k8s实战案例之部署redis单机和redis cluster

1、在k8s上部署redis单机

1.1、redis简介

redis是一款基于BSD协议，开源的非关系型数据库（nosql数据库），作者是意大利开发者Salvatore Sanfilippo在2009年发布，使用C语言编写；redis是基于内存存储，而且是目前比较流行的键值数据库（key-value database），它提供将内存通过网络远程共享的一种服务，提供类似功能的还有memcache，但相比 memcache，redis 还提供了易扩展、高性能、具备数据持久性等功能。主要的应用场景有session共享，常用于web集群中的tomcat或PHP中多web服务器的session共享；消息队列，ELK的日志缓存，部分业务的订阅发布系统；计数器，常用于访问排行榜，商品浏览数等和次数相关的数值统计场景；缓存，常用于数据查询、电商网站商品信息、新闻内容等；相对memcache，redis支持数据的持久化，可以将内存的数据保存在磁盘中，重启redis服务或者服务器之后可以从备份文件中恢复数据到内存继续使用；
1.2、PV/PVC 及 Redis 单机

由于redis的数据（主要是redis快照）都存放在存储系统中，即便redis pod挂掉，对应数据都不会丢；因为在k8s上部署redis单机，redis pod挂了，k8s会将对应pod重建，重建时会把对应pvc挂载至pod中，加载快照，从而使得redis的数据不被pod的挂掉而丢数据；

1.3、构建redis镜像

1. root@k8s-master01:~/k8s-data/dockerfile/web/magedu/redis# ll
2. total 1784
3. drwxr-xr-x  2 root root    4096 Jun  5 15:22 ./
4. drwxr-xr-x 11 root root    4096 Aug  9  2022 ../
5. -rw-r--r--  1 root root     717 Jun  5 15:20 Dockerfile
6. -rwxr-xr-x  1 root root     235 Jun  5 15:21 build-command.sh*
7. -rw-r--r--  1 root root 1740967 Jun 22  2021 redis-4.0.14.tar.gz
8. -rw-r--r--  1 root root   58783 Jun 22  2021 redis.conf
9. -rwxr-xr-x  1 root root      84 Jun  5 15:21 run_redis.sh*
10. root@k8s-master01:~/k8s-data/dockerfile/web/magedu/redis# cat Dockerfile
11. #Redis Image
12. # 导入自定义centos基础镜像
13. FROM harbor.ik8s.cc/baseimages/magedu-centos-base:7.9.2009
14. # 添加redis源码包至/usr/local/src
15. ADD redis-4.0.14.tar.gz /usr/local/src
16. # 编译安装redis
17. RUN ln -sv /usr/local/src/redis-4.0.14 /usr/local/redis && cd /usr/local/redis && make && cp src/redis-cli /usr/sbin/ && cp src/redis-server  /usr/sbin/ && mkdir -pv /data/redis-data
18. # 添加redis配置文件
19. ADD redis.conf /usr/local/redis/redis.conf
20. # 暴露redis服务端口
21. EXPOSE 6379
23. #ADD run_redis.sh /usr/local/redis/run_redis.sh
24. #CMD ["/usr/local/redis/run_redis.sh"]
25. # 添加启动脚本
26. ADD run_redis.sh /usr/local/redis/entrypoint.sh
27. # 启动redis
28. ENTRYPOINT ["/usr/local/redis/entrypoint.sh"]
29. root@k8s-master01:~/k8s-data/dockerfile/web/magedu/redis# cat build-command.sh
30. #!/bin/bash
31. TAG=$1
32. #docker build -t harbor.ik8s.cc/magedu/redis:${TAG} .
33. #sleep 3
34. #docker push  harbor.ik8s.cc/magedu/redis:${TAG}
36. nerdctl build -t  harbor.ik8s.cc/magedu/redis:${TAG} .
37. nerdctl push harbor.ik8s.cc/magedu/redis:${TAG}
38. root@k8s-master01:~/k8s-data/dockerfile/web/magedu/redis# cat run_redis.sh
39. #!/bin/bash
40. # Redis启动命令
41. /usr/sbin/redis-server /usr/local/redis/redis.conf
42. # 使用tail -f 在pod内部构建守护进程
43. tail -f  /etc/hosts
44. root@k8s-master01:~/k8s-data/dockerfile/web/magedu/redis# grep -v '^#\|^$' redis.conf
45. bind 0.0.0.0
46. protected-mode yes
47. port 6379
48. tcp-backlog 511
49. timeout 0
50. tcp-keepalive 300
51. daemonize yes
52. supervised no
53. pidfile /var/run/redis_6379.pid
54. loglevel notice
55. logfile ""
56. databases 16
57. always-show-logo yes
58. save 900 1
59. save 5 1
60. save 300 10
61. save 60 10000
62. stop-writes-on-bgsave-error no
63. rdbcompression yes
64. rdbchecksum yes
65. dbfilename dump.rdb
66. dir /data/redis-data
67. slave-serve-stale-data yes
68. slave-read-only yes
69. repl-diskless-sync no
70. repl-diskless-sync-delay 5
71. repl-disable-tcp-nodelay no
72. slave-priority 100
73. requirepass 123456
74. lazyfree-lazy-eviction no
75. lazyfree-lazy-expire no
76. lazyfree-lazy-server-del no
77. slave-lazy-flush no
78. appendonly no
79. appendfilename "appendonly.aof"
80. appendfsync everysec
81. no-appendfsync-on-rewrite no
82. auto-aof-rewrite-percentage 100
83. auto-aof-rewrite-min-size 64mb
84. aof-load-truncated yes
85. aof-use-rdb-preamble no
86. lua-time-limit 5000
87. slowlog-log-slower-than 10000
88. slowlog-max-len 128
89. latency-monitor-threshold 0
90. notify-keyspace-events ""
91. hash-max-ziplist-entries 512
92. hash-max-ziplist-value 64
93. list-max-ziplist-size -2
94. list-compress-depth 0
95. set-max-intset-entries 512
96. zset-max-ziplist-entries 128
97. zset-max-ziplist-value 64
98. hll-sparse-max-bytes 3000
99. activerehashing yes
100. client-output-buffer-limit normal 0 0 0
101. client-output-buffer-limit slave 256mb 64mb 60
102. client-output-buffer-limit pubsub 32mb 8mb 60
103. hz 10
104. aof-rewrite-incremental-fsync yes
105. root@k8s-master01:~/k8s-data/dockerfile/web/magedu/redis#

