kubernetes搭建redis集群

IT那活兒發(fā)布于2023-01-11 13:19 / 606人閱讀

kubernetes搭建redis集群

點(diǎn)擊上方“IT那活兒”公眾號(hào)，關(guān)注后了解更多內(nèi)容，不管IT什么活兒，干就完了！！！

一

知識(shí)點(diǎn)描述

通過(guò)k8s中statefulset類型的pod控制器和pv存儲(chǔ)卷，搭建redis集群。

二

標(biāo)準(zhǔn)指導(dǎo)操作

注意事項(xiàng)：本次演示為本地測(cè)試環(huán)境，k8s集群為單節(jié)點(diǎn)一主兩從集群；在生產(chǎn)環(huán)境需要按照需求搭建。

2.1 主機(jī)規(guī)劃

主機(jī)名	ip	擔(dān)任角色	k8s版本
master	192.168.158.136	master	v1.17.4
node01	192.168.158.137	node	v1.17.4
node02	192.168.158.138	node	v1.17.4

2.2 安裝步驟

1）默認(rèn)以安裝好k8s集群就不做敘述，需要準(zhǔn)備nfs單機(jī)環(huán)境，若k8s為多master集群，可以準(zhǔn)備nfs集群環(huán)境。

# 在master上安裝nfs服務(wù)，設(shè)置開(kāi)機(jī)自啟動(dòng)。

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

# 在node上安裝nfs服務(wù)，注意不需要啟動(dòng)。

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

2）準(zhǔn)備共享目錄，將目錄以讀寫(xiě)權(quán)限暴露給192.168.158.0/24網(wǎng)段中的所有主機(jī)（在生產(chǎn)環(huán)境應(yīng)該只將共享目錄暴露給集群所在機(jī)器）；必須暴漏共享目錄，否則在創(chuàng)建pod時(shí)會(huì)出現(xiàn)報(bào)錯(cuò)無(wú)法找到pv！

# 創(chuàng)建共享目錄。

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

# 將共享目錄暴露，暴露給指定主機(jī)將“192.168.158.0/24”改成主機(jī)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)

3）創(chuàng)建6個(gè)pv，kubectl apply -f redis-pv.yaml，下面為redis-pv.yaml內(nèi)容

apiVersion: v1
kind: PersistentVolume # 創(chuàng)建pv
metadata:
 name: redis-pv1 # 名稱
spec:
 capacity:
   storage: 3Gi # 3G磁盤(pán)空間
 accessModes:
 - ReadWriteOnce # 權(quán)限讀寫(xiě)
 persistentVolumeReclaimPolicy: Recycle # 回收策略，清除數(shù)據(jù)
 storageClassName: "redis-cluster" # 存儲(chǔ)類別為“redis-cluster”，只能允許存儲(chǔ)類別相同的pvc使用
 nfs:
   path: /data/redis-cluster/pv1
   server: 192.168.158.136

# ...
# 中間省略一直到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

# 創(chuàng)建成功可以通過(guò)kubectl命令看到下列6個(gè)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

4）創(chuàng)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: # 標(biāo)簽選擇器
   matchLabels:
     app: redis-cluster
 template: # pod創(chuàng)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

[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

5）創(chuàng)建service對(duì)外暴露端口，kubectl apply -f redis-svc.yaml，若需要集群外機(jī)器訪問(wèn)，將下列yaml中type值改為NodePort即可。

---
apiVersion: v1
kind: Service
metadata:
 name: redis-cluster
spec:
 type: ClusterIP # 集群內(nèi)部機(jī)器可訪問(wèn)，如需要集群外部訪問(wèn)，將類型改為NodePort即可
 clusterIP: 10.96.97.97 # 不寫(xiě)會(huì)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 10.96.97.97      6379/TCP,16379/TCP 124m

6）初始化redis，創(chuàng)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 10.244.2.10:6379 to 10.244.2.8:6379
Adding replica 10.244.1.10:6379 to 10.244.1.8:6379
Adding replica 10.244.1.9:6379 to 10.244.2.9:6379
M: aaf12abf3906e40d7c1084fa7228e99a49fc02df 10.244.2.8:6379
  slots:[0-5460] (5461 slots) master
M: 547619c817623c71502e52413e46bf33bfb307bc 10.244.1.8:6379
  slots:[5461-10922] (5462 slots) master
M: cd3abc406759315a814820dc0a6ce53b93a919a8 10.244.2.9:6379
  slots:[10923-16383] (5461 slots) master
S: b0e40a1d30b397bacefd0f4c4d8584246ce52fc6 10.244.1.9:6379
  replicates cd3abc406759315a814820dc0a6ce53b93a919a8
S: d8f1a35fc156598c4d9871a607f2639206072782 10.244.2.10:6379
  replicates aaf12abf3906e40d7c1084fa7228e99a49fc02df
S: 372b0086cccdcdea81be571df557b958712529a5 10.244.1.10:6379
  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 10.244.2.8:6379)
M: aaf12abf3906e40d7c1084fa7228e99a49fc02df 10.244.2.8:6379
  slots:[0-5460] (5461 slots) master
  1 additional replica(s)
S: 372b0086cccdcdea81be571df557b958712529a5 10.244.1.10:6379
  slots: (0 slots) slave
  replicates 547619c817623c71502e52413e46bf33bfb307bc
S: d8f1a35fc156598c4d9871a607f2639206072782 10.244.2.10:6379
  slots: (0 slots) slave
  replicates aaf12abf3906e40d7c1084fa7228e99a49fc02df
M: 547619c817623c71502e52413e46bf33bfb307bc 10.244.1.8:6379
  slots:[5461-10922] (5462 slots) master
  1 additional replica(s)
S: b0e40a1d30b397bacefd0f4c4d8584246ce52fc6 10.244.1.9:6379
  slots: (0 slots) slave
  replicates cd3abc406759315a814820dc0a6ce53b93a919a8
M: cd3abc406759315a814820dc0a6ce53b93a919a8 10.244.2.9:6379
  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.

7）驗(yàn)證集群 redis-cli cluster info，可以看到集群有6個(gè)節(jié)點(diǎn)，創(chuàng)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

8）配置修改，將configmap中的redis.conf修改即可，重啟pod可將配置生效。

[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