# Redis部署
#### 登录云服务器ECS
创建命名空间
```
# kubectl create ns applife
# kubectl create ns daas
# kubectl create ns daas-web
# kubectl create ns kong
# kubectl create ns openkube
# kubectl create ns nfs
```
#### RedisGraph安装
安装路径约定，安装机器：中间件服务器
安装目录：/opt/redis-6.0.0
配置文件：/opt/redis-6.0.0/config/redis.conf 和 /opt/redis-6.0.0/config/sentinel/sentinel.conf
数据目录：/data/redis-6.0.0/data
日志目录：/data/redis-6.0.0/logs
插件目录：/opt/redis-6.0.0/modules/redisgraph.so
#### 安装Redis主从
1. 解压，上传压缩包redis-6.0.0-linux.tar.gz到/soft/redis并解压
   ```
   # tar -vxf redis-6.0.0-linux.tar.gz
   # mv redis-6.0.0 /opt
   # mkdir /opt/redis-6.0.0/bin
   # cd /opt/redis-6.0.0/src
   # cp redis-benchmark redis-check-aof redis-check-rdb redis-cli redis-sentinel redis-server -t ../bin/
   ```
   
2. 创建数据目录及授权
   ```
   # mkdir -p /data/redis-6.0.0/{data,logs}
   # chmod 755 /opt/redis-6.0.0/modules/redisgraph.so
   ```
   ![](https://support.huaweicloud.com/deepexidip-dma/public_sys-resources/note_3.0-zh-cn.png)
   redisgraph.so要授权可执行，否则会导致redis启动失败。
   
3. 配置环境变量
   ```
   # vim /etc/profile
   export REDIS_HOME=/opt/redis-6.0.0
   export PATH=$REDIS_HOME/bin:$PATH
   ```
   图1配置环境变量   
   ![](https://support.huaweicloud.com/deepexidip-dma/zh-cn_image_0000001862232852.png) 
4. 新增主redis配置文件
   ![](https://support.huaweicloud.com/deepexidip-dma/public_sys-resources/note_3.0-zh-cn.png)
   /opt/redis-6.0.0/config/redis.conf文件不存在，注意文件不要有特殊空格字符，踩过坑！
   ```
   # vim /opt/redis-6.0.0/config/redis.conf
   bind 0.0.0.0
   port 6379
   daemonize yes
   pidfile /var/run/redis.pid
   loglevel notice
   logfile /data/redis-6.0.0/logs/redis.log
   dir /data/redis-6.0.0/data
   masterauth Mypwd@123456
   requirepass Mypwd@123456
   appendonly yes
   loadmodule /opt/redis-6.0.0/modules/redisgraph.so
   # 确保有1个从节点写入，且延时不超过10s，否则主节点会停止写入请求（防止数据丢失）
   min-replicas-to-write 1
   min-replicas-max-lag 10
   ```
   
5. 新增从redis配置文件 在安装redis-slave机器上，重复上面5.4.2.1 到5.4.2.4过程，并使用下面配置。
   ![](https://support.huaweicloud.com/deepexidip-dma/public_sys-resources/note_3.0-zh-cn.png)
   如果/opt/redis-6.0.0/config/redis.conf文件不存在就新增。
   ```
   # vim /opt/redis-6.0.0/config/redis.conf
   bind 0.0.0.0
   port 6379
   daemonize yes
   pidfile /var/run/redis.pid
   loglevel notice
   logfile /data/redis-6.0.0/logs/redis.log
   dir /data/redis-6.0.0/data
   masterauth Mypwd@123456
   requirepass Mypwd@123456
   appendonly yes
   loadmodule /opt/redis-6.0.0/modules/redisgraph.so
   # 确保有1个从节点写入，且延时不超过10s，否则主节点会停止写入请求（防止数据丢失）
   min-replicas-to-write 1
   min-replicas-max-lag 10
   #配置连接主redis
   replicaof 192.168.0.82 6379
   ```
   
