Cassandra是一款非常优秀的开源的分布式NoSQL数据库,被许多优秀的大公司采用,具有高可用、弹性扩展、性能好等特点。
正应Cassandra的优势,我们经常需要在云上服务使用,则需要部署Cassandra到K8s上,这就有了K8ssandra。K8ssandra不仅帮助我们可以快速可靠地在Kubernetes上部署Cassandra,同时提供了许多组件,如监控、备份、同步、访问等。而这些都是一个Production-Ready的产品不可或缺的。
K8ssandra的组件架构图如下:
如何在Ubuntu上通过Minikube快速启动一个Kubernetes,在文章《服务网格Istio入门-详细记录Kubernetes安装Istio并使用》已经有详细的介绍,这里不再赘述。为了更好的兼容性,我们指定了Kubernetes的版本,命令如下:
minikube start --driver=none --kubernetes-version=v1.19.13
因为要用到PVC,我们创建一个StorageClass:
$ kubectl apply -f https://raw.githubusercontent.com/rancher/local-path-provisioner/master/deploy/local-path-storage.yaml
我们需要使用Helm来部署K8ssandra,下载Helm3如下:
# 下载安装包 curl -LO https://get.helm.sh/helm-v3.7.0-linux-amd64.tar.gz # 解压 tar -zxvf helm-v3.7.0-linux-amd64.tar.gz # 移动到bin目录 mv linux-amd64/helm /usr/local/bin/helm
添加Helm的仓库:
helm repo add k8ssandra https://helm.k8ssandra.io/stable $ helm repo list NAME URL k8ssandra https://helm.k8ssandra.io/stable traefik https://helm.traefik.io/traefik
查找一下K8ssandra相关的包:
$ helm search repo k8ssandra NAME CHART VERSION APP VERSION DESCRIPTION k8ssandra/k8ssandra 1.3.1 Provisions and configures an instance of the en... k8ssandra/k8ssandra-common 0.28.4 Helper library containing functions used by man... k8ssandra/k8ssandra-operator 0.31.0 1.0.0 Kubernetes operator which handles the provision... k8ssandra/backup 0.26.0 Creates a CassandraBackup custom resource insta... k8ssandra/cass-operator 0.31.0 1.8.0 Kubernetes operator which handles the provision... k8ssandra/medusa-operator 0.30.1 0.1.0 Installs and configures the Medusa Operator for... k8ssandra/reaper-operator 0.32.1 0.1.0 Configures and installs the Reaper Operator for... k8ssandra/restore 0.27.1 Creates a CassandraRestore custom resource inst...
我们安装k8ssandra/k8ssandra就可以了。
Helm是一个Chart+Value的管理方式,我们准备一个yaml文件(k8ssandra-values.yaml)来放一些变量:
cassandra: version: "4.0.0" cassandraLibDirVolume: storageClass: local-path size: 5Gi allowMultipleNodesPerWorker: true heap: size: 1G newGenSize: 1G resources: requests: cpu: 1000m memory: 2Gi limits: cpu: 1000m memory: 2Gi datacenters: - name: dc1 size: 1 racks: - name: default kube-prometheus-stack: grafana: adminUser: admin adminPassword: admin123 stargate: enabled: true replicas: 1 heapMB: 256 cpuReqMillicores: 200 cpuLimMillicores: 1000
安装K8ssandra:
$ helm install -f k8ssandra-values.yaml k8ssandra k8ssandra/k8ssandra NAME: k8ssandra LAST DEPLOYED: Thu Sep 30 21:20:49 2021 NAMESPACE: default STATUS: deployed REVISION: 1
做一些必要的检查如下:
$ helm list NAME NAMESPACE REVISION UPDATED STATUS CHART APP VERSION k8ssandra default 1 2021-09-30 21:20:49.409672869 +0800 CST deployed k8ssandra-1.3.1 $ kubectl get cassandradatacenters NAME AGE dc1 4m34s $ kubectl describe CassandraDataCenter dc1 | grep "Cassandra Operator Progress:" Cassandra Operator Progress: Ready
检查一下Pod和Service:
获取K8ssandra超级用户的用户名和密码
$ kubectl get secret k8ssandra-superuser -o jsonpath="{.data.username}" | base64 --decode ; echo k8ssandra-superuser $ kubectl get secret k8ssandra-superuser -o jsonpath="{.data.password}" | base64 --decode ; echo TNE5xOk45C1aQsj29qxw
我们为了高可用和容量,创建更多的Cassandra Node,直接修改k8ssandra-values.yaml如下:
cassandra: version: "4.0.0" cassandraLibDirVolume: storageClass: local-path size: 5Gi allowMultipleNodesPerWorker: true heap: size: 1G newGenSize: 1G resources: requests: cpu: 1000m memory: 2Gi limits: cpu: 1000m memory: 2Gi datacenters: - name: dc1 size: 3 racks: - name: racks1 - name: racks2 - name: racks3 kube-prometheus-stack: grafana: adminUser: admin adminPassword: admin123 stargate: enabled: true replicas: 1 heapMB: 256 cpuReqMillicores: 200 cpuLimMillicores: 1000
修改完之后,升级配置:
$ helm upgrade -f k8ssandra-values.yaml k8ssandra k8ssandra/k8ssandra Release "k8ssandra" has been upgraded. Happy Helming! NAME: k8ssandra LAST DEPLOYED: Fri Oct 1 00:40:08 2021 NAMESPACE: default STATUS: deployed REVISION: 2
查看Kubernetes的相关资源:
我们对外暴露Grafana服务来看看:
kubectl expose deployment k8ssandra-grafana --type=NodePort --name=grafana-out
找到对应的NodePort的端口30348,访问:http://外网IP:30348
账号:admin/admin123
界面如下,提供良好的监控界面:
K8ssandra真是一个不错的开源项目,后续再介绍如何在开发中通过K8ssandra使用Cassandra吧。
代码请查看:https://github.com/LarryDpk/pkslow-samples