Kubernets包管理工具—>Helm
什么是Helm?
我们都知道,Linux系统各发行版都有自己的包管理工具,比如Centos的YUM,再如Ubuntu的APT。
Kubernetes也有自己的集群包管理工具,那就是Helm。
Helm本质就是让K8S的应用管理(Deployment,Service等)可配置,能动态生成。通过动态生成K8S资源清单文件(deployment.yaml,service.yaml),然后调用kubectl自动执行K8S部署。
Helm有两个重要的概念,chart 和 release
- chart 是创建一个应用的信息集合,包括各种Kubernetes对象的配置模板、参数定义、依赖关系和文档说明等,chart是应用部署的自包含逻辑单元。可以将 chart 想象成apt、yum中的软件安装包。
- release 是chart的运行实例,代表了一个正在运行的应用,当chart被安装到kubernetes集群,就生成一个release,chart能够多次安装到同一个集群,每次安装都是一个release。
Helm包含两个组件,Helm 客户端和 Tiller 服务器
- Helm客户端负责chart和release的创建和管理,以及和Tiller的交互。
- Tiller服务运行在 Kubernetes 集群中,它会处理Helm客户端的请求,与 Kubernetes API Server 交互
Helm 部署
越来越多的公司开始使用Helm这个Kubernetes包管理工具,Helm的安装也十分简单,下载 helm 命令行工具到Master节点即可,以下示例安装为Helm v2.16.10 版本,包下载地址:https://github.com/helm/helm/releases
[root@Centos8 heml]# wget https://get.helm.sh/helm-v2.16.10-linux-amd64.tar.gz
[root@Centos8 heml]# tar zxvf helm-v2.16.10-linux-amd64.tar.gz -C /usr/local/
[root@Centos8 heml]# cd /usr/local/linux-amd64/
[root@Centos8 linux-amd64]# ln -s pwd/helm /usr/local/bin/
以上Helm命令安装完成,官方文档:https://helm.sh/docs/intro/install/#helm
为了安装tiller,还需要在这台机器上配置好kubectl工具和kubeconfig文件,确保kubectl工具可以在这台机器访问apiserver且正常使用。
因为Kubernetes ApiServer开启了RBAC访问控制,所以需要创建tiller使用的 service account:tiller并分配合适的角色给它。这里简单起见直接分配cluster-admin这个集群内置的CluserRole给它。创建rbac-config.yaml文件:
vimrbac-config.yaml
apiVersion: v1
kind: ServiceAccount
metadata:
name: tiller
namespace: kube-system
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
name: tiller
subjects:
- kind: ServiceAccount
name: tiller
namespace: kube-system
roleRef:
kind: ClusterRole
name: cluster-admin
apiGroup: rbac.authorization.k8s.io
[root@Centos8 rbac]# kubectl create -f rbac-config.yaml
serviceaccount/tiller created
clusterrolebinding.rbac.authorization.k8s.io/tiller created
在K8S集群中初始化helm
[root@Centos8 rbac]# helm init --service-account tiller --skip-refresh
Creating /root/.helm
Creating /root/.helm/repository
Creating /root/.helm/repository/cache
Creating /root/.helm/repository/local
Creating /root/.helm/plugins
Creating /root/.helm/starters
Creating /root/.helm/cache/archive
Creating /root/.helm/repository/repositories.yaml
Adding stable repo with URL: https://kubernetes-charts.storage.googleapis.com
Adding local repo with URL: http://127.0.0.1:8879/charts
$HELM_HOME has been configured at /root/.helm.
Tiller (the Helm server-side component) has been installed into your Kubernetes Cluster.
Please note: by default, Tiller is deployed with an insecure 'allow unauthenticated users' policy.
To prevent this, run helm init with the --tiller-tls-verify flag.
For more information on securing your installation see: https://v2.helm.sh/docs/securing_installation/
[root@Centos8 rbac]# kubectl get pod -n kube-system
NAME READY STATUS RESTARTS AGE
tiller-deploy-8487d94bcf-nfc74 0/1 ContainerCreating 0 98s
[root@Centos8 ~]# kubectl describe pod tiller-deploy-8487d94bcf-nfc74 -n kube-system
Back-off pulling image "gcr.io/kubernetes-helm/tiller:v2.16.10"
会发现tiller的Pod Running不起来,是因为导入镜像失败,因为网络的原因访问gcr.io访问不到,于是到docker hub中查询此image,发现确实有相同的image,pull过来改个名即可。
[root@Centos8 ~]# docker pull jessestuart/tiller:v2.16.10
Status: Downloaded newer image for jessestuart/tiller:v2.16.10
docker.io/jessestuart/tiller:v2.16.10
docker tag jessestuart/tiller:v2.16.10 gcr.io/kubernetes-helm/tiller:v2.16.10
然后传输到每一个node节点上:
[root@Centos8 ~]# docker save gcr.io/kubernetes-helm/tiller -o /usr/local/install-k8s/heml/tiller.tgz
[root@Centos8 ~]# scp /usr/local/install-k8s/heml/tiller.tgz 192.168.152.253:/usr/local/install-k8s/
node节点接收到后,再导入成image即可:
[root@TestCentos7 install-k8s]# docker load < tiller.tgz
Loaded image: gcr.io/kubernetes-helm/tiller:v2.16.10
再次查看tiller Pod的状态,已经变为Running:
[root@Centos8 ~]# kubectl get pod -n kube-system
tiller-deploy-8487d94bcf-nfc74 1/1 Running 0 1h
Helm的使用
Helm的使用与yum、apt等工具如出一辙,可以事先去helm hub中寻找想要安装的工具或应用:https://hub.helm.sh/,其页面会有具体的安装方法及步骤。
以安装redis为例:https://hub.helm.sh/charts/choerodon/redis
1、 先添加redis的repo源;
helm repo add choerodon https://openchart.choerodon.com.cn/choerodon/c7n
"choerodon" has been added to your repositories
2、 更新一下helmrepo;
[root@Centos8 ~]# helm repo update
Hang tight while we grab the latest from your chart repositories...
