k8s部署redis6节点，组成3主3从集群模式
一般来说，redis部署有三种模式。
单实例模式，一般用于测试环境。
哨兵模式
集群模式
后两者用于生产部署
哨兵模式
在redis3.0以前，要实现集群一般是借助哨兵sentinel工具来监控master节点的状态。
如果master节点异常，则会做主从切换，将某一台slave作为master。
引入了哨兵节点，部署更复杂，维护成本也比较高，并且性能和高可用性等各方面表现一般。
集群模式
3.0 后推出的 Redis 分布式集群解决方案
主节点提供读写操作，从节点作为备用节点，不提供请求，只作为故障转移使用
如果master节点异常，也是会自动做主从切换，将slave切换为master。
总的来说，集群模式明显优于哨兵模式
一、redis集群搭建
RC、Deployment、DaemonSet都是面向无状态的服务，它们所管理的Pod的IP、名字，启停顺序等都是随机的，而StatefulSet是什么？顾名思义，有状态的集合，管理所有有状态的服务，比如MySQL、MongoDB集群等。
StatefulSet本质上是Deployment的一种变体，在v1.9版本中已成为GA版本，它为了解决有状态服务的问题，它所管理的Pod拥有固定的Pod名称，启停顺序，在StatefulSet中，Pod名字称为网络标识(hostname)，还必须要用到共享存储。
在Deployment中，与之对应的服务是service，而在StatefulSet中与之对应的headless service，headless service，即无头服务，与service的区别就是它没有Cluster IP，解析它的名称时将返回该Headless Service对应的全部Pod的Endpoint列表。
除此之外，StatefulSet在Headless Service的基础上又为StatefulSet控制的每个Pod副本创建了一个DNS域名，这个域名的格式为：
( p o d . n a m e ) . (pod.name). (pod.name).(headless server.name).${namespace}.svc.cluster.local
也即是说，对于有状态服务，我们最好使用固定的网络标识（如域名信息）来标记节点，当然这也需要应用程序的支持（如Zookeeper就支持在配置文件中写入主机域名）。
StatefulSet基于Headless Service（即没有Cluster IP的Service）为Pod实现了稳定的网络标志（包括Pod的hostname和DNS Records），在Pod重新调度后也保持不变。同时，结合PV/PVC，StatefulSet可以实现稳定的持久化存储，就算Pod重新调度后，还是能访问到原先的持久化数据。
以下为使用StatefulSet部署Redis的架构，无论是Master还是Slave，都作为StatefulSet的一个副本，并且数据通过PV进行持久化，对外暴露为一个Service，接受客户端请求。
1.redis.conf配置文件参考
说明：我们知道，redis默认目录是/var/lib/redis/和/etc/redis/，同时官方在构建redis镜像时，默认工作目录在/data目录，所以本篇为了规范redis数据存放目录，将redis.conf挂载到/etc/redis/下，其他redis日志文件、数据文件全部放到/data目录下。
#[root@master redis]# vi redis.conf #编写一个redis.conf配置文件
#[root@master redis]# grep -Ev "$^|#" redis.conf #下面是redis.conf配置文件
bind 0.0.0.0
protected-mode yes
port 6379 #redis端口,为了安全设置为6379端口
tcp-backlog 511
timeout 0
tcp-keepalive 300
daemonize no #redis是否以后台模式运行,必须设置no
supervised no
pidfile /data/redis.pid #redis的pid文件,放到/data目录下
loglevel notice
logfile /data/redis_log #redis日志文件,放到/data目录下
databases 16
always-show-logo yes
save 900 1
save 300 10
save 60 10000
stop-writes-on-bgsave-error yes
rdbcompression yes
rdbchecksum yes
dbfilename dump.rdb #这个文件会放在dir定义的/data目录
dir /data #数据目录
masterauth iloveyou #redis集群各节点相互认证的密码,必须配置和下面的requirepass一致
replica-serve-stale-data yes
replica-read-only yes
repl-diskless-sync no
repl-diskless-sync-delay 5
repl-disable-tcp-nodelay no
replica-priority 100
requirepass iloveyou #redis的密码
lazyfree-lazy-eviction no
lazyfree-lazy-expire no
lazyfree-lazy-server-del no
replica-lazy-flush no
appendonly no
appendfilename "appendonly.aof" #这个文件会放在dir定义的/data目录
appendfsync everysec
no-appendfsync-on-rewrite no
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb
aof-load-truncated yes
aof-use-rdb-preamble yes
lua-time-limit 5000
cluster-enabled yes #是否启用集群模式，必须去掉注释设为yes
cluster-config-file nodes.conf #这个文件会放在dir定义的/data目录
cluster-node-timeout 15000
slowlog-log-slower-than 10000
slowlog-max-len 128
latency-monitor-threshold 0
notify-keyspace-events ""
