Xtradb+haproxy high-Availability DB cluster (iii) SYSBENCH performance test Chapter

XTRADB sysbench Performance Test

Install sysbench, the latest version is 0.5

Yum Install Sysbench-y

Initializing the test table

Sysbench--test=/usr/share/doc/sysbench/tests/db/oltp.lua--oltp-table-size=1000000--mysql-table-engine=innodb-- Mysql-user=sysbench--mysql-password=sysbench--mysql-port=3306--mysql-host=10.0.37.123--mysql-db=test Prepare

Parameter description:

--oltp-table-size     Specifies the size of the table, which is the number of rows in the table

--mysql-table-engine myisam innodb heap ndbcluster bdb maria falcon pbxt

--mysql-db : Specifies the database in which to create the test table, which defaults to the sbtest Library, which needs to be created in advance.

--test : Specify Lua scripts, parameter options most of the same as the old version of --TEST=OLTP Help

--db-driver : Specify driver, default to Mysql

--myisam-max-rows : Specify Myisam of the table Max_Rows Options

--oltp-secondary       key XID (ID) replace PRIMARY KEY (ID) innodb inside the engine create a unique for each table 6

--oltp-auto-inc : set ID column auto-incrementalwith a value of on or off , the default is on

SelectTest

Sysbench--test=/usr/share/doc/sysbench/tests/db/select.lua--oltp-table-size=1000000--mysql-table-engine=innodb --mysql-user=sbtest--mysql-password=sbpass--mysql-port=3306--mysql-host=192.168.70.72--mysql-db=sbtest-- Max-requests=0--max-time=120--oltp-tables-count=1--report-interval=10--num-threads=8 Run

Parameter description:

--max-time to specify the length of the test

--oltp-tables-count Specify the number of test tables

--max-requests Specify the maximum number of requests, default 10000 , 0 is not limited

--num-threads Specify the number of threads

Prepare is a preparation process, such as measuring OLTP need to Load Data to Table

Run It's a real test process.

Cleanup is the purge process

UpdateTest

Sysbench--test=/usr/share/doc/sysbench/tests/db/update_index.lua--oltp-table-size=1000000--mysql-table-engine= InnoDB--mysql-user=sbtest--mysql-password=sbpass--mysql-port=3306--mysql-host=192.168.70.72--mysql-db=sbtest-- Max-requests=0--max-time=120--oltp-tables-count=1--report-interval=10--num_threads=8 Run

InsertTest

Sysbench--test=/usr/share/doc/sysbench/tests/db/insert.lua--oltp-table-size=1000000--mysql-table-engine=innodb --mysql-user=sbtest--mysql-password=sbpass--mysql-port=3306--mysql-host=192.168.70.72--mysql-db=sbtest-- Max-requests=0--max-time=120--oltp-tables-count=1--report-interval=10--num_threads=8 Run

Test Results

8 thread	read 120s	insert 120s	120s
A single	578859	76204	70780
Two units	531659	90487	91068
Three	459249	88791	81759
A single but Haproxy	718150	76268	69842

In writing, over the Haproxy, one with more than a small gap, more than the performance of the writing has little impact.

In reading, supposedly, multiple units should be doubled performance, but from the test results, there is no change, do not know why, some people know why also please inform.

This article is from the "Yang Cloud" blog, please be sure to keep this source http://yangrong.blog.51cto.com/6945369/1684140

Xtradb+haproxy high-Availability DB cluster (iii) SYSBENCH performance test Chapter