Redis深入浅出
- 1. Redis 深入浅出 Http://www.hoterran.info Http://Weibo.com/hoterran
- 2. agenda use feature architecture admin replication presistence misc
- 3. use Download http://redis.googlecode.com/files/redis-2.2.11.tar.gz install Redis-server > /dev/null & Redis-cli http://try.redis-db.com/
- 4. use set/get lpush/lrem/rpop/ hset/hgetall sadd/srem multi/discard/exec
- 5. feature Advanced key value store like memcachedb single threaded, no lock Event demultiplexer, no use libevent 2w lines writen in c without 3 rd library faster
- 6. feature Difference with memached Data structure No only memcached k-v Linklist k-v1->v2->v3->v4... Set k - (v1,v2,v3,v4) Hash k – (k1 -v1, k2 – v2, k3 – v3) Need design a little Replication Presistence not support LRU Need multi instance use multi core
- 7. feature Difference with MySQL Not Heap table store Not need index, so not execute plan Not compute statistic No sql No join Capacity size limited by cache a little design before dev High rw
- 8. architecture
- 9. architecture
- 10. admin Redis.conf parameter intro Replication intro Persistence intro
- 11. replication procedure map Dual master mode? Master slave slave mode defects Without resume broken transfer 可以架两级 slave 解决 ? Without lag(slave position)
- 12. replication
- 13. replication Case replication switch old redis version running on, say, redis:6379. Install and start a new redis on redis:6380 with a different dump file location. Execute SLAVE OF redis 6379 against redis:6380. Wait for first SYNC to complete. echo "enable server redis/redis-6380" | socat stdio unix-connect:/var/run/haproxy/admin.sock echo "disable server redis/redis-6379" | socat stdio unix-connect:/var/run/haproxy/admin.sock Execute SLAVE OF no one on redis:6380. Execute SHUTDOWN on redis:6379. Assuming that you've got a HAProxy config looking something like listen redis :6378 server redis-6379 localhost:6379 check weight 256 server redis-6380 localhost:6380 check disabled
- 14. persistence Snapshot Fork process, loop hash table, save on file dump.rdb Yes!!! sequential write Write dump.rdb need O_DIRECT? Aof Like binlog, for recover after crash appendfsync Aof file is big use bgrewriteaof
- 15. Login user Design a login user system Heap table userid login_times last_login_time 1 5 2011-1-1 2 1 2011-1-2 3 2 2011-1-3 Last login man? Max login man?
- 16. Login user data Set userid:1:login_times 5 Set userid:2:login_times 1 Set userid:3:login_times 2 Set userid:1:last_login 2011-1-1 Set userid:2:last_login 2011-1-2 Set userid:3:last_login 2011-1-3 Last login lpush user_last_login 1 lpush user_last_login 2 lpush user_last_login 3 ltrim user_last_login 0 1
- 17. Login user Max login man zadd user:login_times 5 1 zadd user:login_times 1 2 zadd user:login_times 2 3 zcard user:login_times zrangebyscore user:login_times 3 +inf withscores Column store data?
- 18. tag Relational is suitable? book id book author 1 <<Diving into Python>> gun 2 <<Ruby on rail>> japan Tag Tagname id Web 1 Web 2 Python 2 Sql select id from tag t1, tag t2 where t1.tagname = 'web' and t2.tagname = 'python' and t1.id = t2.id
- 19. tag Kv design set book:1 ”diving into python” Set book:2 ”ruby on rail” sadd tag:web 2 sadd tag:web 1 sadd tag:python 1 sinter 'tag:web' 'tag:python' sunion tag:web 'tag:python' sdiff tag:web 'tag:python' 反思？
- 20. Sina weibo Recent record List, Lrange( 关注的人列表、粉丝列表、 feeds 列表 ) Set Sadd|, sdiff, sinter ( 我的好友，我可能感兴趣的人 ) Score Zadd, zset(feeds 排名 ) Hash Hincrby, hgetall( 用户资料，关注数，粉丝数，微博数 ) Delete history data expire
- 21. Sina weibo 我关注的人 sadd user:hoterran:follows user:foolishceo sadd user:hoterran:follows user:logzgh sadd user:hoterran:follows user:sqlrush smembers user:hoterran:follows 英杰关注的人 sadd user:sqlrush:follows user:logzgh sadd user:sqlrush:follows user:hoterran
- 22. Sina weibo 我们共通关注的人 sinter user:sqlrush:follows user:hoterran:follows 我的粉丝 sadd user:logzgh:follows user:hoterran sadd user:hoterran.follower user:logzgh sadd user:hoterran.follower user:sqlrush smembers user:hoterran:follower
- 23. Sina weibo 我的好友 sinter user:hoterran:follows user:hoterran:follower 我们都关注的人 sinter user:sqlrush:follows user:hoterran:follows
- 24. Sina weibo 可能感兴趣的的人 sadd user:sqlrush:follows user:free_yz sdiff user:sqlrush:follows user:hoterran:follows 我关注的人数，我的粉丝数，我的 Hset hoterran:profile follows 1 Hset hoterran:profile follower 2 Hset hoterran:profile blog 3 Hgetall hoterran:profile Hincrby hoterran:profile blog 1
- 25. Suitable scene 轻量级的高性能消息队列服务 生产者消费者 Producer lpush Consumer blpop Redis 的主要缺点是数据库容量受到物理内存的限制，不能用作海量数据的高性能读写，并且它没有原生的可扩展机制，不具有 scale （可扩展）能力，要依赖客户端来实现分布式读写，因此 Redis 适合的场景主要局限在较小数据量的高性能操作和运算上。
- 26. summary Defects Failover Replication Dba level Memory limit size Advantage High qps Data structure
- 27. qa Replace aliWangwang k-v(blob) Can satisfy Online Log requirement?