hash-max-ziplist-entries 512
hash-max-ziplist-value 64
list-max-ziplist-size -2
list-compress-depth 0
set-max-intset-entries 512
zset-max-ziplist-entries 128
zset-max-ziplist-value 64
hll-sparse-max-bytes 3000
stream-node-max-bytes 4096
stream-node-max-entries 100
activerehashing yes
client-output-buffer-limit normal 0 0 0
client-output-buffer-limit replica 256mb 64mb 60
client-output-buffer-limit pubsub 32mb 8mb 60
hz 10
dynamic-hz yes
aof-rewrite-incremental-fsync yes
rdb-save-incremental-fsync yes2.创建statefulsets有状态应用
redis集群一般可以使用deployment和statefulsets，这里使用statefulsets有状态应用来创建redis，创建sts有状态应用需要有一个headless service，同时在sts中挂载configmap卷，使用动态分配pv用于redis数据持久化。
cat 03-redis-cluster-sts.yaml
cat > 03-redis-cluster-sts.yaml << 'eof'
apiVersion: v1
kind: ConfigMap
metadata:name: redis-cluster-config
data:redis-cluster.conf: |daemonize nosupervised noprotected-mode nobind 0.0.0.0port 6379cluster-announce-bus-port 16379cluster-enabled yesappendonly yescluster-node-timeout 5000dir /datacluster-config-file /data/nodes.confrequirepass iloveyoumasterauth iloveyoupidfile /data/redis.pid loglevel noticelogfile /data/redis_log
---
apiVersion: v1
kind: Service
metadata:name: redis-cluster-service
spec:selector:app: redis-clusterclusterIP: Noneports:- name: redis-6379port: 6379- name: redis-16379port: 16379
---
apiVersion: v1
kind: Service
metadata:name: redis-cluster-service-access
spec:selector:app: redis-clustertype: NodePortsessionAffinity: Noneports:- name: redis-6379port: 6379targetPort: 6379nodePort: 30202
---
apiVersion: apps/v1
kind: StatefulSet
metadata:labels:app: redis-clustername: redis-cluster
spec:serviceName: redis-cluster-servicereplicas: 6selector:matchLabels:app: redis-clustertemplate:metadata:labels:app: redis-clusterspec:terminationGracePeriodSeconds: 30containers:- name: redisimage: redis:6.2.6imagePullPolicy: IfNotPresentterminationMessagePath: /dev/termination-logterminationMessagePolicy: Filecommand: [ "redis-server", "/etc/redis/redis-cluster.conf" ]args:- "--cluster-announce-ip"- "$(POD_IP)"env:- name: HOST_IPvalueFrom:fieldRef:fieldPath: status.hostIP- name: POD_IPvalueFrom:fieldRef:fieldPath: status.podIP- name: POD_NAMEvalueFrom:fieldRef:fieldPath: metadata.name- name: TZvalue: "Asia/Shanghai"ports:- name: rediscontainerPort: 6379protocol: TCP- name: clustercontainerPort: 16379protocol: TCPvolumeMounts:- name: redis-confmountPath: /etc/redis- name: pvc-datamountPath: /datavolumes:- name: timezonehostPath:path: /usr/share/zoneinfo/Asia/Shanghai- name: redis-confconfigMap:name: redis-cluster-configitems:- key: redis-cluster.confpath: redis-cluster.confvolumeClaimTemplates:- metadata:name: pvc-dataspec:accessModes: [ "ReadWriteOnce" ]resources:requests:storage: 200MaccessModes:- ReadWriteOnce #访问模式为RWOstorageClassName: "nfs-boge" #指定使用的存储类，实现动态分配pv
eofkubectl apply -f 03-redis-cluster-sts.yaml
二、构建3主3从集群模式
