kubernetes(k8s)：包管理工具Helm
程序员文章站
2022-07-10 10:39:28
...
1.Helm简介
- Helm是Kubernetes 应用的包管理工具，主要用来管理 Charts，类似Linux系统的yum。
- Helm Chart 是用来封装 Kubernetes 原生应用程序的一系列 YAML 文件。可以在你部署应用的时候自定义应用程序的一些Metadata，以便于应用程序的分发。
- 对于应用发布者而言，可以通过 Helm 打包应用、管理应用依赖关系、管理应用版本并发布应用到软件仓库。
- 对于使用者而言，使用 Helm 后不用需要编写复杂的应用部署文件，可以以简单的方式在 Kubernetes上查找、安装、升级、回滚、卸载应用程序。
Helm当前最新版本 v3.2.0 官网：https://helm.sh/docs/intro/Helm V3 与 V2 最大的区别在于去掉了tiller，见下图：
本节使用的是 Helm V3
2. Helm的部署
Helm当前最新版本 v3.1.0 官网：https://helm.sh/docs/intro/
Helm安装：
下载软件包：helm-v3.1.1-linux-amd64.tar.gz
[aaa@qq.com ~]# tar zxf helm-v3.1.1-linux-amd64.tar.gz
[aaa@qq.com ~]# cd linux-amd64/
[aaa@qq.com linux-amd64]# ls
helm LICENSE README.md
[aaa@qq.com ~]# cp helm /usr/local/bin/
设置helm命令补齐：
[aaa@qq.com ~]$ echo "source <(helm completion bash)" >> ~/.bashrc
重新登陆生效
[aaa@qq.com ~]# su - kubeadm
Last login: Mon May 11 09:16:14 CST 2020 on pts/1
Last failed login: Mon May 11 11:38:46 CST 2020 on pts/1
There was 1 failed login attempt since the last successful login.
[aaa@qq.com ~]$ helm //helm命令
completion dependency get install list plugin repo search status test upgrade version
create env history lint package pull rollback show template uninstall verify
Helm 添加第三方 Chart 库：
[aaa@qq.com ~]$ helm repo add stable http://mirror.azure.cn/kubernetes/charts/
"stable" has been added to your repositories
[aaa@qq.com ~]$ helm repo list
NAME URL
stable http://mirror.azure.cn/kubernetes/charts/
[aaa@qq.com ~]$ helm search repo redis
NAME CHART VERSION APP VERSION DESCRIPTION
stable/prometheus-redis-exporter 3.4.0 1.3.4 Prometheus exporter for Redis metrics
stable/redis 10.5.7 5.0.7 DEPRECATED Open source, advanced key-value stor...
stable/redis-ha 4.4.4 5.0.6 Highly available Kubernetes implementation of R...
stable/sensu 0.2.3 0.28 Sensu monitoring framework backed by the Redis ...
3. Helm 部署应用(拉取仓库的redis-ha并部署)
[aaa@qq.com ~]$ helm search repo redis-ha -l // 可以查看redis-ha的所有版本
NAME CHART VERSION APP VERSION DESCRIPTION
stable/redis-ha 4.4.4 5.0.6 Highly available Kubernetes implementation of R...
stable/redis-ha 4.4.3 5.0.6 Highly available Kubernetes implementation of R...
stable/redis-ha 4.4.2 5.0.6 Highly available Kubernetes implementation of R...
stable/redis-ha 4.4.1 5.0.6 Highly available Kubernetes implementation of R...
stable/redis-ha 4.4.0 5.0.6 Highly available Kubernetes implementation of R...
stable/redis-ha 4.3.4 5.0.6 Highly available Kubernetes implementation of R...
[aaa@qq.com ~]$ helm pull stable/redis-ha // 拉取应用到本地，默认拉取最新版本
[aaa@qq.com ~]$ ls // 拉取了一个redis-ha的压缩包
components.yaml dashrbac.yaml kube-flannel.yml manifest recommended.yaml redis-ha-4.4.4.tgz
[aaa@qq.com ~]$ mkdir helm
[aaa@qq.com ~]$ tar zxf redis-ha-4.4.4.tgz -C helm/ // 解压到helm目录
[aaa@qq.com redis-ha]$ sudo yum install tree -y
[aaa@qq.com redis-ha]$ tree . // 查看目录结构
.
