看看对应文件是否能够正常被访问到？
root@harbor:/data/volumes# lsmyserver-myserver-myapp-dynamic-pvc-pvc-01709c7f-0cf9-4554-9ae9-72db89e7308croot@harbor:/data/volumes# cd myserver-myserver-myapp-dynamic-pvc-pvc-01709c7f-0cf9-4554-9ae9-72db89e7308c/root@harbor:/data/volumes/myserver-myserver-myapp-dynamic-pvc-pvc-01709c7f-0cf9-4554-9ae9-72db89e7308c# lsroot@harbor:/data/volumes/myserver-myserver-myapp-dynamic-pvc-pvc-01709c7f-0cf9-4554-9ae9-72db89e7308c# echo "this page from nfs-server /data/volumes" >> index.htmlroot@harbor:/data/volumes/myserver-myserver-myapp-dynamic-pvc-pvc-01709c7f-0cf9-4554-9ae9-72db89e7308c# cat index.html this page from nfs-server /data/volumesroot@harbor:/data/volumes/myserver-myserver-myapp-dynamic-pvc-pvc-01709c7f-0cf9-4554-9ae9-72db89e7308c#
提示：在nfs服务器上的/data/volumes目录下会自动生成一个使用pvcpod所在名称空间+pvc名字+pv名字的一个目录，这个目录就是由provision创建；
访问pod
root@harbor:~# curl 192.168.0.36:30015/statics/index.htmlthis page from nfs-server /data/volumesroot@harbor:~#
提示：能够访问到我们刚才创建的文件，说明pod正常挂载nfs服务器对应目录；
PV/PVC总结
PV是对底层网络存储的抽象，即将网络存储定义为一种存储资源，将一个整体的存储资源拆分成多份后给不同的业务使用。
PVC是对PV资源的申请调用，pod是通过PVC将数据保存至PV，PV再把数据保存至真正的硬件存储。
PersistentVolume参数
Capacity： #当前PV空间大小，kubectl explain PersistentVolume.spec.capacity
accessModes ：访问模式，#kubectl explain PersistentVolume.spec.accessModes
ReadWriteOnce – PV只能被单个节点以读写权限挂载，RWO
ReadOnlyMany – PV以可以被多个节点挂载但是权限是只读的,ROX ReadWriteMany – PV可以被多个节点是读写方式挂载使用,RWX
persistentVolumeReclaimPolicy #删除机制即删除存储卷卷时候，已经创建好的存储卷由以下删除操作：
Retain – 删除PV后保持原装，最后需要管理员手动删除
Recycle – 空间回收，及删除存储卷上的所有数据(包括目录和隐藏文件),目前仅支持NFS和hostPath
Delete – 自动删除存储卷
volumeMode #卷类型，kubectl explain PersistentVolume.spec.volumeMode；定义存储卷使用的文件系统是块设备还是文件系统，默认为文件系统
mountOptions #附加的挂载选项列表，实现更精细的权限控制；
官方文档：持久卷 | Kubernetes；
PersistentVolumeClaim创建参数
accessModes ：PVC 访问模式，#kubectl explain PersistentVolumeClaim.spec.volumeMode
ReadWriteOnce – PVC只能被单个节点以读写权限挂载，RWO
ReadOnlyMany – PVC以可以被多个节点挂载但是权限是只读的,ROX
ReadWriteMany – PVC可以被多个节点是读写方式挂载使用,RWX
resources： #定义PVC创建存储卷的空间大小
selector： #标签选择器，选择要绑定的PV
matchLabels #匹配标签名称
matchExpressions #基于正则表达式匹配
volumeName #要绑定的PV名称
volumeMode #卷类型，定义PVC使用的文件系统是块设备还是文件系统，默认为文件系统
Volume- 存储卷类型
static：静态存储卷 ,需要在使用前手动创建PV、然后创建PVC并绑定到PV然后挂载至pod使用，适用于PV和PVC相对比较固定的业务场景。
dynamin：动态存储卷,先创建一个存储类storageclass，后期pod在使用PVC的时候可以通过存储类动态创建PVC，适用于有状态服务集群如MySQL一主多从、zookeeper集群等。
存储类官方文档：存储类 | Kubernetes