...Skip local chart repository
...Successfully got an update from the "choerodon" chart repository
...Successfully got an update from the "stable" chart repository
Update Complete.
3、 开始安装;
[root@Centos8 ~]# helm install choerodon/redis --version 0.2.5
NAME: exhaling-yak
LAST DEPLOYED: Sun Sep 6 22:57:51 2020
NAMESPACE: default
STATUS: DEPLOYED
RESOURCES:
==> v1/ConfigMap
NAME DATA AGE
exhaling-yak-cm 1 0s
==> v1/Deployment
NAME READY UP-TO-DATE AVAILABLE AGE
exhaling-yak 0/1 0 0 0s
==> v1/Pod(related)
4、 可以看到,在default名称空间生成了ConfigMap、Deployment和Pod;
[root@Centos8 ~]# kubectl get pod
NAME READY STATUS RESTARTS AGE
exhaling-yak-cdc8cf8f9-xqtk9 0/1 ImagePullBackOff 0 40s
[root@Centos8 ~]# kubectl get deployment
NAME READY UP-TO-DATE AVAILABLE AGE
exhaling-yak 0/1 1 0 85s
[root@Centos8 ~]# kubectl get cm
NAME DATA AGE
exhaling-yak-cm 1 109s
Pod ImagePullBackOff 的原因是redis镜像没导入成功,再自行pull即可
3、 Helm常用命令,大家可以通过helm--help进行了解;
Helm自定义模板
以上拉取的都是别人自定义好的模板,自己也可以做一些模板上传或者收藏起来。在此测试创建hello-world模板
1.创建好模板所有文件所放置的目录
mkdir charts
cd charts/
mkdir templates 必须创建一个名字为 templates 的目录
2.编辑Chart.yaml
vim Chart.yaml # 必须创建一个名为 Chart.yaml 的文件,并指定 name 和 version 两个key的值
name: hello-world
version: 1.0.0
3.在 templates 目录下创建 deployment 及 service
vimtemplates/deployments.yaml
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
name: hello-world
spec:
replicas: 1
template:
metadata:
labels:
app: hello-world
spec:
containers:
- name: hello-world
image: nginx:1.2.1
imagePullPolicy: IfNotPresent
ports:
- containerPort: 80
vimservices.yaml
apiVersion: v1
kind: Service
metadata:
name: hello-world
spec:
type: NodePort
ports:
- port: 80
containerPort: 80
nodePort: 30001
selector:
app: hello-world
此时整体目录结构为:
[root@Centos8 charts]# tree
.
├── Chart.yaml
└── templates
├── deployments.yaml
└── services.yaml
4.安装此自定义chart
[root@Centos8 charts]# helm install .
NAME: wishing-badger
LAST DEPLOYED: Mon Sep 7 20:55:42 2020
NAMESPACE: default
STATUS: DEPLOYED
RESOURCES:
==> v1/Pod(related)
NAME READY STATUS RESTARTS AGE
hello-world-767c98894d-7lrzt 0/1 ContainerCreating 0 1s
==> v1/Service
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
hello-world NodePort 10.100.108.217 <none> 80:30001/TCP 0s
==> v1beta1/Deployment
NAME READY UP-TO-DATE AVAILABLE AGE
hello-world 0/1 1 0 0s
查看Pod、Deployment、Service
[root@Centos8 charts]# kubectl get pod
NAME READY STATUS RESTARTS AGE
hello-world-767c98894d-7lrzt 1/1 Running 0 67s
[root@Centos8 charts]# kubectl get deployment
NAME READY UP-TO-DATE AVAILABLE AGE
hello-world 1/1 1 1 78s
[root@Centos8 charts]# kubectl get svc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
hello-world NodePort 10.100.108.217 <none> 80:30001/TCP 81s
Helm常用命令及用法
1.更新镜像
第一种:手动更新
进入deployments.yaml 修改 image 行,然后helm upgrade
vimtemplates/deployments.yaml
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
name: hello-world
spec:
replicas: 1
template:
metadata:
labels:
app: hello-world
spec:
containers:
- name: hello-world
image: hub.vfancloud.com/test/myapp
imagePullPolicy: IfNotPresent
ports:
- containerPort: 80
[root@Centos8 charts]# helm upgrade wishing-badger .