复制代码

1.3.1、验证rdis镜像是否上传至harbor?


1.4、测试redis 镜像

1.4.1、验证将redis镜像运行为容器，看看是否正常运行?


1.4.2、远程连接redis，看看是否可正常连接？

能够将redis镜像运行为容器，并且能够通过远程主机连接至redis进行数据读写，说明我们构建的reids镜像没有问题；

1.5、创建PV和PVC

1.5.1、在nfs服务器上准备redis数据存储目录

1. root@harbor:~# mkdir -pv /data/k8sdata/magedu/redis-datadir-1
2. mkdir: created directory '/data/k8sdata/magedu/redis-datadir-1'
3. root@harbor:~# cat /etc/exports
4. # /etc/exports: the access control list for filesystems which may be exported
5. #               to NFS clients.  See exports(5).
6. #
7. # Example for NFSv2 and NFSv3:
8. # /srv/homes       hostname1(rw,sync,no_subtree_check) hostname2(ro,sync,no_subtree_check)
9. #
10. # Example for NFSv4:
11. # /srv/nfs4        gss/krb5i(rw,sync,fsid=0,crossmnt,no_subtree_check)
12. # /srv/nfs4/homes  gss/krb5i(rw,sync,no_subtree_check)
13. #
14. /data/k8sdata/kuboard *(rw,no_root_squash)
15. /data/volumes *(rw,no_root_squash)
16. /pod-vol *(rw,no_root_squash)
17. /data/k8sdata/myserver *(rw,no_root_squash)
18. /data/k8sdata/mysite *(rw,no_root_squash)
20. /data/k8sdata/magedu/images *(rw,no_root_squash)
21. /data/k8sdata/magedu/static *(rw,no_root_squash)
24. /data/k8sdata/magedu/zookeeper-datadir-1 *(rw,no_root_squash)
25. /data/k8sdata/magedu/zookeeper-datadir-2 *(rw,no_root_squash)
26. /data/k8sdata/magedu/zookeeper-datadir-3 *(rw,no_root_squash)
29. /data/k8sdata/magedu/redis-datadir-1 *(rw,no_root_squash)
31. root@harbor:~# exportfs -av
32. exportfs: /etc/exports [1]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/kuboard".
33.   Assuming default behaviour ('no_subtree_check').
34.   NOTE: this default has changed since nfs-utils version 1.0.x
36. exportfs: /etc/exports [2]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/volumes".
37.   Assuming default behaviour ('no_subtree_check').
38.   NOTE: this default has changed since nfs-utils version 1.0.x
40. exportfs: /etc/exports [3]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/pod-vol".
41.   Assuming default behaviour ('no_subtree_check').
42.   NOTE: this default has changed since nfs-utils version 1.0.x
44. exportfs: /etc/exports [4]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/myserver".
45.   Assuming default behaviour ('no_subtree_check').
46.   NOTE: this default has changed since nfs-utils version 1.0.x
48. exportfs: /etc/exports [5]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/mysite".
49.   Assuming default behaviour ('no_subtree_check').
50.   NOTE: this default has changed since nfs-utils version 1.0.x
52. exportfs: /etc/exports [7]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/images".
53.   Assuming default behaviour ('no_subtree_check').
54.   NOTE: this default has changed since nfs-utils version 1.0.x
56. exportfs: /etc/exports [8]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/static".
57.   Assuming default behaviour ('no_subtree_check').
58.   NOTE: this default has changed since nfs-utils version 1.0.x
60. exportfs: /etc/exports [11]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/zookeeper-datadir-1".
61.   Assuming default behaviour ('no_subtree_check').
62.   NOTE: this default has changed since nfs-utils version 1.0.x
64. exportfs: /etc/exports [12]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/zookeeper-datadir-2".
65.   Assuming default behaviour ('no_subtree_check').
66.   NOTE: this default has changed since nfs-utils version 1.0.x
68. exportfs: /etc/exports [13]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/zookeeper-datadir-3".
69.   Assuming default behaviour ('no_subtree_check').
70.   NOTE: this default has changed since nfs-utils version 1.0.x
72. exportfs: /etc/exports [16]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/redis-datadir-1".
73.   Assuming default behaviour ('no_subtree_check').
74.   NOTE: this default has changed since nfs-utils version 1.0.x
76. exporting *:/data/k8sdata/magedu/redis-datadir-1
77. exporting *:/data/k8sdata/magedu/zookeeper-datadir-3
78. exporting *:/data/k8sdata/magedu/zookeeper-datadir-2
79. exporting *:/data/k8sdata/magedu/zookeeper-datadir-1
80. exporting *:/data/k8sdata/magedu/static
81. exporting *:/data/k8sdata/magedu/images
82. exporting *:/data/k8sdata/mysite
83. exporting *:/data/k8sdata/myserver
84. exporting *:/pod-vol
85. exporting *:/data/volumes
86. exporting *:/data/k8sdata/kuboard
87. root@harbor:~#

