K8S部署Redis Cluster集群（三主三从模式） - 部署笔记

一、Redis 介绍

• Redis代表REmote DIctionary Server是一种开源的内存中数据存储，通常用作数据库，缓存或消息代理。它可以存储和操作高级数据类型，例如列表，地图，集合和排序集合。
• 由于Redis接受多种格式的密钥，因此可以在服务器上执行操作，从而减少了客户端的工作量。
• 它仅将磁盘用于持久性，而将数据完全保存在内存中。
• Redis是一种流行的数据存储解决方案，并被GitHub，Pinterest，Snapchat，Twitter，StackOverflow，Flickr等技术巨头所使用。

  二、为什么要用Redis

• 它的速度非常快。它是用ANSI C编写的，并且可以在POSIX系统上运行，例如Linux，Mac OS X和Solaris。
• Redis通常被排名为最流行的键/值数据库和最流行的与容器一起使用的NoSQL数据库。
• 其缓存解决方案减少了对云数据库后端的调用次数。
• 应用程序可以通过其客户端API库对其进行访问。
• 所有流行的编程语言都支持Redis。
• 它是开源且稳定的。

  三、什么是Redis Cluster集群

• Redis Cluster是一组Redis实例，旨在通过对数据库进行分区来扩展数据库，从而使其更具弹性。
• 群集中的每个成员（无论是主副本还是辅助副本）都管理哈希槽的子集。如果主机无法访问，则其从机将升级为主机。在由三个主节点组成的最小Redis群集中，每个主节点都有一个从节点（以实现最小的故障转移），每个主节点都分配有一个介于0到16,383之间的哈希槽范围。节点A包含从0到5000的哈希槽，节点B从5001到10000，节点C从10001到16383。
• 群集内部的通信是通过内部总线进行的，使用协议传播有关群集的信息或发现新节点。

  四、在Kubernetes中部署Redis Cluster集群过程记录 在Kubernetes中部署Redis集群面临挑战，因为每个Redis实例都依赖于一个配置文件，该文件可以跟踪其他集群实例及其角色。为此，我们需要结合使用StatefulSets控制器和PersistentVolumes持久化存储。   StatefulSet的设计原理模型:

• 拓扑状态：

应用的多个实例之间不是完全对等的关系,这个应用实例的启动必须按照某些顺序启动,比如应用的主节点 A 要先于从节点 B 启动。而如果你把 A 和 B 两个Pod删除掉,他们再次被创建出来是也必须严格按照这个顺序才行,并且,新创建出来的Pod,必须和原来的Pod的网络标识一样,这样原先的访问者才能使用同样的方法,访问到这个新的Pod  

• 存储状态：

应用的多个实例分别绑定了不同的存储数据.对于这些应用实例来说,Pod A第一次读取到的数据,和隔了十分钟之后再次读取到的数据,应该是同一份,哪怕在此期间Pod A被重新创建过.一个数据库应用的多个存储实例。   存储卷 了解statefulset状态后，应该知道要为数据准备一个存储卷了，创建方式有静态方式和动态方式，静态方式就是手动创建PV、PVC，然后POD进行进行调用即可。这里使用动态NFS作为挂载卷，需要部署NFS动态StorageClass   1、使用NFS配置StatefulSet的动态持久化存储 1）在NFS服务器端（172.16.60.238）通过nfs创建Redis Cluster集群的共享目录

