Skip to main content

redis

Overviw of Redis​

Redis is an open source. BSD licensed, advanced key-value cache and tore. It is ofternreferred to as a data structure server since keys can contain string, hashes, lits, sets, sorted sets, bitmaps and hyperloglogs.

  • The word Redis means REmote DIctionary Server
  • Initial release in 2009
  • It is an advanced key-value store or a data structure store
  • Runs entirely in memory
    • All data is kept in memory
    • Quick dta access since it is maintained in memory
    • Data can be backed up to disk periodically
    • Single threaded server
  • Extensible via Lua Scripts
  • Able to replicate data between servers
  • Clustering also available(v3.0+)

Redis features​

  • KV cache and store

    • Redis is an in-memory but persisten on disk datase
      • 1 Million small Key -> String value pairs use ~ 100 MB of memory
      • Single threaded - but CPU should not be the bottleneck
        • Average Linux system can deliver even 500k(50万并发) requests per second
      • Limit is likely the available memory in your system
        • max 323 keys
    • Persistence,持久化
      • Snapshotting
        • Data is asynchronously transferred from memory to disk
      • AOF(Append Only File)
        • Each modifying operation is written to a file
        • Can recreae data store by replaying operations
        • Without interrupting service, will rebuild AOF as the shortest sequence of commands needed to rebuild the current dataset in memory.
    • replication 主从(借助于sentinel实现一定意义上的 HA)
      • Redis supports master-slave replication
      • Master-slave replication can be chained
      • Be carefull
        • Slaved are writeable
        • Potential for data inconsistency
      • Fully compatible with Pub/Sub features
    • clustering(分布式)

  • Data Structure Server

    • String
    • List
    • Hash(关联数组)
    • Set
    • Sorted Set
    • Bitmaps
    • Hyperloglogs

Redis VS Memcache​

  • Memcached is a "distributed memory object caching system"
  • Redis persis data to disk eventually
  • Memcached in a LRU cache
  • Redis has difference data types and more features
  • Memcached is multithreaded
  • Similar speed

Redis 优势​

  • 丰富的（资料形态）操作
    • Hashes, Lists, Sets, Sorted Sets, HyperLogLog 等
  • 内建 replication 及 cluster
  • 就地更新(in-place update)操作
  • 支持持久化（磁盘）
    • 避免雪崩效应

Memcache 优势​

  • 多线程
    • 善用多核 CPU
    • 更少的阻塞操作
  • 更少的内存开销
  • 更少的内存分配压力
  • 可能有更少的内存碎片

Prominent Adopters​

  • Twitter
  • Pinterest
  • Tumblr
  • Github
  • Stack Overflow
  • digg
  • Blizard
  • flickr
  • Weibo

Redis 3.0​

  • 2015-4-1 release
  • Redis Cluster
  • 新的 “embedded string”
  • LRU 演算法的改进
    • 预设随机取5个样本，插入并排序至一个 pool, 移除最佳者，如此反复，知道内存用最小于 maxmemory 的设定
    • 样本 5 比先前 3 多
    • 从局部优化趋向全局最优

Redis 特性​

存储系统​

  • RDBMS：Oracle, DB2, MySQL, SQL Server
  • NoSQL: HBase, Memcached, MongoDB, Redis
  • NewSQL(分布式): Aerospike, FoundationDB, RethinkDB

NoSQL​

  • key-value NoSQL
    • Memcached, Redis
  • Column family NoSQL
    • Cassandra, HBase
  • Documatation NoSQL
    • MongoDB
  • Graph NoSQL
    • Neo4j

Redis 组件​

  • 早期只有10k行代码（作者：意大利人）

  • redis.io

  • redis-server

  • redis-clie

    • command line interface

  • redis-benchmark 性能压力测试工具

    • benchmark utility

  • redis-check-dump & redis-check-aof

    • corrupted RDB/AOF files utilities

epel 源
# yum info redis

www.redis.io

pkgs.org

redis-3.0.0.2-1.el6....

# rpm -i redis-3...
# rpm -ql redis

# cp /etc/redis.conf{,bak}
# vim /etc/redis.conf
  daemonize no 守护进程
  pidfile /var/run/redis.pid
  port 6379 监听端口
  tcp-backlog 511 等待队列
  bind 127.0.0.1 [172.16.100.6]
  unixsocketperm 700 本地socket
  timeout 0 客户端连接不超时
  tcp-keepalive 0
  loglevel notice
  logfile /var/log/redis/redis.log
  databases 16 支持16个数据库，集群仅支持 0 号数据库

  save 900 1 : 900秒1个记录变化一次快照
  save 300 10 : 300秒内10个记录发生变化一次快照
  save 60 10000 : 60秒内10000个记录发生变化一次快照

  maxmemory 最大使用内存


# service redis start
# ss -tnl | grep 6379

# redis-cli -h
# redis-cli -h 172.16.100.6

# redis-cli

> help

> help @STRING
> help APPEND
> help @transaction
> help @server

> CLIENT LIST

Data Type​

Keys​

  • Arbitrary ASCII strings
    • Define some format convention and adhere to it
    • Key length matters
  • Multiple name spaces are available
    • Separate DBs indexed by an integer value
      • select command
      • multiples DBs vs. Single DB + key prefixes
  • Keys can expire automatically
名称空间或数据库会解决内存使用量
> select 0
> select 2