复制代码

1.5.2、创建pv

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis/pv# cat redis-persistentvolume.yaml     
2. ---
3. apiVersion: v1
4. kind: PersistentVolume
5. metadata:
6.   name: redis-datadir-pv-1
7. spec:
8.   capacity:
9.     storage: 10Gi
10.   accessModes:
11.     - ReadWriteOnce
12.   nfs:
13.     path: /data/k8sdata/magedu/redis-datadir-1
14.     server: 192.168.0.42
15. root@k8s-master01:~/k8s-data/yaml/magedu/redis/pv#

复制代码

1.5.3、创建pvc

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis/pv# cat redis-persistentvolumeclaim.yaml
2. ---
3. apiVersion: v1
4. kind: PersistentVolumeClaim
5. metadata:
6.   name: redis-datadir-pvc-1
7.   namespace: magedu
8. spec:
9.   volumeName: redis-datadir-pv-1
10.   accessModes:
11.     - ReadWriteOnce
12.   resources:
13.     requests:
14.       storage: 10Gi
15. root@k8s-master01:~/k8s-data/yaml/magedu/redis/pv#

复制代码

1.6、部署redis服务

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis# cat redis.yaml
2. kind: Deployment
3. #apiVersion: extensions/v1beta1
4. apiVersion: apps/v1
5. metadata:
6.   labels:
7.     app: devops-redis
8.   name: deploy-devops-redis
9.   namespace: magedu
10. spec:
11.   replicas: 1
12.   selector:
13.     matchLabels:
14.       app: devops-redis
15.   template:
16.     metadata:
17.       labels:
18.         app: devops-redis
19.     spec:
20.       containers:
21.         - name: redis-container
22.           image: harbor.ik8s.cc/magedu/redis:v4.0.14
23.           imagePullPolicy: Always
24.           volumeMounts:
25.           - mountPath: "/data/redis-data/"
26.             name: redis-datadir
27.       volumes:
28.         - name: redis-datadir
29.           persistentVolumeClaim:
30.             claimName: redis-datadir-pvc-1
32. ---
33. kind: Service
34. apiVersion: v1
35. metadata:
36.   labels:
37.     app: devops-redis
38.   name: srv-devops-redis
39.   namespace: magedu
40. spec:
41.   type: NodePort
42.   ports:
43.   - name: http
44.     port: 6379
45.     targetPort: 6379
46.     nodePort: 36379
47.   selector:
48.     app: devops-redis
49.   sessionAffinity: ClientIP
50.   sessionAffinityConfig:
51.     clientIP:
52.       timeoutSeconds: 10800
53. root@k8s-master01:~/k8s-data/yaml/magedu/redis#

复制代码

上述报错说我们的服务端口超出范围，这是因为我们在初始化k8s集群时指定的服务端口范围；

1.6.1、修改nodeport端口范围

编辑/etc/systemd/system/kube-apiserver.service,将其--service-node-port-range选项指定的值修改即可；其他两个master节点也需要修改哦

1.6.2、重载kube-apiserver.service,重启kube-apiserver

1. root@k8s-master01:~# systemctl daemon-reload                 
2. root@k8s-master01:~# systemctl restart kube-apiserver.service
3. root@k8s-master01:~#

复制代码

再次部署redis

1.7、验证redis数据读写

1.7.1、连接k8s任意节点的36376端口，测试redis读写数据


1.8、验证redis pod 重建对应数据是否丢失？

1.8.1、查看redis快照文件是否存储到存储上呢？

1. root@harbor:~# ll /data/k8sdata/magedu/redis-datadir-1
2. total 12
3. drwxr-xr-x 2 root root 4096 Jun  5 16:29 ./
4. drwxr-xr-x 8 root root 4096 Jun  5 15:53 ../
5. -rw-r--r-- 1 root root  116 Jun  5 16:29 dump.rdb
6. root@harbor:~#

复制代码

可以看到刚才我们向redis写入数据，对应redis在规定时间内发现key的变化就做了快照，因为redis数据目录时通过pv/pvc挂载的nfs，所以我们在nfs对应目录里时可以正常看到这个快照文件的；

1.8.2、删除redis pod 等待k8s重建redis pod


1.8.3、验证重建后的redis pod数据

可以看到k8s重建后的redis pod 还保留着原有pod的数据；这说明k8s重建时挂载了前一个pod的pvc;