1	[root@k8s-harbor01 ~]# mkdir -p /data/storage/k8s/redis

2）创建nfs的rbac

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

[root@k8s-master01 ~]# mkdir -p /opt/k8s/k8s_project/redis [root@k8s-master01 ~]# cd /opt/k8s/k8s_project/redis [root@k8s-master01 redis]# vim nfs-rbac.yaml --- apiVersion: v1 kind: ServiceAccount metadata:   name: nfs-provisioner   namespace: wiseco --- kind: ClusterRole apiVersion: rbac.authorization.k8s.io/v1 metadata:    name: nfs-provisioner-runner    namespace: wiseco rules:    -  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: ["watch", "create", "update", "patch"]    -  apiGroups: [""]       resources: ["services", "endpoints"]       verbs: ["get","create","list", "watch","update"]    -  apiGroups: ["extensions"]       resources: ["podsecuritypolicies"]       resourceNames: ["nfs-provisioner"]       verbs: ["use"] --- kind: ClusterRoleBinding apiVersion: rbac.authorization.k8s.io/v1 metadata:   name: run-nfs-provisioner subjects:   - kind: ServiceAccount     name: nfs-provisioner     namespace: wiseco roleRef:   kind: ClusterRole   name: nfs-provisioner-runner   apiGroup: rbac.authorization.k8s.io

创建并查看

1 2 3 4 5 6 7 8 9 10 11

[root@k8s-master01 redis]# kubectl apply -f nfs-rbac.yaml serviceaccount/nfs-provisioner created clusterrole.rbac.authorization.k8s.io/nfs-provisioner-runner created clusterrolebinding.rbac.authorization.k8s.io/run-nfs-provisioner created   [root@k8s-master01 redis]# kubectl get sa -n wiseco|grep nfs nfs-provisioner                1         24s [root@k8s-master01 redis]# kubectl get clusterrole -n wiseco|grep nfs nfs-provisioner-runner                                                 2021-02-04T02:21:11Z [root@k8s-master01 redis]# kubectl get clusterrolebinding -n wiseco|grep nfs run-nfs-provisioner                                    ClusterRole/nfs-provisioner-runner                                                 34s

3）创建redis cluster集群的storageclass

1 2 3 4 5 6 7 8 9 10 11 12

[root@k8s-master01 redis]# ll total 4 -rw-r--r-- 1 root root 1216 Feb  4 10:20 nfs-rbac.yaml   [root@k8s-master01 redis]# vim redis-nfs-class.yaml apiVersion: storage.k8s.io/v1beta1 kind: StorageClass metadata:   name: redis-nfs-storage   namespace: wiseco provisioner: redis/nfs reclaimPolicy: Retain

创建并查看

1 2 3 4 5 6

[root@k8s-master01 redis]# kubectl apply -f redis-nfs-class.yaml storageclass.storage.k8s.io/redis-nfs-storage created   [root@k8s-master01 redis]# kubectl get sc -n wiseco NAME                PROVISIONER   RECLAIMPOLICY   VOLUMEBINDINGMODE   ALLOWVOLUMEEXPANSION   AGE redis-nfs-storage   redis/nfs     Retain          Immediate           false

4）创建redis cluster集群的nfs-client-provisioner

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

[root@k8s-master01 redis]# ll total 8 -rw-r--r-- 1 root root 1216 Feb  4 10:20 nfs-rbac.yaml -rw-r--r-- 1 root root  155 Feb  4 10:24 redis-nfs-class.yaml   [root@k8s-master01 redis]# vim redis-nfs.yml apiVersion: apps/v1 kind: Deployment metadata:   name: redis-nfs-client-provisioner   namespace: wiseco spec:   replicas: 1   selector:     matchLabels:       app: redis-nfs-client-provisioner   strategy:     type: Recreate   template:     metadata:       labels:         app: redis-nfs-client-provisioner     spec:       serviceAccount: nfs-provisioner       containers:         - name: redis-nfs-client-provisioner           image: registry.cn-hangzhou.aliyuncs.com/open-ali/nfs-client-provisioner           imagePullPolicy: IfNotPresent           volumeMounts:             - name: nfs-client-root               mountPath:  /persistentvolumes           env:             - name: PROVISIONER_NAME               value: redis/nfs             - name: NFS_SERVER               value: 172.16.60.238             - name: NFS_PATH               value: /data/storage/k8s/redis       volumes:         - name: nfs-client-root           nfs:             server: 172.16.60.238             path: /data/storage/k8s/redis