├── Chart.yaml
├── ci
│ └── haproxy-enabled-values.yaml
├── OWNERS
├── README.md
├── templates
│ ├── _configs.tpl
│ ├── _helpers.tpl
│ ├── NOTES.txt
│ ├── redis-auth-secret.yaml
│ ├── redis-ha-announce-service.yaml
│ ├── redis-ha-configmap.yaml
│ ├── redis-ha-exporter-script-configmap.yaml
│ ├── redis-ha-pdb.yaml
│ ├── redis-haproxy-deployment.yaml
│ ├── redis-haproxy-serviceaccount.yaml
│ ├── redis-haproxy-servicemonitor.yaml
│ ├── redis-haproxy-service.yaml
│ ├── redis-ha-rolebinding.yaml
│ ├── redis-ha-role.yaml
│ ├── redis-ha-serviceaccount.yaml
│ ├── redis-ha-servicemonitor.yaml
│ ├── redis-ha-service.yaml
│ ├── redis-ha-statefulset.yaml
│ └── tests
│ ├── test-redis-ha-configmap.yaml
│ └── test-redis-ha-pod.yaml
└── values.yaml
3 directories, 25 files
我们重点关注的是 values.yaml文件 进行修改和编辑
案例：部署redis-ha，实现高可用
（1）将镜像拉取到本地harbor之后
需要redis:5.0.6-alpine的镜像
[aaa@qq.com harbor]# docker pull redis:5.0.6-alpine
[aaa@qq.com harbor]# docker tag redis:5.0.6-alpine reg.westos.org/library/redis:5.0.6-alpine
[aaa@qq.com harbor]# docker push reg.westos.org/library/redis
[aaa@qq.com redis-ha]$ kubectl create namespace redis //创建redis命名空间
[aaa@qq.com redis-ha]$ helm -n redis install redis //在redisnamespace下安装redis
NAME: redis-ha
LAST DEPLOYED: Mon May 11 13:04:20 2020
NAMESPACE: redis
STATUS: deployed
REVISION: 1
NOTES:
Redis can be accessed via port 6379 and Sentinel can be accessed via port 26379 on the following DNS name from within your cluster:
redis-ha.redis.svc.cluster.local
To connect to your Redis server:
1. Run a Redis pod that you can use as a client:
kubectl exec -it redis-ha-server-0 sh -n redis
2. Connect using the Redis CLI:
redis-cli -h redis-ha.redis.svc.cluster.local
（2）创建后查看：
[aaa@qq.com redis-ha]$ kubectl -n redis get all // redis-ha部署完成
NAME READY STATUS RESTARTS AGE
pod/redis-ha-server-0 2/2 Running 0 2m16s
pod/redis-ha-server-1 2/2 Running 0 107s
pod/redis-ha-server-2 0/2 Pending 0 77s
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/redis-ha ClusterIP None <none> 6379/TCP,26379/TCP 2m19s
service/redis-ha-announce-0 ClusterIP 10.100.4.107 <none> 6379/TCP,26379/TCP 2m18s
service/redis-ha-announce-1 ClusterIP 10.109.211.70 <none> 6379/TCP,26379/TCP 2m18s
service/redis-ha-announce-2 ClusterIP 10.105.156.107 <none> 6379/TCP,26379/TCP 2m18s
NAME READY AGE
statefulset.apps/redis-ha-server 2/3 2m17s
[aaa@qq.com redis-ha]$ kubectl -n redis describe pod redis-ha-server-0 //pod中由两个容器redis和sentinel
Containers:
redis:
sentinel: // 负责主从切换
可以看到创建了两个pod，其中server-0为master，还有两个svc和一个statefulset控制器。
（3）测试redis集群高可用
每个pod中有两个容器redis和sentinel，我们进入redis容器查看：
[aaa@qq.com redis-ha]$ kubectl -n redis exec -it redis-ha-server-0 -c redis -- sh // 进入redis-ha-server-0，-c 指定进入的容器时redis
/data $ redis-cli