2、在k8s上部署redis集群

2.1、PV/PVC及Redis Cluster-StatefulSet

redis cluster相比redis单机要稍微复杂一点，我们也是通过pv/pvc将redis cluster数据存放在存储系统中，不同于redis单机，redis cluster对存入的数据会做crc16计算，然后和16384做取模计算，得出一个数字，这个数字就是存入redis cluster的一个槽位；即redis cluster将16384个槽位，平均分配给集群所有master节点，每个master节点存放整个集群数据的一部分；这样一来就存在一个问题，如果master宕机，那么对应槽位的数据也就不可用，为了防止master单点故障，我们还需要对master做高可用，即专门用一个slave节点对master做备份，master宕机的情况下，对应slave会接管master继续向集群提供服务，从而实现redis cluster master的高可用；如上图所示，我们使用3主3从的redis cluster，redis0,1,2为master，那么3，4，5就对应为0，1，2的slave，负责备份各自对应的master的数据；这六个pod都是通过k8s集群的pv/pvc将数据存放在存储系统中；

2.2、创建PV

2.2.1、在nfs上准备redis cluster 数据目录

1. root@harbor:~# mkdir -pv /data/k8sdata/magedu/redis{0,1,2,3,4,5}
2. mkdir: created directory '/data/k8sdata/magedu/redis0'
3. mkdir: created directory '/data/k8sdata/magedu/redis1'
4. mkdir: created directory '/data/k8sdata/magedu/redis2'
5. mkdir: created directory '/data/k8sdata/magedu/redis3'
6. mkdir: created directory '/data/k8sdata/magedu/redis4'
7. mkdir: created directory '/data/k8sdata/magedu/redis5'
8. root@harbor:~# tail -6 /etc/exports
9. /data/k8sdata/magedu/redis0 *(rw,no_root_squash)
10. /data/k8sdata/magedu/redis1 *(rw,no_root_squash)
11. /data/k8sdata/magedu/redis2 *(rw,no_root_squash)
12. /data/k8sdata/magedu/redis3 *(rw,no_root_squash)
13. /data/k8sdata/magedu/redis4 *(rw,no_root_squash)
14. /data/k8sdata/magedu/redis5 *(rw,no_root_squash)
15. root@harbor:~# exportfs  -av
16. exportfs: /etc/exports [1]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/kuboard".
17.   Assuming default behaviour ('no_subtree_check').
18.   NOTE: this default has changed since nfs-utils version 1.0.x
20. exportfs: /etc/exports [2]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/volumes".
21.   Assuming default behaviour ('no_subtree_check').
22.   NOTE: this default has changed since nfs-utils version 1.0.x
24. exportfs: /etc/exports [3]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/pod-vol".
25.   Assuming default behaviour ('no_subtree_check').
26.   NOTE: this default has changed since nfs-utils version 1.0.x
28. exportfs: /etc/exports [4]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/myserver".
29.   Assuming default behaviour ('no_subtree_check').
30.   NOTE: this default has changed since nfs-utils version 1.0.x
32. exportfs: /etc/exports [5]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/mysite".
33.   Assuming default behaviour ('no_subtree_check').
34.   NOTE: this default has changed since nfs-utils version 1.0.x
36. exportfs: /etc/exports [7]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/images".
37.   Assuming default behaviour ('no_subtree_check').
38.   NOTE: this default has changed since nfs-utils version 1.0.x
40. exportfs: /etc/exports [8]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/static".
41.   Assuming default behaviour ('no_subtree_check').
42.   NOTE: this default has changed since nfs-utils version 1.0.x
44. exportfs: /etc/exports [11]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/zookeeper-datadir-1".
45.   Assuming default behaviour ('no_subtree_check').
46.   NOTE: this default has changed since nfs-utils version 1.0.x
48. exportfs: /etc/exports [12]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/zookeeper-datadir-2".
49.   Assuming default behaviour ('no_subtree_check').
50.   NOTE: this default has changed since nfs-utils version 1.0.x
52. exportfs: /etc/exports [13]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/zookeeper-datadir-3".
53.   Assuming default behaviour ('no_subtree_check').
54.   NOTE: this default has changed since nfs-utils version 1.0.x
56. exportfs: /etc/exports [16]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/redis-datadir-1".
57.   Assuming default behaviour ('no_subtree_check').
58.   NOTE: this default has changed since nfs-utils version 1.0.x
60. exportfs: /etc/exports [18]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/redis0".
61.   Assuming default behaviour ('no_subtree_check').
62.   NOTE: this default has changed since nfs-utils version 1.0.x
64. exportfs: /etc/exports [19]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/redis1".
65.   Assuming default behaviour ('no_subtree_check').
66.   NOTE: this default has changed since nfs-utils version 1.0.x
68. exportfs: /etc/exports [20]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/redis2".
69.   Assuming default behaviour ('no_subtree_check').
70.   NOTE: this default has changed since nfs-utils version 1.0.x
72. exportfs: /etc/exports [21]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/redis3".
73.   Assuming default behaviour ('no_subtree_check').
74.   NOTE: this default has changed since nfs-utils version 1.0.x
76. exportfs: /etc/exports [22]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/redis4".
77.   Assuming default behaviour ('no_subtree_check').
78.   NOTE: this default has changed since nfs-utils version 1.0.x
80. exportfs: /etc/exports [23]: Neither 'subtree_check' or 'no_subtree_check' specified for export "*:/data/k8sdata/magedu/redis5".
81.   Assuming default behaviour ('no_subtree_check').
82.   NOTE: this default has changed since nfs-utils version 1.0.x
84. exporting *:/data/k8sdata/magedu/redis5
85. exporting *:/data/k8sdata/magedu/redis4
86. exporting *:/data/k8sdata/magedu/redis3
87. exporting *:/data/k8sdata/magedu/redis2
88. exporting *:/data/k8sdata/magedu/redis1
89. exporting *:/data/k8sdata/magedu/redis0
90. exporting *:/data/k8sdata/magedu/redis-datadir-1
91. exporting *:/data/k8sdata/magedu/zookeeper-datadir-3
92. exporting *:/data/k8sdata/magedu/zookeeper-datadir-2
93. exporting *:/data/k8sdata/magedu/zookeeper-datadir-1
94. exporting *:/data/k8sdata/magedu/static
95. exporting *:/data/k8sdata/magedu/images
96. exporting *:/data/k8sdata/mysite
97. exporting *:/data/k8sdata/myserver
98. exporting *:/pod-vol
99. exporting *:/data/volumes
100. exporting *:/data/k8sdata/kuboard
101. root@harbor:~#