创建并查看

1 2 3 4 5

[root@k8s-master01 redis]# kubectl apply -f redis-nfs.yml deployment.apps/redis-nfs-client-provisioner created   [root@k8s-master01 redis]# kubectl get pods -n wiseco|grep nfs redis-nfs-client-provisioner-58b46549dd-h87gg   1/1     Running   0          40s

2、部署Redis Cluster集群
本案例部署采用的namespace命名空间是wiseco

1）准备image镜像
redis-trib.rb工具可以去redis源码中拷贝一个到当前目录,然后构建镜像。

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

[root@k8s-master01 redis]# pwd /opt/k8s/k8s_project/redis [root@k8s-master01 redis]# ll total 12 -rw-r--r-- 1 root root 1216 Feb  4 15:31 nfs-rbac.yaml -rw-r--r-- 1 root root  155 Feb  4 15:32 redis-nfs-class.yaml -rw-r--r-- 1 root root 1006 Feb  4 15:32 redis-nfs.yml   [root@k8s-master01 redis]# mkdir image [root@k8s-master01 redis]# cd image [root@k8s-master01 image]# ll total 64 -rw-r--r-- 1 root root   191 Feb  4 18:14 Dockerfile -rwxr-xr-x 1 root root 60578 Feb  4 15:49 redis-trib.rb   [root@k8s-master01 image]# cat Dockerfile FROM redis:4.0.11 RUN apt-get update -y RUN apt-get install -y  ruby \ rubygems RUN apt-get clean all RUN gem install redis RUN apt-get install dnsutils -y COPY redis-trib.rb /usr/local/bin/

创建镜像并上传到Harbor仓库

1 2	[root@k8s-master01 image]# docker build -t 172.16.60.238/wiseco/redis:4.0.11 . [root@k8s-master01 image]# docker push 172.16.60.238/wiseco/redis:4.0.11

2）创建configmap redis配置文件使用configmap方式进行挂载，如果将配置封装到docker image中的话，俺么每次修改配置就需要重新docker build。个人觉得比较麻烦，所以使用configmap方式挂载配置。

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81

[root@k8s-master01 redis]# pwd /opt/k8s/k8s_project/redis [root@k8s-master01 redis]# ll total 12 drwxr-xr-x 2 root root   45 Feb  4 18:14 image -rw-r--r-- 1 root root 1216 Feb  4 15:31 nfs-rbac.yaml -rw-r--r-- 1 root root  155 Feb  4 15:32 redis-nfs-class.yaml -rw-r--r-- 1 root root 1006 Feb  4 15:32 redis-nfs.yml   [root@k8s-master01 redis]# mkdir conf [root@k8s-master01 redis]# cd conf/   [root@k8s-master01 conf]# vim redis-configmap.yaml apiVersion: v1 kind: ConfigMap metadata:   name: redis-cluster   namespace: wiseco data:   fix-ip.sh: |     #!/bin/sh     CLUSTER_CONFIG="/data/nodes.conf"     if [ -f ${CLUSTER_CONFIG} ]; then       if [ -z "${POD_IP}" ]; then         echo "Unable to determine Pod IP address!"         exit 1       fi       echo "Updating my IP to ${POD_IP} in ${CLUSTER_CONFIG}"       sed -i.bak -e '/myself/ s/[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}/'${POD_IP}'/' ${CLUSTER_CONFIG}     fi     exec "$@"   redis.conf: |     cluster-enabled yes     cluster-config-file /data/nodes.conf     cluster-node-timeout 10000     protected-mode no     daemonize no     pidfile /var/run/redis.pid     port 6379     tcp-backlog 511     bind 0.0.0.0     timeout 3600     tcp-keepalive 1     loglevel verbose     logfile /data/redis.log     databases 16     save 900 1     save 300 10     save 60 10000     stop-writes-on-bgsave-error yes     rdbcompression yes     rdbchecksum yes     dbfilename dump.rdb     dir /data     #requirepass yl123456     appendonly yes     appendfilename "appendonly.aof"     appendfsync everysec     no-appendfsync-on-rewrite no     auto-aof-rewrite-percentage 100     auto-aof-rewrite-min-size 64mb     lua-time-limit 20000     slowlog-log-slower-than 10000     slowlog-max-len 128     #rename-command FLUSHALL  ""     latency-monitor-threshold 0     notify-keyspace-events ""     hash-max-ziplist-entries 512     hash-max-ziplist-value 64     list-max-ziplist-entries 512     list-max-ziplist-value 64     set-max-intset-entries 512     zset-max-ziplist-entries 128     zset-max-ziplist-value 64     hll-sparse-max-bytes 3000     activerehashing yes     client-output-buffer-limit normal 0 0 0     client-output-buffer-limit slave 256mb 64mb 60     client-output-buffer-limit pubsub 32mb 8mb 60     hz 10     aof-rewrite-incremental-fsync yes

