前言

我負(fù)責(zé)的有幾個(gè)系統(tǒng)隨著業(yè)務(wù)量的增長(zhǎng)，存儲(chǔ)在MySQL中的數(shù)據(jù)日益劇增，我當(dāng)時(shí)就想現(xiàn)在的業(yè)務(wù)方不講武德，搞偷襲，趁我沒(méi)反應(yīng)過(guò)來(lái)把很多表，很快，很快啊都打到了億級(jí)別，我大意了，沒(méi)有閃，這就導(dǎo)致跟其Join的表的SQL變得很慢，對(duì)的應(yīng)用接口的response time也變長(zhǎng)了，影響了用戶體驗(yàn)。

事后我找到業(yè)務(wù)方，我批評(píng)了他們跟他們說(shuō)要講武德，連忙跟我道歉，這個(gè)事情才就此作罷，走的時(shí)候我對(duì)他們說(shuō)下次不要這樣了，耗子尾汁，好好反思。

罵歸罵，事情還是得解決，時(shí)候我分析原因發(fā)現(xiàn)，發(fā)現(xiàn)有些表的數(shù)據(jù)量增長(zhǎng)很快，對(duì)應(yīng)SQL掃描了很多無(wú)效數(shù)據(jù)，導(dǎo)致SQL慢了下來(lái)，通過(guò)確認(rèn)之后，這些大表都是一些流水、記錄、日志類型數(shù)據(jù)，只需要保留1到3個(gè)月，此時(shí)需要對(duì)表做數(shù)據(jù)清理實(shí)現(xiàn)瘦身，一般都會(huì)想到用insert + delete的方式去清理。

這篇文章我會(huì)從InnoDB存儲(chǔ)空間分布，delete對(duì)性能的影響，以及優(yōu)化建議方面解釋為什么不建議delete刪除數(shù)據(jù)。

InnoDB存儲(chǔ)架構(gòu)

從這張圖可以看到，InnoDB存儲(chǔ)結(jié)構(gòu)主要包括兩部分：邏輯存儲(chǔ)結(jié)構(gòu)和物理存儲(chǔ)結(jié)構(gòu)。

邏輯上是由表空間tablespace —> 段segment或者inode —> 區(qū)Extent ——>數(shù)據(jù)頁(yè)P(yáng)age構(gòu)成，Innodb邏輯管理單位是segment，空間分配的最小單位是extent，每個(gè)segment都會(huì)從表空間FREE_PAGE中分配32個(gè)page，當(dāng)這32個(gè)page不夠用時(shí)，會(huì)按照以下原則進(jìn)行擴(kuò)展：如果當(dāng)前小于1個(gè)extent，則擴(kuò)展到1個(gè)extent；當(dāng)表空間小于32MB時(shí)，每次擴(kuò)展一個(gè)extent；表空間大于32MB，每次擴(kuò)展4個(gè)extent。

物理上主要由系統(tǒng)用戶數(shù)據(jù)文件，日志文件組成，數(shù)據(jù)文件主要存儲(chǔ)MySQL字典數(shù)據(jù)和用戶數(shù)據(jù)，日志文件記錄的是data page的變更記錄，用于MySQL Crash時(shí)的恢復(fù)。

Innodb表空間

InnoDB存儲(chǔ)包括三類表空間：系統(tǒng)表空間，用戶表空間，Undo表空間。

**系統(tǒng)表空間：**主要存儲(chǔ)MySQL內(nèi)部的數(shù)據(jù)字典數(shù)據(jù)，如information_schema下的數(shù)據(jù)。

**用戶表空間：**當(dāng)開(kāi)啟innodb_file_per_table=1時(shí)，數(shù)據(jù)表從系統(tǒng)表空間獨(dú)立出來(lái)存儲(chǔ)在以table_name.ibd命令的數(shù)據(jù)文件中，結(jié)構(gòu)信息存儲(chǔ)在table_name.frm文件中。

**Undo表空間：**存儲(chǔ)Undo信息，如快照一致讀和flashback都是利用undo信息。

從MySQL 8.0開(kāi)始允許用戶自定義表空間，具體語(yǔ)法如下：

CREATE TABLESPACE tablespace_name ADD DATAFILE ’file_name’#數(shù)據(jù)文件名 USE LOGFILE GROUP logfile_group #自定義日志文件組，一般每組2個(gè)logfile。 [EXTENT_SIZE [=] extent_size] #區(qū)大小 [INITIAL_SIZE [=] initial_size] #初始化大小 [AUTOEXTEND_SIZE [=] autoextend_size] #自動(dòng)擴(kuò)寬尺寸 [MAX_SIZE [=] max_size]#單個(gè)文件最大size，最大是32G。 [NODEGROUP [=] nodegroup_id] #節(jié)點(diǎn)組 [WAIT] [COMMENT [=] comment_text] ENGINE [=] engine_name