6个pod已经创建完毕，状态都是running，下面将6个pod 组成redis集群，3主3从模式。
命令说明：
- –replicas 1或者–cluster-replicas 1 ：指定集群中每个master的副本个数为1，此时节点总数 ÷ (replicas + 1) 得到的就是master的数量。因此节点列表中的前n个就是master，其它节点都是slave节点，随机分配到不同master
#在redis任意一个pod执行初始化命令,可以进入到pod里面执行也可以直接在外面执行
#其中为了获取每个pod的ip,使用
# kubectl get pods -l app=redis-cluster -o jsonpath='{range.items[*]}{.status.podIP}:6379 {end}'
#本次采用自动创建redis的形式,也就是说不指定哪个是主哪个是从节点,让redis自动分配,生产环境中也建议使用该种方式[root@master redis]# kubectl exec -it redis-cluster-0 -- redis-cli -a iloveyou --cluster create --cluster-replicas 1 $(kubectl get pods -l app=redis-cluster -o jsonpath='{range.items[*]}{.status.podIP}:6379 {end}')
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
>>> Performing hash slots allocation on 6 nodes...
Master[0] -> Slots 0 - 5460
Master[1] -> Slots 5461 - 10922
Master[2] -> Slots 10923 - 16383
Adding replica 10.244.2.22:6379 to 10.244.2.20:6379
Adding replica 10.244.1.20:6379 to 10.244.1.18:6379
Adding replica 10.244.2.21:6379 to 10.244.1.19:6379
M: 972b376e8cc658b8bf5f2a1a3294cbe2c84ee852 10.244.2.20:6379slots:[0-5460] (5461 slots) master
M: ac6cc9dd3a86cf370333d36933c99df5f13f42ab 10.244.1.18:6379slots:[5461-10922] (5462 slots) master
M: 18b4ceacd3222e546ab59e041e4ae50e736c5c26 10.244.1.19:6379slots:[10923-16383] (5461 slots) master
S: 8394ceff0b32fc7119b65704ea78e9b5bbc2fbd7 10.244.2.21:6379replicates 18b4ceacd3222e546ab59e041e4ae50e736c5c26
S: 565f9f9931323f8ac0376b7a7ec701f0a2955e8b 10.244.2.22:6379replicates 972b376e8cc658b8bf5f2a1a3294cbe2c84ee852
S: 5c1270743b6a5f81003da4402f39c360631a2d0f 10.244.1.20:6379replicates ac6cc9dd3a86cf370333d36933c99df5f13f42ab
Can I set the above configuration? (type 'yes' to accept): yes #输入yes表示接受redis给我们自动分配的槽点,主从也是redis任意指定的
>>> 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 10.244.2.20:6379)
M: 972b376e8cc658b8bf5f2a1a3294cbe2c84ee852 10.244.2.20:6379slots:[0-5460] (5461 slots) master1 additional replica(s)
S: 5c1270743b6a5f81003da4402f39c360631a2d0f 10.244.1.20:6379slots: (0 slots) slavereplicates ac6cc9dd3a86cf370333d36933c99df5f13f42ab
M: 18b4ceacd3222e546ab59e041e4ae50e736c5c26 10.244.1.19:6379slots:[10923-16383] (5461 slots) master1 additional replica(s)
M: ac6cc9dd3a86cf370333d36933c99df5f13f42ab 10.244.1.18:6379slots:[5461-10922] (5462 slots) master1 additional replica(s)
S: 565f9f9931323f8ac0376b7a7ec701f0a2955e8b 10.244.2.22:6379slots: (0 slots) slavereplicates 972b376e8cc658b8bf5f2a1a3294cbe2c84ee852
S: 8394ceff0b32fc7119b65704ea78e9b5bbc2fbd7 10.244.2.21:6379slots: (0 slots) slavereplicates 18b4ceacd3222e546ab59e041e4ae50e736c5c26
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
[root@master redis]# 1.验证redis
#往集群里面写入一百条数据。
for line in {1..100};do kubectl exec -it redis-cluster-0 -- redis-cli -c -p 6379 -a iloveyou -c set ops_${line} ${line}; done#读取数据看一下是否能读取到。
for line in {1..100};do kubectl exec -it redis-cluster-1 -- redis-cli -c -p 6379 -a iloveyou -c get ops_${line}; done#登入redis