需要注意：fix-ip.sh 脚本的作用用于当redis集群某pod重建后Pod IP发生变化，在/data/nodes.conf中将新的Pod IP替换原Pod IP。不然集群会出问题。   创建和查看

1 2 3 4

[root@k8s-master01 conf]# kubectl apply -f redis-configmap.yaml   [root@k8s-master01 conf]# kubectl get cm -n wiseco|grep redis redis-cluster                     2      8m55s

2）准备StatefulSet volumeClaimTemplates 用于StatefulSet控制器场景：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84

[root@k8s-master01 redis]# pwd /opt/k8s/k8s_project/redis [root@k8s-master01 redis]# ll total 12 drwxr-xr-x 2 root root   34 Feb  4 18:52 conf drwxr-xr-x 2 root root   45 Feb  4 18:14 image -rw-r--r-- 1 root root 1216 Feb  4 15:31 nfs-rbac.yaml -rw-r--r-- 1 root root  155 Feb  4 15:32 redis-nfs-class.yaml -rw-r--r-- 1 root root 1006 Feb  4 15:32 redis-nfs.yml   [root@k8s-master01 redis]# mkdir deploy [root@k8s-master01 redis]# cd deploy/ [root@k8s-master01 deploy]# cat redis-cluster.yml --- apiVersion: v1 kind: Service metadata:   namespace: wiseco   name: redis-cluster spec:   clusterIP: None   ports:   - port: 6379     targetPort: 6379     name: client   - port: 16379     targetPort: 16379     name: gossip   selector:     app: redis-cluster --- apiVersion: apps/v1 kind: StatefulSet metadata:   namespace: wiseco   name: redis-cluster spec:   serviceName: redis-cluster   replicas: 6   selector:     matchLabels:       app: redis-cluster   template:     metadata:       labels:         app: redis-cluster     spec:       containers:       - name: redis         image: 172.16.60.238/wiseco/redis:4.0.11         ports:         - containerPort: 6379           name: client         - containerPort: 16379           name: gossip         command: ["/etc/redis/fix-ip.sh", "redis-server", "/etc/redis/redis.conf"]         env:         - name: POD_IP           valueFrom:             fieldRef:               fieldPath: status.podIP         volumeMounts:         - name: conf           mountPath: /etc/redis/           readOnly: false         - name: data           mountPath: /data           readOnly: false       volumes:       - name: conf         configMap:           name: redis-cluster           defaultMode: 0755   volumeClaimTemplates:   - metadata:       name: data       annotations:         volume.beta.kubernetes.io/storage-class: "redis-nfs-storage"     spec:       accessModes:         - ReadWriteMany       resources:         requests:           storage: 10Gi

创建并查看

1 2 3 4 5 6 7 8 9 10 11 12

[root@k8s-master01 deploy]# kubectl apply -f redis-cluster.yml   [root@k8s-master01 deploy]# kubectl get pods -n wiseco|grep redis-cluster redis-cluster-0                                 1/1     Running   0          10m redis-cluster-1                                 1/1     Running   0          10m redis-cluster-2                                 1/1     Running   0          10m redis-cluster-3                                 1/1     Running   0          10m redis-cluster-4                                 1/1     Running   0          9m35s redis-cluster-5                                 1/1     Running   0          9m25s   [root@k8s-master01 deploy]# kubectl get svc -n wiseco|grep redis-cluster redis-cluster    ClusterIP   None                     6379/TCP,16379/TCP           10m