复制代码

2.2.2、创建pv

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster# cat pv/redis-cluster-pv.yaml
2. apiVersion: v1
3. kind: PersistentVolume
4. metadata:
5.   name: redis-cluster-pv0
6. spec:
7.   capacity:
8.     storage: 5Gi
9.   accessModes:
10.     - ReadWriteOnce
11.   nfs:
12.     server: 192.168.0.42
13.     path: /data/k8sdata/magedu/redis0
15. ---
16. apiVersion: v1
17. kind: PersistentVolume
18. metadata:
19.   name: redis-cluster-pv1
20. spec:
21.   capacity:
22.     storage: 5Gi
23.   accessModes:
24.     - ReadWriteOnce
25.   nfs:
26.     server: 192.168.0.42
27.     path: /data/k8sdata/magedu/redis1
29. ---
30. apiVersion: v1
31. kind: PersistentVolume
32. metadata:
33.   name: redis-cluster-pv2
34. spec:
35.   capacity:
36.     storage: 5Gi
37.   accessModes:
38.     - ReadWriteOnce
39.   nfs:
40.     server: 192.168.0.42
41.     path: /data/k8sdata/magedu/redis2
43. ---
44. apiVersion: v1
45. kind: PersistentVolume
46. metadata:
47.   name: redis-cluster-pv3
48. spec:
49.   capacity:
50.     storage: 5Gi
51.   accessModes:
52.     - ReadWriteOnce
53.   nfs:
54.     server: 192.168.0.42
55.     path: /data/k8sdata/magedu/redis3
57. ---
58. apiVersion: v1
59. kind: PersistentVolume
60. metadata:
61.   name: redis-cluster-pv4
62. spec:
63.   capacity:
64.     storage: 5Gi
65.   accessModes:
66.     - ReadWriteOnce
67.   nfs:
68.     server: 192.168.0.42
69.     path: /data/k8sdata/magedu/redis4
71. ---
72. apiVersion: v1
73. kind: PersistentVolume
74. metadata:
75.   name: redis-cluster-pv5
76. spec:
77.   capacity:
78.     storage: 5Gi
79.   accessModes:
80.     - ReadWriteOnce
81.   nfs:
82.     server: 192.168.0.42
83.     path: /data/k8sdata/magedu/redis5
84. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster#

复制代码

2.3、部署redis cluster

2.3.1、基于redis.conf文件创建configmap

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster# cat redis.conf
2. appendonly yes
3. cluster-enabled yes
4. cluster-config-file /var/lib/redis/nodes.conf
5. cluster-node-timeout 5000
6. dir /var/lib/redis
7. port 6379
8. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster#

复制代码

2.3.2、创建configmap

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster# kubectl create cm redis-conf --from-file=./redis.conf -n magedu
2. configmap/redis-conf created
3. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster# kubectl get cm -n magedu
4. NAME               DATA   AGE
5. kube-root-ca.crt   1      35h
6. redis-conf         1      6s
7. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster#

复制代码

2.3.3、验证configmap

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster# kubectl describe cm redis-conf -n magedu
2. Name:         redis-conf
3. Namespace:    magedu
4. Labels:       <none>
5. Annotations:  <none>
7. Data
8. ====
9. redis.conf:
10. ----
11. appendonly yes
12. cluster-enabled yes
13. cluster-config-file /var/lib/redis/nodes.conf
14. cluster-node-timeout 5000
15. dir /var/lib/redis
16. port 6379
19. BinaryData
20. ====
22. Events:  <none>
23. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster#

