K8S搭建redis集群

IT那活儿 2023-11-28

444

点击上方“IT那活儿”公众号--专注于企业全栈运维技术分享，不管IT什么活儿，干就完了！！！

本次演示为本地测试环境，k8s集群为单节点一主两从集群；

在生产环境需要按照需求搭建。

主机规划

安装步骤

2.1 安装nfs服务

默认以安装好k8s集群就不做叙述，需要准备nfs单机环境，若k8s为多master集群，可以准备nfs集群环境。

1）在master上安装nfs服务

注：设置开机自启动。

[root@nfs ~]# yum install nfs-utils -y
[root@nfs ~]# systemctl restart nfs
[root@nfs ~]# systemctl enable nfs

2）在node上安装nfs服务

注：不需要启动。

[root@k8s-master01 ~]# yum install nfs-utils -y

2.2 准备共享目录

将目录以读写权限暴露给192.168.158.0/24网段中的所有主机（在生产环境应该只将共享目录暴露给集群所在机器）；

必须暴漏共享目录，否则在创建pod时会出现报错无法找到pv！

1）创建共享目录

for x in $(seq 1 6);\
> do \
> mkdir -p data/redis-cluster/pv${x}
> done

2）将共享目录暴露

暴露给指定主机将“192.168.158.0/24”改成主机ip即可。

[root@master ~]# vim /etc/exports
/data/redis-cluster/pv1 192.168.158.0/24(rw,no_root_squash)
/data/redis-cluster/pv2 192.168.158.0/24(rw,no_root_squash)
/data/redis-cluster/pv3 192.168.158.0/24(rw,no_root_squash)
/data/redis-cluster/pv4 192.168.158.0/24(rw,no_root_squash)
/data/redis-cluster/pv5 192.168.158.0/24(rw,no_root_squash)
/data/redis-cluster/pv6 192.168.158.0/24(rw,no_root_squash)

2.3 创建6个pv，kubectl apply -f redis-pv.yaml

下面为redis-pv.yaml内容：

apiVersion: v1
kind: PersistentVolume # 创建pv
metadata:
 name: redis-pv1 # 名称
spec:
 capacity:
   storage: 3Gi # 3G磁盘空间
 accessModes: 
 - ReadWriteOnce # 权限读写
 persistentVolumeReclaimPolicy: Recycle # 回收策略，清除数据
 storageClassName: "redis-cluster" # 存储类别为“redis-cluster”，只能允许存储类别相同的pvc使用
 nfs:
   path: data/redis-cluster/pv1
   server: 192.168.158.136
# ...

中间省略一直到redis-pv6 修改metadata:name 和 spec:nfs:path即可。

---
apiVersion: v1
kind: PersistentVolume
metadata:
 name: redis-pv6
spec:
 capacity:
   storage: 3Gi
 accessModes:
 - ReadWriteOnce
 persistentVolumeReclaimPolicy: Recycle
 storageClassName: "redis-cluster" 
 nfs:
   path: data/redis-cluster/pv6
   server: 192.168.158.136

[root@master ~]# kubectl apply -f redis-pv.yaml

创建成功。

可以通过kubectl命令看到下列6个pv：

[root@master ~]# kubectl get pv 
NAME CAPACITY   ACCESS MODES RECLAIM POLICY STATUS CLAIM                          STORAGECLASS REASON AGE
redis-pv1 3Gi RWO Recycle Bound default/data-redis-cluster-0   redis-cluster 115m
redis-pv2 3Gi RWO Recycle Bound default/data-redis-cluster-2   redis-cluster 115m
redis-pv3 3Gi RWO Recycle Bound default/data-redis-cluster-3   redis-cluster 115m
redis-pv4 3Gi RWO Recycle Bound default/data-redis-cluster-4   redis-cluster 115m
redis-pv5 3Gi RWO Recycle Bound default/data-redis-cluster-1   redis-cluster 115m
redis-pv6 3Gi RWO Recycle Bound default/data-redis-cluster-5   redis-cluster 115m

2.4 创建ConfigMap和StatefulSet

---
apiVersion: v1
kind: ConfigMap
metadata:
 name: redis-cluster
data:
 update-node.sh: |
   #!/bin/sh
   REDIS_NODES="/data/nodes.conf"
   sed -i -e "/myself/ s/[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}/${POD_IP}/" ${REDIS_NODES}
   exec "$@"
 redis.conf: |+
   cluster-enabled yes
   cluster-require-full-coverage no
   cluster-node-timeout 15000
   cluster-config-file /data/nodes
   cluster-migration-barrier 1
   Maxmemory 3GB
   port 6379
   appendonly yes
   protected-mode no
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
 name: redis-cluster
spec:
 serviceName: redis-cluster
 replicas: 6
 selector: # 标签选择器
   matchLabels:
     app: redis-cluster
 template: # pod创建模板
   metadata:
     labels:
       app: redis-cluster
   spec:
     containers:
     - name: redis
       image: redis:5.0.5-alpine
       ports:
       - containerPort: 6379
         name: client
       - containerPort: 16379
         name: gossip
       command: ["/conf/update-node.sh", "redis-server", "/conf/redis.conf"]
       env:
       - name: POD_IP
         valueFrom:
           fieldRef:
             fieldPath: status.podIP
       volumeMounts:
       - name: conf
         mountPath: /conf
         readOnly: false
       - name: data
         mountPath: /data
         readOnly: false
     volumes:
     - name: conf
       configMap:
         name: redis-cluster
         defaultMode: 0755
 volumeClaimTemplates:
 - metadata:
     name: data
   spec:
     accessModes: [ "ReadWriteOnce" ]
     resources:
       requests:
         storage: 3Gi
     storageClassName: redis-cluster  YAML