kubectl exec -it redis-cluster-0 -- redis-cli -c -p 6379 -a iloveyou#查看redis信息
127.0.0.1:6379> cluster info#检查redis集群信息
127.0.0.1:6379> cluster nodes#插入数据
127.0.0.1:6379> set num 111#遍历所有的key
127.0.0.1:6379> keys *#删除key
127.0.0.1:6379> del num#登入redis0
kubectl exec -it redis-cluster-0 -- redis-cli -c -p 6379 -a iloveyou#遍历所有的key
127.0.0.1:6379> keys *
#登入redis1
kubectl exec -it redis-cluster-1 -- redis-cli -c -p 6379 -a iloveyou#遍历所有的key
127.0.0.1:6379> keys *
#登入redis2
kubectl exec -it redis-cluster-2 -- redis-cli -c -p 6379 -a iloveyou#遍历所有的key
127.0.0.1:6379> keys *
2.删除pod验证
2.1 验证集群信息
[root@k8s-master01 redis]# kubectl get po -owide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
redis-cluster-0 1/1 Running 0 61m 100.125.152.57 k8s-node02 <none> <none>
redis-cluster-1 1/1 Running 0 33m 100.97.125.39 k8s-node01 <none> <none>
redis-cluster-2 1/1 Running 0 60m 100.125.152.22 k8s-node02 <none> <none>
redis-cluster-3 1/1 Running 0 60m 100.97.125.10 k8s-node01 <none> <none>
redis-cluster-4 1/1 Running 0 60m 100.125.152.41 k8s-node02 <none> <none>
redis-cluster-5 1/1 Running 0 60m 100.97.125.57 k8s-node01 <none> <none>
[root@k8s-master01 redis]##登入redis
kubectl exec -it redis-cluster-0 – redis-cli -c -p 6379 -a iloveyou
#检查redis集群信息
127.0.0.1:6379> cluster nodes
e4189ba4094320d4f379fb0cbc78cac6cdb3abe6 100.97.125.10:6379@16379 slave 0e3cae56f05f211c35b4401f6a0e61454ca4f64a 0 1693891070028 3 connected
9cdff4df51646dd440e5005bac7f0cb5f4470d9b 100.125.152.41:6379@16379 slave 0a228f6fbd7fc75247180ac31516146e8abd0466 0 1693891069927 1 connected
93ccc4c795f74da6acf84821bf372a2414375f55 100.97.125.39:6379@16379 master - 0 1693891070430 2 connected 5461-10922
ff960d1a1aca448a391b69da155947b4c627677e 100.97.125.57:6379@16379 slave 93ccc4c795f74da6acf84821bf372a2414375f55 0 1693891069000 2 connected
0e3cae56f05f211c35b4401f6a0e61454ca4f64a 100.125.152.22:6379@16379 master - 0 1693891069426 3 connected 10923-16383
0a228f6fbd7fc75247180ac31516146e8abd0466 100.125.152.57:6379@16379 myself,master - 0 1693891069000 1 connected 0-54602.2.删除pod
[root@k8s-master01 redis]# kubectl get po -owide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
redis-cluster-0 1/1 Running 0 61m 100.125.152.57 k8s-node02 <none> <none>
redis-cluster-1 1/1 Running 0 33m 100.97.125.39 k8s-node01 <none> <none>
redis-cluster-2 1/1 Running 0 60m 100.125.152.22 k8s-node02 <none> <none>
redis-cluster-3 1/1 Running 0 60m 100.97.125.10 k8s-node01 <none> <none>
redis-cluster-4 1/1 Running 0 60m 100.125.152.41 k8s-node02 <none> <none>
redis-cluster-5 1/1 Running 0 60m 100.97.125.57 k8s-node01 <none> <none>
[root@k8s-master01 redis]# kubectl delete po redis-cluster-5 --force
warning: Immediate deletion does not wait for confirmation that the running resource has been terminated. The resource may continue to run on the cluster indefinitely.