Strings​

> help set
> SET disto fedora
> GET disto
"fedora"
> SET disto centos
> GET disto
"centos"
> append disto slackware
(integer) 15
> GET disto
"centosslackware"

> help @string

> STRLEN disto
(integer) 15

> SET count 0
OK

> INCR count
(integer) 1

> INCR count
(integer) 2

> INCR count
(integer) 3

> DECR count
(integer) 2

> DECR count
(integer) 1

> DECR count
(integer) 0

> DECR count
(integer) -1

自动过期
> HELP SET
key seconds value

如果键不存在，则可以设置键
> SET disto gentoo NX
(nil)

> GET disto
"centosslackware"

如果键存在，则可以设置键
> SET foo bar XX
(nil)

键是否存在
> EXISTS foo

List​

['A','B','C']

> help @list
> rpush 向右添加
> lpush 向前添加
> lpop 左边弹出元素
> rpop 右边弹出元素

> LPUSH L0  mon
(integer) 1
> LINDEX L0 0
"mon"
> LPUSH L0 sun
(integer) 2
> LINDEX L0 0
"sun"
> LINDEX L0 1
"mon"
> RPUSH L0 tue
(integer) 3
> LINDEX L0 2
"tue"
> LSET L0 1 fri
> LINDEX L0 1
> RPOP L0
> RPOP L1


Set​

无序数据结构

(A,B,C,D)

> HELP @SET
> SADD w1 mon tue ved thu fri sat sun
(integer) 7
> SADD w2 tue thu day
(integer) 3 

> SINTER w1 w2 交集
1) "thu"
2) "tue"

> SUNION w1 w2 并集
1) "thu"
2) "tue"
3) "sat"
4) "mon"
5) "thu"
6) "wed"
7) "sun"
8) "day"

> SPOP w1 随机弹出
"fre"

> SISMEMBER w1 fri 是否存在元素

Sorted Set​

有序集合

{C:1, D:2, A:3, B:4}

> HELP @SORTED_SET

key score memer : score 排序分数

> ZADD weekday1 1 mon 2 tue 3 wed
> ZCARD weekday1
(integer) 3
> ZRANK weekday1 tue 索引号
> ZRANK weekday1 tue 索引号

> ZSCORE weekday2 tue 根据 score 获取索引
"2"

> ZRANGE weekday1 0 2 起始索引和结束索引

Hash​

{key:value}

{field1: "A", field2:"B"}

> HELP @HASH

> HSET h0 a mon
(integer) 1
> HGET h0 a
> HSET h0 b tue
(integer) 1
> HGET h0 a
"mon"
> HGET h0 b
"tue"

> HKEYS h0 所有键名
1) "a"
2) "b"

> HVALS h0 所有名

> HLEN h0
(integer) 2

Bitmpas​

用于数据统计

HyperLogLog​

认证实现方法​

# vim redis.conf
  requirepass 123456
# service redis restart
# redis-cli -h 172.16.100.5
> SELECT 0
> AUTH 123456
> exit

清空数据库​

> FLUSHDB: 清空当前数据库
> FLUSHALL: 清空所有库

事务​

通过 MULTI, EXEC, WATCH 等命令实现事务功能；将一个或多个命令归并为一个操作提请服务器按顺序执行的机制；不支持回滚操作

执行事务的时候，客户端等待执行事务的完之后才会接受其他客户端请求

  • MULTI：启动一个事务

  • EXEC：执行事务；一次性将事务中的所有操作执行完成返回给客户端

  • WATCH: 乐观锁，在 EXEC 命令执行之前，用于监视指定数量键，如果监视中的某任意键数据被修改，则服务器拒绝执行事务；

> MULTI
> SET ip 192.168.1.1
> GET ip
> SET port 8080
> GET port
> EXEC

watch 监视​

第一步

# redis-cli
> WATCH ip
> MULTI
> SET ip 10.0.0.1
> GET ip

第二步

# redis-cli
> get ip
> SET ip 172.16.100.99

第三步

> EXEC
(nil)

Connectin 相关命令​

  • AUTH
  • ECHO
  • PING
  • QUIT
  • SELECT
AUTH
> help ping
> ECHO "hello world"
> HELP @connection
> HELP @server 服务器常用命令

Server 相关的命令​

  • CLIENT SETNAME localconn
  • CLIENT GETNAME
  • CLIENT KILL ip:port