MySQL Cluster2个数据节点压力测试–mysqlslap工具压400W写
锅巴哥的个人建议：cluster叫电信运营商版本，所以基本上在很大的用户并发量的情况下才会用到，对连接数的线性增长要求高的场景，千兆就不用想了， 没万兆就不用玩了。
很不幸，我的就是千兆网络，我的数据节点是16 core，32G内存，sql节点是16core，16G内存，做了raid10
PS: 水平所限，环境所限，测试结果仅供参考，欢迎大家一起交流提出宝贵的意见，让我更进一步。
1【】一：测试
10.80.40.107 3307
mysql -utim -pxxxxxxxx -P3307 -h10.80.40.107 –default-character-set=utf8
— select 查询测试 100W 数据压力测试
1.1 ndbcluster
(time mysqlslap –engine=ndbcluster -h10.80.40.107 -P3307 -utim -pxxxxxxxx –auto-generate-sql-guid-primary –concurrency=500,1000,1500 –number-of-queries=1000000 –iterations=2 –number-char-cols=10 –number-int-cols=10 –auto-generate-sql –create-schema=rdb10 –auto-generate-sql-load-type=read ) 1> r101.log 2> r102.log
— 20.40.30.31
(time mysqlslap –engine=ndbcluster –auto-generate-sql-guid-primary –concurrency=800,1200,1500 –number-of-queries=1000000 –iterations=2 –number-char-cols=10 –number-int-cols=10 –auto-generate-sql –create-schema=rdb10 –auto-generate-sql-load-type=read ) 1> /root/r101.log 2> /root/r102.log
1.2 innodb
— 20.40.30.37
(time mysqlslap –engine=innodb -h20.40.30.37 -P3307 -utim -pxxxxxxxx –auto-generate-sql-guid-primary –concurrency=800,1200,1500 –number-of-queries=1000000 –iterations=2 –number-char-cols=10 –number-int-cols=10 –auto-generate-sql –create-schema=idb10 –auto-generate-sql-load-type=read ) 1> /root/rin101.log 2> /root/rin102.log
2 mixed 读写混合压力测试
ndbcluster 10W
2.1【===================10W 写 =================】
— ndbcluster 在集群里面测试
2.1.1 ndbcluster 10W测试，并发分别为–concurrency=500，1000，1500
(time mysqlslap –engine=ndbcluster -h10.80.40.107 -P3307 -utim -pxxxxxxxx –auto-generate-sql-write-number=1000 –auto-generate-sql-guid-primary –concurrency=500 –number-of-queries=100000 –iterations=2 –number-char-cols=10 –number-int-cols=10 –auto-generate-sql –create-schema=ndb10 –-commint=100 –auto-generate-sql-load-type=write)
(time mysqlslap –engine=ndbcluster -h10.80.40.107 -P3307 -utim -pxxxxxxxx –auto-generate-sql-write-number=1000 –auto-generate-sql-guid-primary –concurrency=1000 –number-of-queries=100000 –iterations=2 –number-char-cols=10 –number-int-cols=10 –auto-generate-sql –create-schema=ndb10 –-commint=100 –auto-generate-sql-load-type=write)
(time mysqlslap –engine=ndbcluster -h10.80.40.107 -P3307 -utim -pxxxxxxxx –auto-generate-sql-write-number=1000 –auto-generate-sql-guid-primary –concurrency=1500 –number-of-queries=100000 –iterations=2 –number-char-cols=10 –number-int-cols=10 –auto-generate-sql –create-schema=ndb10 –-commint=100 –auto-generate-sql-load-type=write)
Benchmark
Running for engine ndbcluster
Average number of seconds to run all queries: 4.885 seconds
Minimum number of seconds to run all queries: 4.729 seconds
Maximum number of seconds to run all queries: 5.041 seconds
Number of clients running queries: 500
Average number of queries per client: 200
Benchmark
Running for engine ndbcluster
Average number of seconds to run all queries: 5.948 seconds
Minimum number of seconds to run all queries: 5.399 seconds
Maximum number of second