pod "redis-cluster-5" force deleted
[root@k8s-master01 redis]# kubectl get po -owide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
redis-cluster-0 1/1 Running 0 62m 100.125.152.57 k8s-node02 <none> <none>
redis-cluster-1 1/1 Running 0 35m 100.97.125.39 k8s-node01 <none> <none>
redis-cluster-2 1/1 Running 0 62m 100.125.152.22 k8s-node02 <none> <none>
redis-cluster-3 1/1 Running 0 62m 100.97.125.10 k8s-node01 <none> <none>
redis-cluster-4 1/1 Running 0 61m 100.125.152.41 k8s-node02 <none> <none>
redis-cluster-5 1/1 Running 0 6s 100.97.125.37 k8s-node01 <none> <none>
[root@k8s-master01 redis]#2.2 验证redis集群
[root@k8s-master01 redis]# kubectl exec -it redis-cluster-1 -- redis-cli -c -p 6379 -a iloveyou
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
127.0.0.1:6379> cluster nodes
e4189ba4094320d4f379fb0cbc78cac6cdb3abe6 100.97.125.10:6379@16379 slave 0e3cae56f05f211c35b4401f6a0e61454ca4f64a 0 1693891297573 3 connected
9cdff4df51646dd440e5005bac7f0cb5f4470d9b 100.125.152.41:6379@16379 slave 0a228f6fbd7fc75247180ac31516146e8abd0466 0 1693891297073 1 connected
93ccc4c795f74da6acf84821bf372a2414375f55 100.97.125.39:6379@16379 myself,master - 0 1693891297000 2 connected 5461-10922
0e3cae56f05f211c35b4401f6a0e61454ca4f64a 100.125.152.22:6379@16379 master - 0 1693891297573 3 connected 10923-16383
ff960d1a1aca448a391b69da155947b4c627677e 100.97.125.37:6379@16379 slave 93ccc4c795f74da6acf84821bf372a2414375f55 0 1693891297073 2 connected
0a228f6fbd7fc75247180ac31516146e8abd0466 100.125.152.57:6379@16379 master - 0 1693891296571 1 connected 0-5460
127.0.0.1:6379>查询插入数据验证
del num#读取数据看一下是否能读取到。
for line in {1..100};do kubectl exec -it redis-cluster-4 -- redis-cli -c -p 6379 -a iloveyou -c get ops_${line}; done#往集群里面写入一百条数据。
for line in {1..200};do kubectl exec -it redis-cluster-5 -- redis-cli -c -p 6379 -a iloveyou -c set ops_${line} ${line}; donefor line in {1..200};do kubectl exec -it redis-cluster-2 -- redis-cli -c -p 6379 -a iloveyou -c del ops_${line}; done#往集群里面写入32条数据。
for line in {1..32};do kubectl exec -it redis-cluster-5 -- redis-cli -c -p 6379 -a iloveyou -c set ops_${line} ${line}; done