6. 启动redis-server 分别在主从创建redis-server系统启动文件
   ```
   # vim /etc/systemd/system/redis-server.service
   [Unit]
   Description=redis
   After=network.target
   [Service]
   Type=forking
   PIDFile=/var/run/redis.pid
   ExecStart=/opt/redis-6.0.0/bin/redis-server /opt/redis-6.0.0/config/redis.conf
   ExecStop=/opt/redis-6.0.0/bin/redis-cli -h 127.0.0.1 -p 6379 shutdown
   PrivateTmp=true
   [Install]
   WantedBy=multi-user.target
   ```
   
7. 重新加载systemctl配置
   ```
   # systemctl daemon-reload
   ```
   
8. 设置跟随linux启动，并启动es
   ```
   # systemctl start redis-server && systemctl enable redis-server
   ```
   
9. 查看redis角色
   ```
   # redis-cli -h 192.168.0.82 -p 6379 -a Mypwd@123456 INFO | grep role
   ```
   
#### 配置哨兵
1. 创建哨兵数据目录
   ```
   # mkdir /data/redis-6.0.0/data/sentinel
   ```
   
2. 新增配置文件
   ```
   # vim /opt/redis-6.0.0/config/sentinel/sentinel.conf
   ```
   
3. 分别在主从redis创建sentinel配置文件（3个实例一样）
   ```
   port 16379
   daemonize yes
   pidfile /var/run/redis-sentinel.pid
   logfile /data/redis-6.0.0/logs/sentinel.log
   dir /data/redis-6.0.0/data/sentinel
   sentinel deny-scripts-reconfig yes
   #配置主redis的地址,“2”表示集群中有2个Sentinel认为master退出了就切换
   sentinel monitor redisMaster 192.168.0.82 6379 2
   sentinel auth-pass redisMaster Mypwd@123456
   sentinel config-epoch redisMaster 1
   sentinel leader-epoch redisMaster 1
   ```
   
4. 启动哨兵，分别在主从redis创建redis-sentinel系统启动文件
   ```
   # vim /etc/systemd/system/redis-sentinel.service
   [Unit]
   Description=The redis-sentinel Process Manager
   After=syslog.target network.target
   [Service]
   Type=forking
   PIDFile=/var/run/redis-sentinel.pid
   ExecStart=/opt/redis-6.0.0/bin/redis-sentinel /opt/redis-6.0.0/config/sentinel/sentinel.conf
   ExecStop=/opt/redis-6.0.0/bin/redis-cli -h 127.0.0.1 -p 16379 shundown
   PrivateTmp=true
   [Install]
   WantedBy=multi-user.target
   ```
   
5. 重新加载systemctl配置
   ```
   # systemctl daemon-reload
   ```
   
6. 设置跟随linux启动，并启动es
   ```
   # systemctl start redis-sentinel && systemctl enable redis-sentinel
   ```
   
7. 查看启动日志
   ```
   # cat /data/redis-6.0.0/logs/sentinel.log
   ```
   图2查看启动日志   
   ![](https://support.huaweicloud.com/deepexidip-dma/zh-cn_image_0000001862073028.png) 
8. 检查redis主从状态
   ```
   # redis-cli -h 127.0.0.1 -p 6379 -a Mypwd@123456 info replication
   ```
   图3查看启动日志   
   ![](https://support.huaweicloud.com/deepexidip-dma/zh-cn_image_0000001908072465.png)
   ```
   # Replication
   role:master          #角色master或者slave
   connected_slaves:2    #连接的slave个数
   min_slaves_good_slaves:2
   slave0:ip=192.168.0.228 ,port=6379,state=online,offset=288218,lag=0   #slave1节点信息
   slave1:ip=192.168.0.56,port=6379,state=online,offset=288078,lag=1   #slave2节点信息
   master_replid:2ac8c63f0185a38189010b192e42b761e69f549a
   master_replid2:4490ff668a9d41d5636e6a382ff1e80ca31892cf
   master_repl_offset:288218
   master_repl_meaningful_offset:288218
   second_repl_offset:69703
   repl_backlog_active:1
   repl_backlog_size:1048576
   repl_backlog_first_byte_offset:1
   repl_backlog_histlen:288218
   ```