[root@master ~]# kubectl apply -f redis-StatefulSets.yaml
# 查看
[root@master ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
nginx-58777cc9fd-cwj77 1/1     Running 1          29h
redis-cluster-0          1/1     Running 0          76m
redis-cluster-1          1/1     Running 0          76m
redis-cluster-2          1/1     Running 0          76m
redis-cluster-3          1/1     Running 0          75m
redis-cluster-4          1/1     Running 0          74m
redis-cluster-5          1/1     Running 0          74m

2.5 创建service对外暴露端口，kubectl apply -f redis-svc.yaml

若需要集群外机器访问，将下列yaml中type值改为NodePort即可。

---
apiVersion: v1
kind: Service
metadata:
 name: redis-cluster
spec:
 type: ClusterIP # 集群内部机器可访问，如需要集群外部访问，将类型改为NodePort即可
 clusterIP: XX.XX.97.97 # 不写会自己分配
 ports:
 - port: 6379
   targetPort: 6379
   name: client
 - port: 16379
   targetPort: 16379
   name: gossip
 selector:
   app: redis-cluster

[root@master ~]# kubectl apply -f redis-svc.yaml
[root@master ~]# kubectl get svc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
redis-cluster ClusterIP XX.XX.97.97     <none>        6379/TCP,16379/TCP 124m

2.6 初始化redis，创建redis集群

kubectl exec -it redis-cluster-0 -- redis-cli --cluster create --cluster-replicas 1 $(kubectl get pods -l app=redis-cluster -o jsonpath='{range.items[*]}{.status.podIP}:6379 ')
[root@master ~]# kubectl exec -it redis-cluster-0 -- redis-cli --cluster create --cluster-replicas 1 $(kubectl get pods -l app=redis-cluster -o jsonpath='{range.items[*]}{.status.podIP}:6379 ')
>>> Performing hash slots allocation on 6 nodes...
Master[0] -> Slots 0 - 5460
Master[1] -> Slots 5461 - 10922
Master[2] -> Slots 10923 - 16383
Adding replica XX.244.2.10:6379 to XX.244.2.8:6379
Adding replica XX.244.1.10:6379 to XX.244.1.8:6379
Adding replica XX.244.1.9:6379 to XX.244.2.9:6379
M: aaf12abf3906e40d7c1084fa7228e99a49fc02df XX.244.2.8:6379
  slots:[0-5460] (5461 slots) master
M: 547619c817623c71502e52413e46bf33bfb307bc XX.244.1.8:6379
  slots:[5461-10922] (5462 slots) master
M: cd3abc406759315a814820dc0a6ce53b93a919a8 XX.244.2.9:6379
  slots:[10923-16383] (5461 slots) master
S: b0e40a1d30b397bacefd0f4c4d8584246ce52fc6 XX.244.1.9:6379
  replicates cd3abc406759315a814820dc0a6ce53b93a919a8
S: d8f1a35fc156598c4d9871a607f2639206072782 XX.244.2.10:6379
  replicates aaf12abf3906e40d7c1084fa7228e99a49fc02df
S: 372b0086cccdcdea81be571df557b958712529a5 XX.244.1.10:6379
  replicates 547619c817623c71502e52413e46bf33bfb307bc
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 XX.244.2.8:6379)
M: aaf12abf3906e40d7c1084fa7228e99a49fc02df XX.244.2.8:6379
  slots:[0-5460] (5461 slots) master
  1 additional replica(s)
S: 372b0086cccdcdea81be571df557b958712529a5 XX.244.1.10:6379
  slots: (0 slots) slave
  replicates 547619c817623c71502e52413e46bf33bfb307bc
S: d8f1a35fc156598c4d9871a607f2639206072782 XX.244.2.10:6379
  slots: (0 slots) slave
  replicates aaf12abf3906e40d7c1084fa7228e99a49fc02df
M: 547619c817623c71502e52413e46bf33bfb307bc XX.244.1.8:6379
  slots:[5461-10922] (5462 slots) master
  1 additional replica(s)
S: b0e40a1d30b397bacefd0f4c4d8584246ce52fc6 XX.244.1.9:6379
  slots: (0 slots) slave
  replicates cd3abc406759315a814820dc0a6ce53b93a919a8
M: cd3abc406759315a814820dc0a6ce53b93a919a8 XX.244.2.9:6379
  slots:[10923-16383] (5461 slots) master
  1 additional replica(s)
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.

2.7 验证集群 redis-cli cluster info

可以看到集群有6个节点，创建成功。

[root@master ~]# kubectl exec -it redis-cluster-0 -- 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:6454
cluster_stats_messages_pong_sent:6664
cluster_stats_messages_sent:13118
cluster_stats_messages_ping_received:6664
cluster_stats_messages_pong_received:6438
cluster_stats_messages_received:13102

2.8 配置修改

将configmap中的redis.conf修改即可，重启pod可将配置生效。

[root@master ~]# kubectl delete pods redis-cluster-0
[root@master ~]# kubectl delete pods redis-cluster-1
[root@master ~]# kubectl delete pods redis-cluster-2
[root@master ~]# kubectl delete pods redis-cluster-3
[root@master ~]# kubectl delete pods redis-cluster-4
[root@master ~]# kubectl delete pods redis-cluster-5

END

本文作者：IT那活儿(上海新炬中北团队）

本文来源：“IT那活儿”公众号

文章转载自IT那活儿，如果涉嫌侵权，请发送邮件至：contact@modb.pro进行举报，并提供相关证据，一经查实，墨天轮将立刻删除相关内容。

K8S搭建redis集群

本文作者：IT那活儿(上海新炬中北团队）

本文来源：“IT那活儿”公众号

评论