查看PV、PVC

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

[root@k8s-master01 deploy]# kubectl get pv NAME                                       CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS        CLAIM                         STORAGECLASS        REASON   AGE pvc-20bcb3be-90e1-4354-bd11-4f442a3bd562   10Gi       RWX            Delete           Bound         wiseco/data-redis-cluster-0   redis-nfs-storage            19m pvc-3b53a31b-9a53-4bd4-93ff-2cf9fed551de   10Gi       RWX            Delete           Bound         wiseco/data-redis-cluster-2   redis-nfs-storage            12m pvc-43c0cba2-54a9-4416-afb6-8b7730a199dc   10Gi       RWX            Delete           Bound         wiseco/data-redis-cluster-1   redis-nfs-storage            12m pvc-66daade5-1b97-41ce-a9e0-4cf88d63894d   10Gi       RWX            Delete           Terminating   wiseco/data-redis-cluster-5   redis-nfs-storage            11m pvc-dd62a086-1802-446a-9f9d-35620f7f0b4a   10Gi       RWX            Delete           Bound         wiseco/data-redis-cluster-4   redis-nfs-storage            11m pvc-e5aa9802-b983-471c-a7da-32eebc497610   10Gi       RWX            Delete           Bound         wiseco/data-redis-cluster-3   redis-nfs-storage            12m   [root@k8s-master01 deploy]# kubectl get pvc -n wiseco NAME                   STATUS   VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS        AGE data-redis-cluster-0   Bound    pvc-20bcb3be-90e1-4354-bd11-4f442a3bd562   10Gi       RWX            redis-nfs-storage   19m data-redis-cluster-1   Bound    pvc-43c0cba2-54a9-4416-afb6-8b7730a199dc   10Gi       RWX            redis-nfs-storage   12m data-redis-cluster-2   Bound    pvc-3b53a31b-9a53-4bd4-93ff-2cf9fed551de   10Gi       RWX            redis-nfs-storage   12m data-redis-cluster-3   Bound    pvc-e5aa9802-b983-471c-a7da-32eebc497610   10Gi       RWX            redis-nfs-storage   12m data-redis-cluster-4   Bound    pvc-dd62a086-1802-446a-9f9d-35620f7f0b4a   10Gi       RWX            redis-nfs-storage   11m data-redis-cluster-5   Bound    pvc-66daade5-1b97-41ce-a9e0-4cf88d63894d   10Gi       RWX            redis-nfs-storage   11m

3）查看NFS共享存储 NFS服务器（172.16.60.238），查看共享目录/data/storage/k8s/redis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

[root@k8s-harbor01 redis]# pwd /data/storage/k8s/redis [root@k8s-harbor01 redis]# ll total 0 drwxrwxrwx 2 root root 63 Feb  4 18:59 wiseco-data-redis-cluster-0-pvc-20bcb3be-90e1-4354-bd11-4f442a3bd562 drwxrwxrwx 2 root root 63 Feb  4 18:59 wiseco-data-redis-cluster-1-pvc-43c0cba2-54a9-4416-afb6-8b7730a199dc drwxrwxrwx 2 root root 63 Feb  4 18:59 wiseco-data-redis-cluster-2-pvc-3b53a31b-9a53-4bd4-93ff-2cf9fed551de drwxrwxrwx 2 root root 63 Feb  4 19:00 wiseco-data-redis-cluster-3-pvc-e5aa9802-b983-471c-a7da-32eebc497610 drwxrwxrwx 2 root root 63 Feb  4 19:00 wiseco-data-redis-cluster-4-pvc-dd62a086-1802-446a-9f9d-35620f7f0b4a drwxrwxrwx 2 root root 63 Feb  4 19:00 wiseco-data-redis-cluster-5-pvc-66daade5-1b97-41ce-a9e0-4cf88d63894d [root@k8s-harbor01 redis]# ls ./* ./wiseco-data-redis-cluster-0-pvc-20bcb3be-90e1-4354-bd11-4f442a3bd562: appendonly.aof  nodes.conf  redis.log   ./wiseco-data-redis-cluster-1-pvc-43c0cba2-54a9-4416-afb6-8b7730a199dc: appendonly.aof  nodes.conf  redis.log   ./wiseco-data-redis-cluster-2-pvc-3b53a31b-9a53-4bd4-93ff-2cf9fed551de: appendonly.aof  nodes.conf  redis.log   ./wiseco-data-redis-cluster-3-pvc-e5aa9802-b983-471c-a7da-32eebc497610: appendonly.aof  nodes.conf  redis.log   ./wiseco-data-redis-cluster-4-pvc-dd62a086-1802-446a-9f9d-35620f7f0b4a: appendonly.aof  nodes.conf  redis.log   ./wiseco-data-redis-cluster-5-pvc-66daade5-1b97-41ce-a9e0-4cf88d63894d: appendonly.aof  nodes.conf  redis.log

