A, Redis Overview
Redis is an open source (BSD follow protocol) Key-Value memory storage system data structure, as a database, cache and message broker. It supports five data structures: strings string, hash hash, list of list, an ordered collection set and collection zset. Redis Lua scripting support, high availability and clustering mechanism Sentinel. Applicable scene: cache, voting, sweepstakes, distributed session, charts, counting, queue, subscription publishing and so on; the specific description, see Redis official website .
Two, Redis installed
① Download: https: //redis.io/download
② install gcc: yum install gcc
③ to download a good first step redis-5.0.2.tar.gz uploaded to the server / root / svr / packages directory or direct cd / root / svr / packages and then executed: wget http://download.redis.io /releases/redis-5.0.2.tar.gz
④ execution cp redis-5.0.2.tar.gz ../
⑤ CD / the root / SVR then execute: tar -xvf redis-5.0.2.tar.gz:
cd-repeat 5.0.2:
⑥ 执行：make install PREFIX=/root/svr/redis-5.0.2
⑦ start redis execution: bin / redis-server ../redis.conf ( Note: If you need to start background redis.conf configuration inside daemonize changed to yes )
⑧ verify the successful launch ps -ef | grep redis
⑨ into the redis client: bin / redis-cli
⑩ Exit Client: quit
Three, redis.conf main configuration in detail
| parameter | Explanation |
| bind | Redis specified only receives requests from the IP address, if not set, then process all requests |
| port | Listening port, default 6379 |
| timeout | When the timeout setting client connection seconds. When the client does not issue any command during this time, then close the connection |
| daemonize | By default, redis is not running in the background, if you need to run in the background, to change the value of the item to yes |
| loglevel | log divided into three grades 4, debug, verbose, notice, and warning. Production environment in general turned notice |
| logfile | Configuration log file address, the standard default output, i.e., printing on the window displayed on the terminal |
| save | save <seconds> <changes> value of at least 10,000 key changes which come within the meaning save 60 10000 example 60 seconds (1 minute) (the database stored - persistent RDB) |
| dbfilename | Rdb file name |
| to you | Data directory, two kinds of persistence rdb, aof files in this directory |
| replicaof | replicaof <masterip> <masterport>: The master-slave configuration is the configuration represented example is redis masterip: masterport from node |
| masterauth | master password connection |
| replica-serve-stale-data | When the slave master with lost connections or data is being synchronized, slave run in two ways: 1) If the replica-serve-stale-data set yes (the default setting), slave will continue to respond to client requests. 2) If the replica-serve-stale-data set to no, will remove any request other than the specified command returns an error "SYNC with master in progress" |
| replica-read-only | Whether to set a read-only slave |
| repl-diskless-sync | Synchronization strategy: disk or socket, the default disk mode |
| repl-diskless-sync-delay | If a non-synchronized disk mode is turned on, you can configure the synchronization delay time to wait for a child process to produce master socket transmission by RDB data to the slave. The default value is 5 seconds, is set to 0.0 seconds per transmission without delay |
| repl-ping-replica-period | ping requests to slave to master according to the specified interval. 10 seconds by default |
| repl-timeout | Synchronization timeout |
| repl-disable-tcp-nodelay | Whether disabled after the slave socket send SYNC TCP_NODELAY |
| repl-backlog-size | backlog size setting data backup |
| repl-backlog-ttl | After disconnecting slave starts counting the number of seconds, backlog buffer will be released |
| replica-priority | slave priority, when linked to a master, a smaller priority number, priority will be given to enhance the salve is master, 0 as a special priority, the slave can not be identified as a master |
| requirepass | The client should be password validation when processing any command |
| rename-command | Rename command, you can change the name to dangerous command |
| maxclients | Setting a maximum number of clients simultaneously connected, the default limit is 10,000 clients. |
| maxmemory | Set the maximum memory, once the memory usage reaches the maximum memory, redis deletes the key according to the selected recovery strategy (maxmemmory-policy) |
| maxmemory-policy | 最大内存策略：如果达到内存限制了，redis如何选择删除key: |
| maxmemory-samples | 设置样本量的个数 |
| appendonly | 是否开启AOF，如果开启那么在启动时Redis将加载AOF文件，它更能保证数据的可靠性，aof的文件内容就是RESP协议 |
| appendfilename | AOF文件名（默认："appendonly.aof"） |
| appendfsync | 配置 Redis 多久才将数据 fsync 到磁盘一次 |
| auto-aof-rewrite-percentage | 自动重写AOF文件。如果AOF日志文件增大到指定百分比，默认100。Redis能够通过 BGREWRITEAOF 自动重写AOF日志文件 |
| auto-aof-rewrite-min-size | 自动重写AOF文件。如果AOF日志文件到达最小的指定大小，默认64mb |
| aof-use-rdb-preamble | Redis 4.0之后配置混合持久化，需要配置 aof-use-rdb-preamble yes |
| lua-time-limit | Lua脚本的最大执行时间，单位为毫秒 |
| cluster-enabled | 是否开启集群 cluster-enabled yes |
| cluster-config-file | redis自动生成集群配置信息的文件名 |
| cluster-node-timeout | 集群节点超时毫秒数。超时的节点将被视为不可用状态。 |
| aof-rewrite-incremental-fsync | 当一个子进程重写AOF文件时，如果配置aof-rewrite-incremental-fsync yes，则文件每生成32M，数据会被同步 |
| rdb-save-incremental-fsync | 当redis保存RDB文件时，如果启用了以下选项，每生成32MB数据，文件将被fsync到磁盘 |
四、Redis核心数据结构
五、Redis核心原理
1、Redis单线程为什么还能这么快
Redis所有的数据都是在内存中，所有的运算都是内存级别的运算，而且单线程避免了多线程的切换性能损耗的问题。正因为Redis是单线程，所以要小心使用Redis指令，对于那些耗时的指令(比如keys)，一定要谨慎使用，一不小心就可能会导致 Redis 卡顿。
2、Redis单线程为何能处理那么多的并发客户端连接
Redis的IO多路复用：redis利用epoll来实现IO多路复用，将连接信息和事件放到队列中，依次放到文件事件分派器，事件分派器将事件分发给事件处理器。（IO多路复用在后续的netty系列里面详细讲解）
总结：Redis是一个Key-Value数据结构的内存存储系统，他支持5种数据结构，分别是String结构、Hash结构、List结构、Set结构、Zset结构；Redis支持Lua脚本，哨兵机制和集群实现高可用；介绍了安装Redis的步骤，详细的解释了redis.conf的主要配置，以及Redis的核心原理。