mysql經常使用配置注意項與sql優化

時間 2019-11-08

標籤 mysql 經常使用配置注意 sql 優化欄目 MySQL 简体版

原文原文鏈接

創建數據庫：html

- 創建數據庫時編碼字符集採用utf8
- 排序規則：
  - 後綴"_cs"或者"_ci"意思是區分大小寫和不區分大小寫（Case Sensitive & Case Insensitve）
  - 後綴"_bin" 規定每一個字符串用二進制編碼存儲，區分大小寫，能夠直接存儲二進制的內容
  - utf-8有默認的排序規則：命令：SHOW CHARSET LIKE 'utf8%';
  - 注意點：何時須要區分大小寫須要在設計和使用時注意mysql
  - 若是排序規則使用的是不區分大小寫，但部分表字段須要區分大小寫，則能夠對該字段進行修改：
```
ALTER TABLE yourTableName MODIFY COLUMN columnName VARCHAR(255) BINARY CHARACTER SET utf8 COLLATE utf8_bin DEFAULT NULL;
```

建表：linux

　　常見表字段選擇請參考其它sql

　　表創建完成後，能夠插入必定量的，和業務真實基本一致的數據後，經過執行數據庫

SELECT * FROM yourTableName PROCEDURE ANALYSE();

　　根據建議修改表字段定義windows

　　注意：此處建議知識針對表中數據，請合理取捨less

測試環境設置：socket

　　模擬真實場景數據，放大 xxx 倍，做爲上線一段時間後的業務數據預期值tcp

統計數據量以配置緩衝區大小

#統計指定庫表的行數和數據量大小
select TABLE_SCHEMA as DB_NAME,TABLE_NAME,TABLE_ROWS,DATA_LENGTH
from information_schema.TABLES
where table_schema='bdjc_ls'          #指定具體的庫名
-- AND table_name='t_syxm';            #指定具體表名
ORDER BY DATA_LENGTH DESC

#統計整個mysql數據量大小  
select
concat(round(sum(DATA_LENGTH/1024/1024),2),'MB') as ALL_DB_DATA_SIZE,
concat(round(sum(INDEX_LENGTH/1024/1024),2),'MB') as ALL_DB_INDEX_SIZE,
concat(round(sum((DATA_LENGTH + INDEX_LENGTH)/1024/1024),2),'MB') as ALL_DB_TOTAL_SIZE
from information_schema.TABLES;

#查詢所有庫或者指定庫數據量大小
select
table_schema as DB_NAME,
concat(round(sum(DATA_LENGTH/1024/1024),2),'MB') as DB_DATA_SIZE,
concat(round(sum(INDEX_LENGTH/1024/1024),2),'MB') as DB_INDEX_SIZE,
concat(round(sum((DATA_LENGTH + INDEX_LENGTH)/1024/1024),2),'MB') as DB_TOTAL_SIZE
from information_schema.TABLES
where table_schema='bdjc_ls'          #指定具體的庫名
AND table_name='t_syxm';            #指定具體表名
group by table_schema
order by sum(DATA_LENGTH + INDEX_LENGTH) desc ;

#主表數據和索引數據的最大內存緩衝區，分配過大，會使Swap佔用過多，導致Mysql的查詢特慢
SELECT @@innodb_buffer_pool_size;  

#windows下能夠修改my.ini文件進行設置，默認爲8M  my.ini文件能夠經過 select @@datadir 找到數據根路徑， 而後在data上層目錄中會發現my.ini
在[mysqld] 下進行修改，包括常見的端口，慢查詢等各種參數，注意不要超過最大值

好比個人電腦上文件配置爲：

# InnoDB, unlike MyISAM, uses a buffer pool to cache both indexes and
# row data. The bigger you set this the less disk I/O is needed to
# access data in tables. On a dedicated database server you may set this
# parameter up to 80% of the machine physical memory size. Do not set it
# too large, though, because competition of the physical memory may
# cause paging in the operating system. Note that on 32bit systems you
# might be limited to 2-3.5G of user level memory per process, so do not
# set it too high.
innodb_buffer_pool_size=3Gide