這樣的好處是可以做到數(shù)據(jù)的冷熱分離，分別用HDD和SSD來(lái)存儲(chǔ)，既能實(shí)現(xiàn)數(shù)據(jù)的高效訪問(wèn)，又能節(jié)約成本，比如可以添加兩塊500G硬盤，經(jīng)過(guò)創(chuàng)建卷組vg，劃分邏輯卷lv，創(chuàng)建數(shù)據(jù)目錄并mount相應(yīng)的lv，假設(shè)劃分的兩個(gè)目錄分別是/hot_data 和 /cold_data。

這樣就可以將核心的業(yè)務(wù)表如用戶表，訂單表存儲(chǔ)在高性能SSD盤上，一些日志，流水表存儲(chǔ)在普通的HDD上，主要的操作步驟如下：

#創(chuàng)建熱數(shù)據(jù)表空間create tablespace tbs_data_hot add datafile ’/hot_data/tbs_data_hot01.dbf’ max_size 20G;#創(chuàng)建核心業(yè)務(wù)表存儲(chǔ)在熱數(shù)據(jù)表空間create table booking(id bigint not null primary key auto_increment, …… ) tablespace tbs_data_hot;#創(chuàng)建冷數(shù)據(jù)表空間create tablespace tbs_data_cold add datafile ’/hot_data/tbs_data_cold01.dbf’ max_size 20G;#創(chuàng)建日志，流水，備份類的表存儲(chǔ)在冷數(shù)據(jù)表空間create table payment_log(id bigint not null primary key auto_increment, …… ) tablespace tbs_data_cold;#可以移動(dòng)表到另一個(gè)表空間alter table payment_log tablespace tbs_data_hot;Inndob存儲(chǔ)分布創(chuàng)建空表查看空間變化

mysql> create table user(id bigint not null primary key auto_increment, -> name varchar(20) not null default ’’ comment ’姓名’, -> age tinyint not null default 0 comment ’age’, -> gender char(1) not null default ’M’ comment ’性別’, -> phone varchar(16) not null default ’’ comment ’手機(jī)號(hào)’, -> create_time datetime NOT NULL DEFAULT CURRENT_TIMESTAMP COMMENT ’創(chuàng)建時(shí)間’, -> update_time datetime NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT ’修改時(shí)間’ -> ) engine = InnoDB DEFAULT CHARSET=utf8mb4 COMMENT ’用戶信息表’;Query OK, 0 rows affected (0.26 sec)

# ls -lh user1.ibd -rw-r----- 1 mysql mysql 96K Nov 6 12:48 user.ibd

設(shè)置參數(shù)innodb_file_per_table=1時(shí)，創(chuàng)建表時(shí)會(huì)自動(dòng)創(chuàng)建一個(gè)segment，同時(shí)分配一個(gè)extent，包含32個(gè)data page的來(lái)存儲(chǔ)數(shù)據(jù)，這樣創(chuàng)建的空表默認(rèn)大小就是96KB，extent使用完之后會(huì)申請(qǐng)64個(gè)連接頁(yè)，這樣對(duì)于一些小表，或者undo segment，可以在開(kāi)始時(shí)申請(qǐng)較少的空間，節(jié)省磁盤容量的開(kāi)銷。

# python2 py_innodb_page_info.py -v /data2/mysql/test/user.ibdpage offset 00000000, page type <File Space Header>page offset 00000001, page type <Insert Buffer Bitmap>page offset 00000002, page type <File Segment inode>page offset 00000003, page type <B-tree Node>, page level <0000>page offset 00000000, page type <Freshly Allocated Page>page offset 00000000, page type <Freshly Allocated Page>Total number of page: 6: #總共分配的頁(yè)數(shù)Freshly Allocated Page: 2 #可用的數(shù)據(jù)頁(yè)Insert Buffer Bitmap: 1 #插入緩沖頁(yè)File Space Header: 1 #文件空間頭B-tree Node: 1#數(shù)據(jù)頁(yè)File Segment inode: 1 #文件端inonde，如果是在ibdata1.ibd上會(huì)有多個(gè)inode。插入數(shù)據(jù)后的空間變化

mysql> DELIMITER $$mysql> CREATE PROCEDURE insert_user_data(num INTEGER) -> BEGIN -> DECLARE v_i int unsigned DEFAULT 0; -> set autocommit= 0; -> WHILE v_i < num DO -> insert into user(`name`, age, gender, phone) values (CONCAT(’lyn’,v_i), mod(v_i,120), ’M’, CONCAT(’152’,ROUND(RAND(1)*100000000))); -> SET v_i = v_i+1; -> END WHILE; -> commit; -> END $$Query OK, 0 rows affected (0.01 sec)mysql> DELIMITER ;#插入10w數(shù)據(jù)mysql> call insert_user_data(100000);Query OK, 0 rows affected (6.69 sec)