复制代码

2.3.4、部署redis cluster

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster# cat redis.yaml
2. apiVersion: v1
3. kind: Service
4. metadata:
5.   name: redis
6.   namespace: magedu
7.   labels:
8.     app: redis
9. spec:
10.   selector:
11.     app: redis
12.     appCluster: redis-cluster
13.   ports:
14.   - name: redis
15.     port: 6379
16.   clusterIP: None
17.   
18. ---
19. apiVersion: v1
20. kind: Service
21. metadata:
22.   name: redis-access
23.   namespace: magedu
24.   labels:
25.     app: redis
26. spec:
27.   type: NodePort
28.   selector:
29.     app: redis
30.     appCluster: redis-cluster
31.   ports:
32.   - name: redis-access
33.     protocol: TCP
34.     port: 6379
35.     targetPort: 6379
36.     nodePort: 36379
38. ---
39. apiVersion: apps/v1
40. kind: StatefulSet
41. metadata:
42.   name: redis
43.   namespace: magedu
44. spec:
45.   serviceName: redis
46.   replicas: 6
47.   selector:
48.     matchLabels:
49.       app: redis
50.       appCluster: redis-cluster
51.   template:
52.     metadata:
53.       labels:
54.         app: redis
55.         appCluster: redis-cluster
56.     spec:
57.       terminationGracePeriodSeconds: 20
58.       affinity:
59.         podAntiAffinity:
60.           preferredDuringSchedulingIgnoredDuringExecution:
61.           - weight: 100
62.             podAffinityTerm:
63.               labelSelector:
64.                 matchExpressions:
65.                 - key: app
66.                   operator: In
67.                   values:
68.                   - redis
69.               topologyKey: kubernetes.io/hostname
70.       containers:
71.       - name: redis
72.         image: redis:4.0.14
73.         command:
74.           - "redis-server"
75.         args:
76.           - "/etc/redis/redis.conf"
77.           - "--protected-mode"
78.           - "no"
79.         resources:
80.           requests:
81.             cpu: "500m"
82.             memory: "500Mi"
83.         ports:
84.         - containerPort: 6379
85.           name: redis
86.           protocol: TCP
87.         - containerPort: 16379
88.           name: cluster
89.           protocol: TCP
90.         volumeMounts:
91.         - name: conf
92.           mountPath: /etc/redis
93.         - name: data
94.           mountPath: /var/lib/redis
95.       volumes:
96.       - name: conf
97.         configMap:
98.           name: redis-conf
99.           items:
100.           - key: redis.conf
101.             path: redis.conf
102.   volumeClaimTemplates:
103.   - metadata:
104.       name: data
105.       namespace: magedu
106.     spec:
107.       accessModes: [ "ReadWriteOnce" ]
108.       resources:
109.         requests:
110.           storage: 5Gi
111. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster#

复制代码

上述配置清单，主要用sts控制器创建了6个pod副本，每个副本都使用configmap中的配置文件作为redis配置文件，使用pvc模板指定pod在k8s上自动关联pv，并在magedu名称空间创建pvc,即只要k8s上有空余的pv，对应pod就会在magedu这个名称空间按pvc模板信息创建pvc；当然我们可以使用存储类自动创建pvc，也可以提前创建好pvc，一般情况下使用sts控制器，我们可以使用pvc模板的方式来指定pod自动创建pvc(前提是k8s有足够的pv可用)；

应用配置清单部署redis cluster

使用sts控制器创建pod，pod名称是sts控制器的名称-id，使用pvc模板创建pvc的名称为pvc模板名称-pod名称,即pvc模板名-sts控制器名-id;

2.4、初始化redis cluster

2.4.1、在k8s上创建临时容器,安装redis cluster 初始化工具

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster# kubectl run -it ubuntu1804 --image=ubuntu:18.04 --restart=Never -n magedu bash
2. If you don't see a command prompt, try pressing enter.
3. root@ubuntu1804:/#
4. root@ubuntu1804:/# apt update
5. # 安装必要工具
6. root@ubuntu1804:/# apt install python2.7 python-pip redis-tools dnsutils iputils-ping net-tools
7. # 更新pip
8. root@ubuntu1804:/# pip install --upgrade pip
9. # 使用pip安装redis cluster初始化工具redis-trib
10. root@ubuntu1804:/# pip install redis-trib==0.5.1
11. root@ubuntu1804:/#

复制代码

2.4.2、初始化redis cluster

1. root@ubuntu1804:/# redis-trib.py create \
2. `dig +short redis-0.redis.magedu.svc.cluster.local`:6379 \
3. `dig +short redis-1.redis.magedu.svc.cluster.local`:6379 \
4. `dig +short redis-2.redis.magedu.svc.cluster.local`:6379

复制代码

在k8s上我们使用sts创建pod，对应pod的名称是固定不变的，所以我们初始化redis 集群就直接使用redis pod名称就可以直接解析到对应pod的IP地址；在传统虚拟机或物理机上初始化redis集群，我们可用直接使用IP地址，原因是物理机或虚拟机IP地址是固定的，在k8s上pod的IP地址是不固定的；

2.4.3、给master指定slave

• 给redis-0指定slave为 redis-3
  

1. root@ubuntu1804:/# redis-trib.py replicate \
2. --master-addr `dig +short redis-0.redis.magedu.svc.cluster.local`:6379 \
3. --slave-addr `dig +short redis-3.redis.magedu.svc.cluster.local`:6379

复制代码

• 给redis-1指定slave为 redis-4
  

1. root@ubuntu1804:/# redis-trib.py replicate \
2. --master-addr `dig +short redis-1.redis.magedu.svc.cluster.local`:6379 \
3. --slave-addr `dig +short redis-4.redis.magedu.svc.cluster.local`:6379

复制代码

• 给redis-2指定slave为 redis-5
  

1. root@ubuntu1804:/# redis-trib.py replicate \
2. --master-addr `dig +short redis-2.redis.magedu.svc.cluster.local`:6379 \
3. --slave-addr `dig +short redis-5.redis.magedu.svc.cluster.local`:6379