3、初始化Redis Cluster集群
接下来是形成Redis Cluster集群，运行以下命令并键入yes以接受配置。
集群形式：前三个节点成为主节点，后三个节点成为从节点。

需要注意：
redis-trib.rb必须使用ip进行初始化redis集群，使用域名会报如下错误：*******/redis/client.rb:126:in `call’: ERR Invalid node address specified: redis-cluster-0.redis-headless.sts-app.svc.cluster.local:6379 (Redis::CommandError)

这里进行Redis Cluster集群初始化的命令：
以下命令并键入yes以接受配置。前三个节点成为主节点，后三个节点成为从节点。
kubectl exec -it redis-cluster-0 -n wiseco -- redis-trib.rb create --replicas 1 $(kubectl get pods -l app=redis-cluster -n wiseco -o jsonpath='{range.items[*]}{.status.podIP}:6379 ')

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

先获取Redis Cluster集群的6个节点Pod的ip地址 [root@k8s-master01 redis]# kubectl get pods -n wiseco -o wide|grep redis-cluster redis-cluster-0                                 1/1     Running   0          4h34m   172.30.217.83    k8s-node04              redis-cluster-1                                 1/1     Running   0          4h34m   172.30.85.217    k8s-node01              redis-cluster-2                                 1/1     Running   0          4h34m   172.30.135.181   k8s-node03              redis-cluster-3                                 1/1     Running   0          4h34m   172.30.58.251    k8s-node02              redis-cluster-4                                 1/1     Running   0          4h33m   172.30.85.216    k8s-node01              redis-cluster-5                                 1/1     Running   0          4h33m   172.30.217.82    k8s-node04                  [root@k8s-master01 redis]# kubectl get pods -l app=redis-cluster -n wiseco -o jsonpath='{range.items[*]}{.status.podIP}:6379 ' 172.30.217.83:6379 172.30.85.217:6379 172.30.135.181:6379 172.30.58.251:6379 172.30.85.216:6379 172.30.217.82:6379   这里特别注意一下： 上面命令最后一个单引号前面一定要有空格！！ 因为接下来进行Redis Cluster集群初始化的时候，集群节点间的ip+port之间要通过空格隔开。   [root@k8s-master01 redis]# kubectl exec -it redis-cluster-0 -n wiseco -- redis-trib.rb create --replicas 1 $(kubectl get pods -l app=redis-cluster -n wiseco -o jsonpath='{range.items[*]}{.status.podIP}:6379 ') >>> Creating cluster >>> Performing hash slots allocation on 6 nodes... Using 3 masters: 172.30.217.83:6379 172.30.85.217:6379 172.30.135.181:6379 Adding replica 172.30.58.251:6379 to 172.30.217.83:6379 Adding replica 172.30.85.216:6379 to 172.30.85.217:6379 Adding replica 172.30.217.82:6379 to 172.30.135.181:6379 M: e5a3154a17131075f35fb32953b8cf8d6cfc7df0 172.30.217.83:6379    slots:0-5460 (5461 slots) master M: 961398483262f505a115957e7e4eda7ff3e64900 172.30.85.217:6379    slots:5461-10922 (5462 slots) master M: 2d1440e37ea4f4e9f6d39d240367deaa609d324d 172.30.135.181:6379    slots:10923-16383 (5461 slots) master S: 0d7bf40bf18d474509116437959b65551cd68b03 172.30.58.251:6379    replicates