Redis 主从同步

一个架构师

发布于 2022-06-20 19:54:31

4680

发布于 2022-06-20 19:54:31

文章被收录于专栏：从码农的全世界路过从码农的全世界路过

Redis的主从配置可以说是哨兵或者集群模式的最简单元.今天我们就来看下Redis是如何做主从同步,以及要注意什么.

一. 主从配置

配置方式有两种:

(1) 配置Redis.conf

#slaveof <masterip> <masterport>
slaveof 172.17.0.2 6379

(2) 使用slaveof命令配置

SLAVEOF 172.17.0.2 6379

二. 全量同步

主从节点数据同步流程如下:

可见主从同步会有两个主要过程

1. 同步全量RDB信息

2. 同步增量repl_back_buffer信息,buffer的默认大小为1M,可以通过redis.conf中repl-backlog-size配置修改

通过主从节点的日志了解数据同步过程

主节点日志

1:M 19 May 02:30:25.225 * Slave 172.17.0.3:6380 asks for synchronization
1:M 19 May 02:30:25.226 * Partial resynchronization not accepted: Replication ID mismatch (Slave asked for '225ce4bd8f8cebae2035325e51123099852bb7dd', my replication IDs are 'e0aafc1924a836d84d1a1643b07bef6d085e6d63' and '0000000000000000000000000000000000000000')
1:M 19 May 02:30:25.226 * Starting BGSAVE for SYNC with target: disk
1:M 19 May 02:30:25.228 * Background saving started by pid 36
36:C 19 May 02:30:25.233 * DB saved on disk
36:C 19 May 02:30:25.235 * RDB: 0 MB of memory used by copy-on-write
1:M 19 May 02:30:25.323 * Background saving terminated with success
1:M 19 May 02:30:25.324 * Synchronization with slave 172.17.0.3:6380 succeeded

从节点日志

1:M 19 May 02:29:03.020 * Ready to accept connections
1:S 19 May 02:30:24.898 * Before turning into a slave, using my master parameters to synthesize a cached master: I may be able to synchronize with the new master with just a partial transfer.
1:S 19 May 02:30:24.898 * SLAVE OF 172.17.0.2:6379 enabled (user request from 'id=3 addr=127.0.0.1:36280 fd=7 name= age=46 idle=0 flags=N db=0 sub=0 psub=0 multi=-1 qbuf=0 qbuf-free=32768 obl=0 oll=0 omem=0 events=r cmd=slaveof')
1:S 19 May 02:30:25.220 * Connecting to MASTER 172.17.0.2:6379
1:S 19 May 02:30:25.221 * MASTER <-> SLAVE sync started
1:S 19 May 02:30:25.221 * Non blocking connect for SYNC fired the event.
1:S 19 May 02:30:25.223 * Master replied to PING, replication can continue...
1:S 19 May 02:30:25.224 * Trying a partial resynchronization (request 225ce4bd8f8cebae2035325e51123099852bb7dd:1).
1:S 19 May 02:30:25.229 * Full resync from master: e8063053a7483b6356511217e29c5db13d483ef4:0
1:S 19 May 02:30:25.229 * Discarding previously cached master state.
1:S 19 May 02:30:25.324 * MASTER <-> SLAVE sync: receiving 187 bytes from master
1:S 19 May 02:30:25.324 * MASTER <-> SLAVE sync: Flushing old data
1:S 19 May 02:30:25.324 * MASTER <-> SLAVE sync: Loading DB in memory
1:S 19 May 02:30:25.324 * MASTER <-> SLAVE sync: Finished with success

三. 增量同步

1. 从节点发送psync master_runid offset请求

2. 主节点判断runid是否和本机的id一致, offset是否超过了buffer缓存区大小;

3. 如果runid不一致,或者offset超过缓存区大小,返回FULLRESYNC runid offset,进行全量同步.

4. 如果runid一致,offset不超过缓存区大小,则返回CONTINUE信息,并向从节点发送增量数据.

四. 主从节点同步状态

使用info 命令,可以观察到主从节点的同步状态

主节点信息

127.0.0.1:6379> info replication
# Replication
role:master
connected_slaves:1
slave0:ip=172.17.0.3,port=6380,state=online,offset=15114,lag=0
master_replid:e8063053a7483b6356511217e29c5db13d483ef4
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:15114
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:15114

从节点信息

127.0.0.1:6380> info replication
# Replication
role:slave
master_host:172.17.0.2
master_port:6379
master_link_status:up
master_last_io_seconds_ago:2
master_sync_in_progress:0
slave_repl_offset:15170
slave_priority:100
slave_read_only:1
connected_slaves:0
master_replid:e8063053a7483b6356511217e29c5db13d483ef4
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:15170
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:15170

五. 注意事项

1. 主从同步会有延迟

2. repl_back_buffer设置不能太小,会影响正常全量同步以及网络闪断时的增量同步.

3. 主从配置要一致

maxmemory不一致,可能会造成丢失数据

数据结构优化参数不一致,会造成主从内存不一致

本文参与腾讯云自媒体分享计划，分享自微信公众号。

原始发表：2020-05-20，如有侵权请联系 cloudcommunity@tencent.com 删除

node.js