复制代码

2.5、验证redis cluster状态

2.5.1、进入redis cluster 任意pod 查看集群信息


2.5.2、查看集群节点

集群节点信息中记录了master节点id和slave id，其中slave后面会对应master的id，表示该slave备份对应master数据；

2.5.3、查看当前节点信息

1. 127.0.0.1:6379> info
2. # Server
3. redis_version:4.0.14
4. redis_git_sha1:00000000
5. redis_git_dirty:0
6. redis_build_id:165c932261a105d7
7. redis_mode:cluster
8. os:Linux 5.15.0-73-generic x86_64
9. arch_bits:64
10. multiplexing_api:epoll
11. atomicvar_api:atomic-builtin
12. gcc_version:8.3.0
13. process_id:1
14. run_id:aa8ef00d843b4f622374dbb643cf27cdbd4d5ba3
15. tcp_port:6379
16. uptime_in_seconds:4303
17. uptime_in_days:0
18. hz:10
19. lru_clock:8272053
20. executable:/data/redis-server
21. config_file:/etc/redis/redis.conf
23. # Clients
24. connected_clients:1
25. client_longest_output_list:0
26. client_biggest_input_buf:0
27. blocked_clients:0
29. # Memory
30. used_memory:2642336
31. used_memory_human:2.52M
32. used_memory_rss:5353472
33. used_memory_rss_human:5.11M
34. used_memory_peak:2682248
35. used_memory_peak_human:2.56M
36. used_memory_peak_perc:98.51%
37. used_memory_overhead:2559936
38. used_memory_startup:1444856
39. used_memory_dataset:82400
40. used_memory_dataset_perc:6.88%
41. total_system_memory:16740012032
42. total_system_memory_human:15.59G
43. used_memory_lua:37888
44. used_memory_lua_human:37.00K
45. maxmemory:0
46. maxmemory_human:0B
47. maxmemory_policy:noeviction
48. mem_fragmentation_ratio:2.03
49. mem_allocator:jemalloc-4.0.3
50. active_defrag_running:0
51. lazyfree_pending_objects:0
53. # Persistence
54. loading:0
55. rdb_changes_since_last_save:0
56. rdb_bgsave_in_progress:0
57. rdb_last_save_time:1685992849
58. rdb_last_bgsave_status:ok
59. rdb_last_bgsave_time_sec:0
60. rdb_current_bgsave_time_sec:-1
61. rdb_last_cow_size:245760
62. aof_enabled:1
63. aof_rewrite_in_progress:0
64. aof_rewrite_scheduled:0
65. aof_last_rewrite_time_sec:-1
66. aof_current_rewrite_time_sec:-1
67. aof_last_bgrewrite_status:ok
68. aof_last_write_status:ok
69. aof_last_cow_size:0
70. aof_current_size:0
71. aof_base_size:0
72. aof_pending_rewrite:0
73. aof_buffer_length:0
74. aof_rewrite_buffer_length:0
75. aof_pending_bio_fsync:0
76. aof_delayed_fsync:0
78. # Stats
79. total_connections_received:7
80. total_commands_processed:17223
81. instantaneous_ops_per_sec:1
82. total_net_input_bytes:1530962
83. total_net_output_bytes:108793
84. instantaneous_input_kbps:0.04
85. instantaneous_output_kbps:0.00
86. rejected_connections:0
87. sync_full:1
88. sync_partial_ok:0
89. sync_partial_err:1
90. expired_keys:0
91. expired_stale_perc:0.00
92. expired_time_cap_reached_count:0
93. evicted_keys:0
94. keyspace_hits:0
95. keyspace_misses:0
96. pubsub_channels:0
97. pubsub_patterns:0
98. latest_fork_usec:853
99. migrate_cached_sockets:0
100. slave_expires_tracked_keys:0
101. active_defrag_hits:0
102. active_defrag_misses:0
103. active_defrag_key_hits:0
104. active_defrag_key_misses:0
106. # Replication
107. role:master
108. connected_slaves:1
109. slave0:ip=10.200.155.175,port=6379,state=online,offset=1120,lag=1
110. master_replid:60381a28fee40b44c409e53eeef49215a9d3b0ff
111. master_replid2:0000000000000000000000000000000000000000
112. master_repl_offset:1120
113. second_repl_offset:-1
114. repl_backlog_active:1
115. repl_backlog_size:1048576
116. repl_backlog_first_byte_offset:1
117. repl_backlog_histlen:1120
119. # CPU
120. used_cpu_sys:12.50
121. used_cpu_user:7.51
122. used_cpu_sys_children:0.01
123. used_cpu_user_children:0.00
125. # Cluster
126. cluster_enabled:1
128. # Keyspace
129. 127.0.0.1:6379>

复制代码

2.5.4、验证redis cluster读写数据是否正常？

2.5.4.1、手动连接redis cluster 进行数据读写

手动连接redis 集群master节点进行数据读写，存在一个问题就是当我们写入的key经过crc16计算对16384取模后，对应槽位可能不在当前节点，redis它会告诉我们该key该在哪里去写；从上面的截图可用看到，现在redis cluster 是可用正常读写数据的

