最近在看MySQL5.7 Manual,有关Semisynchronous Replication这一块的内容,我们知道,MySQL默认的Replication是异步的,何为异步?何为半同步?废话不多说,直接看官方解释吧:
1.背景知识
Asynchronous replication
the master writes events to its binary log and slaves request them when they are ready. There is no guarantee that any event will ever reach any slave.
--主库只管把events写入binlog中,不管从库有没有收到。
Fully synchronous replication
when a master commits a transaction, all slaves also will have committed the transaction before the master returns to the session that performed the transaction. The drawback of this is that there might be a lot of delay to complete a transaction.
--主库提交一个事物,需要等待所有从库先提交才能返回结果,执行这个事物。这样会造成一个事物延时。
Semisynchronous replication
falls between asynchronous and fully synchronous replication. The master waits only until at least one slave has received and logged the events. It does not wait for all slaves to acknowledge receipt, and it requires only receipt, not that the events have been fully executed and committed on the slave side.
--介于异步复制和全复制之间,主库仅仅只要等待至少一个从库收到和记录events。它不需要等待所有的从库告诉它收到events,也不需要从库执行和提交事物,从库只是收到events就会告诉主库,这样主库就可以提前提交事物了。
此外,半同步也分两种,有参数rpl_semi_sync_master_wait_point控制,这里我就不多做解释了,我们使用默认设置after_sync,这种数据零丢失
AFTER_SYNC (the default): The master writes each transaction to its binary log and the slave, and syncs the binary log to disk. The master waits for slave acknowledgment of transaction receipt after the sync. Upon receiving acknowledgment, the master commits the transaction to the storage engine and returns a result to the client, which then can proceed.
AFTER_COMMIT: The master writes each transaction to its binary log and the slave, syncs the binary log, and commits the transaction to the storage engine. The master waits for slave acknowledgment of transaction receipt after the commit. Upon receiving acknowledgment, the master returns a result to the client, which then can proceed.
2.测试环境
Role | Hostname | IP | CPU | Memory | MySQL Version |
TPCC | sht-sgmhadoopcm-01 | 172.16.101.54 | 2Core | 8G | NO |
master | sht-sgmhadoopdn-01 | 172.16.101.58 | 2Core | 6G | 5.7.21 |
slave1 | sht-sgmhadoopdn-02 | 172.16.101.59 | 2Core | 6G | 5.7.21 |
slave2 | sht-sgmhadoopdn-03 | 172.16.101.60 | 2Core | 6G | 5.7.21 |
一个master,slave1和slave2都是master的直接从库。
分两种情况测试:
(1)当slave1和slave2都是异步复制的时候
(2)当slave1是半同步复制,slave2是异步复制的时候
3.压力测试
使用TPCC压力测试软件,比较TPS判断异步和半同步复制的性能差异到底有多大。
具体如何测试,可以参考之前的博客:MySQL压测--TPCC安装,测试
通过TPCC测试获得的数据库TpmC数据
Thread Count | Asynchronous | Semisynchronous |
4 | 1145.900 | 1081.133 |
8 | 1966.633 | 1687.933 |
16 | 3202.333 | 2538.233 |
32 | 4363.533 | 3302.167 |
64 | 5112.900 | 3638.667 |
128 | 5257.267 | 3738.100 |
256 | 5199.733 | 3823.833 |
512 | 4753.700 | 3538.633 |
1024 | 3268.633 | 2769.667 |
1536 | 2294.000 | 2215.733 |
绘制折线图像
4.总结
从本测试结果可以看出:
无论是半同步复制还是异步复制,数据库的TpmC刚开始随着并发线程的增加而增加,到达200个左右,TpmC开始逐渐下降;
当线程数在16-1024左右之间,异步复制的TpmC明显比半同步的大很多,相差大概1000-2000左右,所以半同步相比异步性能是要下降很多;
如果你的数据库并发并不是很多,又对数据安全要求比较高,可以优先考虑使用半同步复制;
参考链接:
https://dev.mysql.com/doc/refman/5.7/en/replication-semisync.html
转载于:https://blog.51cto.com/darrenmemos/2132288