A First Look at MySQL Performance Tuning

This article related to: the use and performance tuning of MySQL comes with several performance testing tool mysqlslap method

Performance testing tools -mysqlslap

mysqlslap is MySQL comes with a very good performance testing tool. It can be used
to simulate multiple concurrent client issues a query to the server, update request, and then outputs the simple report

Optional parameters

-engines: representative of the engine to be tested, there can be multiple, separated by a delimiter
-iterations: To run these tests on behalf of how many times
-auto-generate-sql: represented by a system of self-generated SQL scripts to test
-auto-generate-sql-load-type: representatives to be tested is a read or write, or a combination of both (read, write, update, mixed)
-number-of-queries: represents the total number of queries to be run. The number of queries each customer can run with the total number of queries / concurrent calculated
Related information represents -debug-info to extra output from the CPU and memory
The number of type int field to create a test table: -number-int-cols
-auto-generate-sql-add-autoincrement: Representative columns are automatically added to the table generated auto_increment
The number of type char field -number-char-cols to create a test table
-create-schema database testing
-query use custom scripts to perform tests, such as a stored procedure can be called from sql statement or defined to perform the test
-only-print if you want only to see what SQL statements, you can use this option

The following is a simple example of use:

[root@syj ~]# mysqlslap -uroot -proot
 --concurrency=500 --iterations=1 
 --create-schema='test' 
 --query='select * from u_trade_pay_order limit1000;' 
 --number-of-queries=10

Factors affecting the performance of MySQL

The number of connections is not enough

When the number of connections using more than 85% there is a connection number has almost not enough

Check connections related method

1 2	show global status like 'Max_used_connections'; show variables like 'max_connections%';

Reason enough connections

Number of connections set too low or too high
IO slow query causes blockage, resulting in a long time not to release the connection
SQL executed, the connection is not released

Modify the number of connections MySQL

1	set global max_connections = 1024;

Slow query

In principle should be single than 100ms execution time of SQL optimization need to consider things

Slow SQL View

1	show global status like 'Slow_queries’;

The reason appears slow queries

No index or index unreasonable
Data Query is too large
Latched

Cache hit rate

If the query cache hit ratio <95% indicates that the cache size may be set too low or there is a lot of debris cache

Cache hit rate calculation method

1	(Qcache_hits – Qcache_inserts) / Qcache_hits * 100%

Enabling query caching

1	set session query_cache_type=on;

Improve the cache utilization

Using the same SQL, such as use in MyBatist # instead of $
The client and server use the same character set
Increasing the size of the buffer
insert and update causes cache invalidation

All blog article first appeared in public No. "Java learning record" reproduced leave
scan code number to receive public attention 2000GJava Learning Resources