2.5.4.2、使用python脚本连接redis cluster 进行数据读写

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster# cat redis-client-test.py
2. #!/usr/bin/env python
3. #coding:utf-8
4. #Author:Zhang ShiJie
5. #python 2.7/3.8
6. #pip install redis-py-cluster
8. import sys,time
9. from rediscluster import RedisCluster
10. def init_redis():
11.     startup_nodes = [
12.         {'host': '192.168.0.34', 'port': 36379},
13.         {'host': '192.168.0.35', 'port': 36379},
14.         {'host': '192.168.0.36', 'port': 36379},
15.         {'host': '192.168.0.34', 'port': 36379},
16.         {'host': '192.168.0.35', 'port': 36379},
17.         {'host': '192.168.0.36', 'port': 36379},
18.     ]
19.     try:
20.         conn = RedisCluster(startup_nodes=startup_nodes,
21.                             # 有密码要加上密码哦
22.                             decode_responses=True, password='')
23.         print('连接成功！！！！！1', conn)
24.         #conn.set("key-cluster","value-cluster")
25.         for i in range(100):
26.             conn.set("key%s" % i, "value%s" % i)
27.             time.sleep(0.1)
28.             data = conn.get("key%s" % i)
29.             print(data)
31.         #return conn
33.     except Exception as e:
34.         print("connect error ", str(e))
35.         sys.exit(1)
37. init_redis()
38. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster#

复制代码

运行脚本，向redis cluster 写入数据

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster# python redis-client-test.py
2. Traceback (most recent call last):
3.   File "/root/k8s-data/yaml/magedu/redis-cluster/redis-client-test.py", line 8, in <module>
4.     from rediscluster import RedisCluster
5. ModuleNotFoundError: No module named 'rediscluster'
6. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster#

复制代码

这里提示没有找到rediscluster模块，解决办法就是通过pip安装redis-py-cluster模块即可；

安装redis-py-cluster模块

运行脚本连接redis cluster进行数据读写

连接redis pod，验证数据是否正常写入？

从上面的截图可用看到三个reids cluster master pod各自都存放了一部分key,并非全部；说明刚才我们用python脚本把数据正常写入了redis cluster;

验证在slave 节点是否可用正常读取数据？

从上面的截图可以了解到在slave节点是不可以读取数据；

到slave对应的master节点读取数据

上述验证说明了redis cluster 只有master可以读写数据，slave只是对master数据做备份，不可以在slave上读写数据；

2.6、验证验证redis cluster高可用

2.6.1、在k8s node节点将redis:4.0.14镜像上传至本地harbor

• 修改镜像tag
  

1. root@k8s-node01:~# nerdctl tag redis:4.0.14 harbor.ik8s.cc/redis-cluster/redis:4.0.14

复制代码

• 上传redis镜像至本地harbor
  

1. root@k8s-node01:~# nerdctl push harbor.ik8s.cc/redis-cluster/redis:4.0.14
2. INFO[0000] pushing as a reduced-platform image (application/vnd.docker.distribution.manifest.list.v2+json, sha256:1ae9e0f790001af4b9f83a2b3d79c593c6f3e9a881b754a99527536259fb6625)
3. WARN[0000] skipping verifying HTTPS certs for "harbor.ik8s.cc"
4. index-sha256:1ae9e0f790001af4b9f83a2b3d79c593c6f3e9a881b754a99527536259fb6625:    done           |++++++++++++++++++++++++++++++++++++++|
5. manifest-sha256:5bd4fe08813b057df2ae55003a75c39d80a4aea9f1a0fbc0fbd7024edf555786: done           |++++++++++++++++++++++++++++++++++++++|
6. config-sha256:191c4017dcdd3370f871a4c6e7e1d55c7d9abed2bebf3005fb3e7d12161262b8:   done           |++++++++++++++++++++++++++++++++++++++|
7. elapsed: 1.4 s                                                                    total:  8.5 Ki (6.1 KiB/s)                                       
8. root@k8s-node01:~#

复制代码

2.6.2、修改redis cluster部署清单镜像和镜像拉取策略

修改镜像为本地harbor镜像和拉取策略是方便我们测试redis cluster的高可用；

2.6.3、重新apply redis cluster部署清单

1. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster# kubectl apply -f redis.yaml
2. service/redis unchanged
3. service/redis-access unchanged
4. statefulset.apps/redis configured
5. root@k8s-master01:~/k8s-data/yaml/magedu/redis-cluster#

复制代码

这里相当于给redis cluster更新，他们之间的集群关系还存在，因为集群关系配置都保存在远端存储之上；

• 验证pod是否都正常running?
  
  
• 验证集群状态和集群关系
  
  

不同于之前，这里rdis-0变成了slave ,redis-3变成了master；从上面的截图我们也发现，在k8s上部署redis cluster pod重建以后（IP地址发生变化），对应集群关系不会发生变化；对应master和salve一对关系始终只是再对应的master和salve两个pod中切换，这其实就是高可用；

2.6.4、停掉本地harbor，删除redis master pod，看看对应slave是否会提升为master？

• 停止harbor服务
  

1. root@harbor:~# systemctl stop harbor

复制代码

• 删除redis-3,看看redis-0是否会提升为master？
  
  

可用看到我们把redis-3删除（相当于master宕机）以后，对应slave提升为master了；

2.6.5、恢复harbor服务，看看对应redis-3恢复会议后是否还是redis-0的slave呢？

• 恢复harbor服务
  
  
• 验证redis-3pod是否恢复？
  
  

再次删除redis-3以后，对应pod正常被重建，并处于running状态；

• 验证redis-3的主从关系、
  
  

可以看到redis-3恢复以后，对应自动加入集群成为redis-0的slave;

        出处：https://www.cnblogs.com/qiuhom-1874/        本文版权归作者和博客园共有，欢迎转载，但未经作者同意必须保留此段声明，且在文章页面明显位置给出原文连接，否则保留追究法律责任的权利.
免责声明：如果侵犯了您的权益，请联系站长，我们会及时删除侵权内容，谢谢合作！