Distributed design - Cluster

• A plurality of nodes together to save data
• effect
  • Extended storage space
  • Increase throughput, improve write performance
• And different points of single
  • No longer distinguish between databases, library only 0, stand-alone default 0-15
  • Does not support transactions / pipeline / multi-value operation
• Feature
  • It requires at least three leading and three from
  • Requirements must be turned AOF persistence
  • Automatically selected cluster nodes for storage
  • Sentinel default integrated, automatic failover
• Configuration

    # 每个节点分别配置ip/端口
    bind 127.0.0.1
    port 6379
    # 集群配置
    cluster-enabled yes   # 开启集群
    cluster-config-file nodes-7000.conf  # 节点日志文件
    cluster-node-timeout 15000  # 节点超时时长 15秒
    # 开启AOF 及相关配置  
    appendonly yes  

• Create a cluster
  • redis installation package contains redis-trib.rb, Use to create a cluster
  • After you create a cluster, restart the cluster starts automatically redis

    # 将命令复制，这样可以在任何⽬录下调⽤此命令
    sudo cp /usr/share/doc/redis-tools/examples/redis-trib.rb /usr/local/bin/
    # 安装ruby环境，因为redis-trib.rb是⽤ruby开发的
    sudo apt-get install ruby
    gem sources --add https://gems.ruby-china.com/ --remove https://rubygems.org/
    sudo gem install redis
  
    # 启动主从数据库 7000-7005.conf
    sudo redis-server 7000.conf
    ...
  
    # 创建集群
    redis-trib.rb create --replicas 1 127.0.0.1:7000 127.0.0.1:7001 127.0.0.1:7002 127.0.0.1:7003 127.0.0.1:7004 127.0.0.1:7005
    # 访问集群  访问集群必须加-c选项, 否则无法进行读写操作
    redis-cli -p 7000 -c

• redis clusters can not support the transaction and WATCH, pessimistic concurrency control can only design your own lock (setnx)

• python operation redis clusters
  • installation pip install redis-py-cluster
    
• Use project
  • Sentinel and the master-slave configuration persistence mechanism for saving user's reading history / search history
  • Cluster is responsible for implementing cache design