Release "wishing-badger" has been upgraded.
LAST DEPLOYED: Mon Sep 7 21:07:04 2020
NAMESPACE: default
STATUS: DEPLOYED
RESOURCES:
==> v1/Pod(related)
NAME READY STATUS RESTARTS AGE
hello-world-7466c45989-cxnps 0/1 Terminating 0 69s
hello-world-864f865db8-zjt79 0/1 ContainerCreating 0 0s
==> v1/Service
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
hello-world NodePort 10.100.108.217 <none> 80:30001/TCP 11m
==> v1beta1/Deployment
NAME READY UP-TO-DATE AVAILABLE AGE
hello-world 0/1 1 0 11m
查看index.html,版本为v1
[root@Centos8 charts]# curl http://10.100.108.217
Hello MyApp | Version: v1 | <a href="hostname.html">Pod Name</a>
第二种:变量更新
创建一个变量文件 values.yaml,保存 image 及 tag
vimvalues.yaml
image:
repository: hub.vfancloud.com/test/myapp
tag: 'v2'
vim templates/deployments.yaml # 将 image 字段更改为以上文件的变量
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
name: hello-world
spec:
replicas: 1
template:
metadata:
labels:
app: hello-world
spec:
containers:
- name: hello-world
image: {{ .Values.image.repository }}:{{ .Values.image.tag }}
imagePullPolicy: IfNotPresent
ports:
- containerPort: 80
开始更新
[root@Centos8 charts]# helm upgrade wishing-badger .
Release "wishing-badger" has been upgraded.
LAST DEPLOYED: Mon Sep 7 21:17:31 2020
NAMESPACE: default
STATUS: DEPLOYED
RESOURCES:
==> v1/Pod(related)
NAME READY STATUS RESTARTS AGE
hello-world-5759c969fc-w9s88 0/1 ContainerCreating 0 0s
hello-world-864f865db8-zjt79 1/1 Terminating 0 10m
==> v1/Service
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
hello-world NodePort 10.100.108.217 <none> 80:30001/TCP 21m
==> v1beta1/Deployment
NAME READY UP-TO-DATE AVAILABLE AGE
hello-world 1/1 1 1 21m
查看index.html,版本为v2
[root@Centos8 charts]# curl http://10.100.108.217
Hello MyApp | Version: v2 | <a href="hostname.html">Pod Name</a>
或者通过命令行直接修改 image 的 tag 标签,来更新镜像版本
[root@Centos8 charts]# helm upgrade wishing-badger --set image.tag='v3' .
Release "wishing-badger" has been upgraded.
LAST DEPLOYED: Mon Sep 7 21:27:04 2020
NAMESPACE: default
STATUS: DEPLOYED
RESOURCES:
==> v1/Pod(related)
NAME READY STATUS RESTARTS AGE
hello-world-5759c969fc-w9s88 1/1 Terminating 0 9m33s
hello-world-6454b8dcc8-pjgk9 0/1 ContainerCreating 0 0s
==> v1/Service
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
hello-world NodePort 10.100.108.217 <none> 80:30001/TCP 31m
==> v1beta1/Deployment
NAME READY UP-TO-DATE AVAILABLE AGE
hello-world 0/1 1 0 31m
查看index.html,已经更新为v3
[root@Centos8 charts]# curl http://10.100.108.217
Hello MyApp | Version: v3 | <a href="hostname.html">Pod Name</a>
2.查看release历史版本
[root@Centos8 charts]# helm history wishing-badger
REVISION UPDATED STATUS CHART APP VERSION DESCRIPTION
1 Mon Sep 7 20:55:42 2020 SUPERSEDED hello-world-1.0.0 Install complete
2 Mon Sep 7 21:07:04 2020 DEPLOYED hello-world-1.0.0 Upgrade complete
3.删除release
[root@Centos8 charts]# helm delete wishing-badger
release "wishing-badger" deleted
以上命令提示此 release 删除,但其实并没有完全“删除”,而是将它放回了“回收站”
原因是怕你将来有一天想要回滚,“回收站”查看方法:
[root@Centos8 charts]# helm list --deleted
NAME REVISION UPDATED STATUS CHART APP VERSION NAMESPACE
wishing-badger 5 Mon Sep 7 21:27:04 2020 DELETED hello-world-1.0.0 default
如果想要彻彻底底的删除,在执行 delete 时加上 --purge 即可
4.回滚release
helm rollback [name] [版本]
[root@Centos8 charts]# helm rollback wishing-badger 2
Rollback was a success.
将wishing-badger 回滚到第二个版本
查看index.html,已回退到第二个版本,version 为 v1
[root@Centos8 charts]# curl http://10.109.145.22
Hello MyApp | Version: v1 | <a href="hostname.html">Pod Name</a>