#linux下經過mysql --help|grep 'my.cnf' 查找配置文件，優先使用考前的，通常爲/etc/my.cnf

設置的innodb_buffer_pool_size 須要爲 innodb_buffer_pool_chunk_size*innodb_buffer_pool_instances 的倍數，
若是不是倍數，設置的innodb_buffer_pool_size會自動調整爲倍數
innodb_buffer_pool_chunk_size是一個只讀值， 當innodb_buffer_pool_size大於1G時，就應該將innodb_buffer_pool_instances值調大

#查看設置是否合理
#Performance = innodb_buffer_pool_reads / innodb_buffer_pool_read_requests * 100
show status like 'innodb_buffer_pool_read%';

若是Performance百分比很小，則表示夠用

參考 https://www.cnblogs.com/wanbin/p/9530833.html

執行
show engine innodb status
查看 Free buffers 大小，若是長時間很大，則能夠調小innodb_buffer_pool_size，不然能夠適當調大innodb_buffer_pool_size


命中率 = innodb_buffer_pool_read_requests / (innodb_buffer_pool_read_requests + innodb_buffer_pool_reads ) * 100

若是命中率低於99%，則能夠考慮增長innodb_buffer_pool_size

開啓慢查詢

=============開啓慢查詢========================
vi /etc/my.cnf
[mysqld]
…
slow_query_log = 1                                         #無效(0或者OFF)、有效(1或者ON)
slow_query_log_file = /data/log/mysql/slow_query.log      #指定日誌文件
long_query_time = 0.5                                    #超過指定時間的SQL會記錄到日誌文件(默認時間爲10秒，默認單位爲秒)

#或者全局設置
-----------------
set global slow_query_log = 1;
set global slow_query_log_file = '/data/log/mysql/slow_query.log';
set global long_query_time = 0.5;

#記錄sql執行日誌
SET GLOBAL general_log=1
SET GLOBAL log_output='FILE';

mysqldumpslow 慢日誌分析工具 
命令：

-s 按照那種方式排序
    c：訪問計數
    l：鎖定時間
    r:返回記錄
    al：平均鎖定時間
    ar：平均訪問記錄數
    at：平均查詢時間
-t 是top n的意思，返回多少條數據。
-g 能夠跟上正則匹配模式，大小寫不敏感。


#獲得返回記錄最多的20個sql
mysqldumpslow -s r -t 20 /data/log/mysql/slow_query.log

#獲得平均訪問次數最多的20條sql
mysqldumpslow -s ar -t 20 /data/log/mysql/slow_query.log

#獲得平均訪問次數最多,而且裏面含有ttt字符的20條sql
mysqldumpslow -s ar -t 20 -g "ttt" /data/log/mysql/slow_query.log

若是出現以下錯誤，Died at /usr/bin/mysqldumpslow line 161, <> chunk 405659.說明要分析的sql日誌太大了，請拆分後再分析
拆分的命令爲：
tail -10000 /data/log/mysql/slow_query.log>/data/log/mysql/slow_query_1_10000.log

能夠經過explain 解析執行對應的慢查詢，經過調整表結構、調整索引、查詢語句等常規方式優化查詢

mysql性能壓力測試 mysqlslap

mysqlslap -h192.168.0.200 -P3306 -uroot -p123456 --number-char-cols=5 --number-int-cols=3 --concurrency=1 --iterations=1 --auto-generate-sql --auto-generate-sql-load-type=mixed --auto-generate-sql-add-autoincrement --engine=innodb,myisam --number-of-queries=1 --auto-generate-sql-write-number=1 --only-print 

mysqlslap -h192.168.0.200 -P3306 -uroot -p123456 --concurrency=100,500,1000 --iterations=1 --auto-generate-sql --auto-generate-sql-load-type=mixed --auto-generate-sql-add-autoincrement --engine=innodb --number-of-queries=5000

mysqlslap -h192.168.0.200 -P3306 -uroot -p123456 --concurrency=2 --iterations=2 --create-schema=mysql --query="select * from user" --engine=innodb --number-of-queries=20

mysqlslap -h192.168.0.200 -P3306 -uroot -p123456 --concurrency=100 --iterations=1 --create-schema=mysql --query=/root/mysql/query.sql --engine=innodb --number-of-queries=5000