e5a3154a17131075f35fb32953b8cf8d6cfc7df0 S: 8cbf699a850c0dafe51524127a594fdbf0a27784 172.30.85.216:6379    replicates 961398483262f505a115957e7e4eda7ff3e64900 S: 2987a33f4ce2e412dcc11c1c1daa2538591cd930 172.30.217.82:6379    replicates 2d1440e37ea4f4e9f6d39d240367deaa609d324d Can I set the above configuration? (type 'yes' to accept): yes >>> Nodes configuration updated >>> Assign a different config epoch to each node >>> Sending CLUSTER MEET messages to join the cluster Waiting for the cluster to join...... >>> Performing Cluster Check (using node 172.30.217.83:6379) M: e5a3154a17131075f35fb32953b8cf8d6cfc7df0 172.30.217.83:6379    slots:0-5460 (5461 slots) master M: 961398483262f505a115957e7e4eda7ff3e64900 172.30.85.217:6379    slots:5461-10922 (5462 slots) master M: 2d1440e37ea4f4e9f6d39d240367deaa609d324d 172.30.135.181:6379    slots:10923-16383 (5461 slots) master M: 0d7bf40bf18d474509116437959b65551cd68b03 172.30.58.251:6379    slots: (0 slots) master    replicates e5a3154a17131075f35fb32953b8cf8d6cfc7df0 M: 8cbf699a850c0dafe51524127a594fdbf0a27784 172.30.85.216:6379    slots: (0 slots) master    replicates 961398483262f505a115957e7e4eda7ff3e64900 M: 2987a33f4ce2e412dcc11c1c1daa2538591cd930 172.30.217.82:6379    slots: (0 slots) master    replicates 2d1440e37ea4f4e9f6d39d240367deaa609d324d [OK] All nodes agree about slots configuration. >>> Check for open slots... >>> Check slots coverage... [OK] All 16384 slots covered.

通过上面初始化信息，可以看出集群关系： redis-cluster-0是master节点，redis-cluster-3是它的从节点。 redis-cluster-1是master节点，redis-cluster-4是它的从节点。 redis-cluster-2是master节点，redis-cluster-5是它的从节点。   4、验证Redis Cluster集群部署

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

[root@k8s-master01 redis]# kubectl exec -it redis-cluster-0 -n wiseco -- redis-cli cluster info                                        cluster_state:ok cluster_slots_assigned:16384 cluster_slots_ok:16384 cluster_slots_pfail:0 cluster_slots_fail:0 cluster_known_nodes:6 cluster_size:3 cluster_current_epoch:6 cluster_my_epoch:1 cluster_stats_messages_ping_sent:130 cluster_stats_messages_pong_sent:137 cluster_stats_messages_sent:267 cluster_stats_messages_ping_received:132 cluster_stats_messages_pong_received:130 cluster_stats_messages_meet_received:5 cluster_stats_messages_received:267   [root@k8s-master01 redis]# for x in $(seq 0 5); do echo "redis-cluster-$x"; kubectl exec redis-cluster-$x -n wiseco -- redis-cli role; echo; done redis-cluster-0 master 168 172.30.58.251 6379 168   redis-cluster-1 master 168 172.30.85.216 6379 168   redis-cluster-2 master 182 172.30.217.82 6379 168   redis-cluster-3 slave 172.30.217.83 6379 connected 182   redis-cluster-4 slave 172.30.85.217 6379 connected 168   redis-cluster-5 slave 172.30.135.181 6379 connected 182