127.0.0.1:6379> info // 查看master的信息
/data $ redis-cli -h 10.109.211.70 // 进入salve
10.109.211.70:6379> info
/data $ nslookup redis-ha.redis.svc.cluster.local
nslookup: can't resolve '(null)': Name does not resolve
Name: redis-ha.redis.svc.cluster.local
Address 1: 10.244.1.147 redis-ha-server-0.redis-ha.redis.svc.cluster.local
Address 2: 10.244.2.131 redis-ha-server-1.redis-ha.redis.svc.cluster.local
/data $ redis-cli -h redis-ha-server-1.redis-ha.redis.svc.cluster.local
redis-ha-server-1.redis-ha.redis.svc.cluster.local:6379> info // 1是slave
# Replication
role:slave
/data $ redis-cli -h redis-ha-server-0.redis-ha.redis.svc.cluster.local
redis-ha-server-0.redis-ha.redis.svc.cluster.local:6379> info // 0是master
# Replication
role:master
helm -n redis upgrade redis-ha // 更新
helm status redis-ha // 查看状态
helm -n redis uninstall redis-ha // 删除卸载
kubectl delete namespaces redis // 删除namespace
helm -n redis history redis-ha // 查看历史
[aaa@qq.com redis-ha]# kubectl -n redis exec -it redis-ha-server-0 -c redis -- sh
/data $ redis-cli
127.0.0.1:6379> info Replication
# Replication
role:master
connected_slaves:1
min_slaves_good_slaves:1
slave0:ip=10.101.152.37,port=6379,state=online,offset=25236,lag=0
master_replid:4baa1ff233604d357f9a5b2d6b100f90192d309c
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:25236
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:25236
使用info Replication来查看主机信息，可以看出redis-ha-server-0为master主机。
执行SHUTDOWN命令来使主机宕机：
127.0.0.1:6379> SHUTDOWN
command terminated with exit code 137
宕机后控制器会马上帮我们启动一个pod
[aaa@qq.com redis-ha]# kubectl get pod -n redis
NAME READY STATUS RESTARTS AGE
redis-ha-server-0 2/2 Running 1 4m57s
redis-ha-server-1 2/2 Running 0 4m48s
等重建好了之后再次查看pod角色
[aaa@qq.com redis-ha]# kubectl -n redis exec -it redis-ha-server-0 -c redis -- sh
/data $ redis-cli
127.0.0.1:6379> info Replication
# Replication
role:slave //宕机后重启切换为slave
master_host:10.101.152.37
master_port:6379
...
可以看出redis-ha-server-0已经变成了从机，而主机的ip为10.101.152.37，这时就实现了主从故障的切换。
主从故障切换依靠 sentinel容器完成。
（4）可以直接更改变量文件中的副本数量实现集群的动态拉伸（5）使用upgrade选项更新集群
[aaa@qq.com redis-ha]# helm -n redis upgrade redis-ha .
更新后查看pod可以看出已经变成了3个
[aaa@qq.com redis-ha]# kubectl -n redis get pod
NAME READY STATUS RESTARTS AGE
redis-ha-server-0 2/2 Running 2 20m
redis-ha-server-1 2/2 Running 0 20m
redis-ha-server-2 0/2 Pending 0 12s
也可以使用以下命令查看更新历史以便回滚
[aaa@qq.com redis-ha]# helm -n redis history
Error: "helm history" requires 1 argument
Usage: helm history RELEASE_NAME [flags]
[aaa@qq.com redis-ha]# helm -n redis history redis-ha
REVISION UPDATED STATUS CHART APP VERSION DESCRIPTION
1 Sun May 10 23:15:35 2020 superseded redis-ha-4.4.4 5.0.6 Install complete
2 Sun May 10 23:35:35 2020 deployed redis-ha-4.4.4 5.0.6 Upgrade complete