--print-defaults Print the program argument list and exit.
--no-defaults Don't read default options from any option file,
except for login file.
--defaults-file=# Only read default options from the given file #.
--defaults-extra-file=# Read this file after the global files are read.
--defaults-group-suffix=#
Also read groups with concat(group, suffix)
--login-path=# Read this path from the login file.
-?, --help Display this help and exit.
-a, --auto-generate-sql
Generate SQL where not supplied by file or command line.
--auto-generate-sql-add-autoincrement
Add an AUTO_INCREMENT column to auto-generated tables.
--auto-generate-sql-execute-number=#
Set this number to generate a set number of queries to run.
--auto-generate-sql-guid-primary
Add GUID based primary keys to auto-generated tables.
--auto-generate-sql-load-type=name
Specify test load type: mixed, update, write, key, or read; default is mixed.
--auto-generate-sql-secondary-indexes=#
Number of secondary indexes to add to auto-generated
tables.
--auto-generate-sql-unique-query-number=#
Number of unique queries to generate for automatic tests.
--auto-generate-sql-unique-write-number=#
Number of unique queries to generate for auto-generate-sql-write-number.
--auto-generate-sql-write-number=#
Number of row inserts to perform for each thread (default is 100).
--commit=# Commit records every X number of statements.
-C, --compress Use compression in server/client protocol.
-c, --concurrency=name
Number of clients to simulate for query to run.
--create=name File or string to use create tables.
--create-schema=name
Schema to run tests in.
--csv[=name] Generate CSV output to named file or to stdout if no file is named.
-#, --debug[=#] This is a non-debug version. Catch this and exit.
--debug-check This is a non-debug version. Catch this and exit.
-T, --debug-info This is a non-debug version. Catch this and exit.
--default-auth=name Default authentication client-side plugin to use.
-F, --delimiter=name
Delimiter to use in SQL statements supplied in file or command line.
--detach=# Detach (close and reopen) connections after X number of requests.
--enable-cleartext-plugin
Enable/disable the clear text authentication plugin.
-e, --engine=name Storage engine to use for creating the table.
-h, --host=name Connect to host.
-i, --iterations=# Number of times to run the tests.
--no-drop Do not drop the schema after the test.
-x, --number-char-cols=name
Number of VARCHAR columns to create in table if specifying --auto-generate-sql.
-y, --number-int-cols=name
Number of INT columns to create in table if specifying --auto-generate-sql.
--number-of-queries=#
Limit each client to this number of queries (this is not exact).
--only-print Do not connect to the databases, but instead print out what would have been done.
-p, --password[=name]
Password to use when connecting to server. If password is not given it's asked from the tty.
-W, --pipe Use named pipes to connect to server.
--plugin-dir=name Directory for client-side plugins.
-P, --port=# Port number to use for connection.
--post-query=name Query to run or file containing query to execute after
tests have completed.
--post-system=name system() string to execute after tests have completed.
--pre-query=name Query to run or file containing query to execute before
running tests.
--pre-system=name system() string to execute before running tests.
--protocol=name The protocol to use for connection (tcp, socket, pipe,
memory).
-q, --query=name Query to run or file containing query to run.
--secure-auth Refuse client connecting to server if it uses old
(pre-4.1.1) protocol. Deprecated. Always TRUE
--shared-memory-base-name=name
Base name of shared memory.
-s, --silent Run program in silent mode - no output.
-S, --socket=name The socket file to use for connection.
--sql-mode=name Specify sql-mode to run mysqlslap tool.
--ssl-mode=name SSL connection mode.
--ssl Deprecated. Use --ssl-mode instead.
(Defaults to on; use --skip-ssl to disable.)
--ssl-verify-server-cert
Deprecated. Use --ssl-mode=VERIFY_IDENTITY instead.
--ssl-ca=name CA file in PEM format.
--ssl-capath=name CA directory.
--ssl-cert=name X509 cert in PEM format.
--ssl-cipher=name SSL cipher to use.
--ssl-key=name X509 key in PEM format.
--ssl-crl=name Certificate revocation list.
--ssl-crlpath=name Certificate revocation list path.
--tls-version=name TLS version to use, permitted values are: TLSv1, TLSv1.1
-u, --user=name User for login if not current user.
-v, --verbose More verbose output; you can use this multiple times to
get even more verbose output.
-V, --version Output version information and exit.