# ls -lh user.ibd-rw-r----- 1 mysql mysql 14M Nov 6 10:58 /data2/mysql/test/user.ibd# python2 py_innodb_page_info.py -v /data2/mysql/test/user.ibdpage offset 00000000, page type <File Space Header>page offset 00000001, page type <Insert Buffer Bitmap>page offset 00000002, page type <File Segment inode>page offset 00000003, page type <B-tree Node>, page level <0001> #增加了一個(gè)非葉子節(jié)點(diǎn)，樹(shù)的高度從1變?yōu)?.........................................................page offset 00000000, page type <Freshly Allocated Page>Total number of page: 896:Freshly Allocated Page: 493Insert Buffer Bitmap: 1File Space Header: 1B-tree Node: 400File Segment inode: 1delete數(shù)據(jù)后的空間變化

mysql> select min(id),max(id),count(*) from user;+---------+---------+----------+| min(id) | max(id) | count(*) |+---------+---------+----------+| 1 | 100000 | 100000 |+---------+---------+----------+1 row in set (0.05 sec)#刪除50000條數(shù)據(jù)，理論上空間應(yīng)該從14MB變長(zhǎng)7MB左右。mysql> delete from user limit 50000;Query OK, 50000 rows affected (0.25 sec) #數(shù)據(jù)文件大小依然是14MB，沒(méi)有縮小。# ls -lh /data2/mysql/test/user1.ibd -rw-r----- 1 mysql mysql 14M Nov 6 13:22 /data2/mysql/test/user.ibd#數(shù)據(jù)頁(yè)沒(méi)有被回收。# python2 py_innodb_page_info.py -v /data2/mysql/test/user.ibdpage offset 00000000, page type <File Space Header>page offset 00000001, page type <Insert Buffer Bitmap>page offset 00000002, page type <File Segment inode>page offset 00000003, page type <B-tree Node>, page level <0001>........................................................page offset 00000000, page type <Freshly Allocated Page>Total number of page: 896:Freshly Allocated Page: 493Insert Buffer Bitmap: 1File Space Header: 1B-tree Node: 400File Segment inode: 1#在MySQL內(nèi)部是標(biāo)記刪除，

mysql> use information_schema;Database changedmysql> SELECT A.SPACE AS TBL_SPACEID, A.TABLE_ID, A.NAME AS TABLE_NAME, FILE_FORMAT, ROW_FORMAT, SPACE_TYPE, B.INDEX_ID , B.NAME AS INDEX_NAME, PAGE_NO, B.TYPE AS INDEX_TYPE FROM INNODB_SYS_TABLES A LEFT JOIN INNODB_SYS_INDEXES B ON A.TABLE_ID =B.TABLE_ID WHERE A.NAME = ’test/user1’;+-------------+----------+------------+-------------+------------+------------+----------+------------+---------+------------+| TBL_SPACEID | TABLE_ID | TABLE_NAME | FILE_FORMAT | ROW_FORMAT | SPACE_TYPE | INDEX_ID | INDEX_NAME | PAGE_NO | INDEX_TYPE |+-------------+----------+------------+-------------+------------+------------+----------+------------+---------+------------+| 1283 | 1207 | test/user | Barracuda | Dynamic | Single | 2236 | PRIMARY | 3 | 3 |+-------------+----------+------------+-------------+------------+------------+----------+------------+---------+------------+1 row in set (0.01 sec)PAGE_NO = 3 標(biāo)識(shí)B-tree的root page是3號(hào)頁(yè)，INDEX_TYPE = 3是聚集索引。 INDEX_TYPE取值如下：0 = nonunique secondary index; 1 = automatically generated clustered index (GEN_CLUST_INDEX); 2 = unique nonclustered index; 3 = clustered index; 32 = full-text index;#收縮空間再后進(jìn)行觀察

MySQL內(nèi)部不會(huì)真正刪除空間，而且做標(biāo)記刪除，即將delflag:N修改為delflag:Y，commit之后會(huì)會(huì)被purge進(jìn)入刪除鏈表，如果下一次insert更大的記錄，delete之后的空間不會(huì)被重用，如果插入的記錄小于等于delete的記錄空會(huì)被重用，這塊內(nèi)容可以通過(guò)知數(shù)堂的innblock工具進(jìn)行分析。

Innodb中的碎片碎片的產(chǎn)生

我們知道數(shù)據(jù)存儲(chǔ)在文件系統(tǒng)上的，總是不能100%利用分配給它的物理空間，刪除數(shù)據(jù)會(huì)在頁(yè)面上留下一些”空洞”，或者隨機(jī)寫(xiě)入（聚集索引非線性增加）會(huì)導(dǎo)致頁(yè)分裂，頁(yè)分裂導(dǎo)致頁(yè)面的利用空間少于50%，另外對(duì)表進(jìn)行增刪改會(huì)引起對(duì)應(yīng)的二級(jí)索引值的隨機(jī)的增刪改，也會(huì)導(dǎo)致索引結(jié)構(gòu)中的數(shù)據(jù)頁(yè)面上留下一些'空洞'，雖然這些空洞有可能會(huì)被重復(fù)利用，但終究會(huì)導(dǎo)致部分物理空間未被使用，也就是碎片。

同時(shí)，即便是設(shè)置了填充因子為100%，Innodb也會(huì)主動(dòng)留下page頁(yè)面1/16的空間作為預(yù)留使用（An innodb_fill_factor setting of 100 leaves 1/16 of the space in clustered index pages free for future index growth）防止update帶來(lái)的行溢出。

mysql> select table_schema, -> table_name,ENGINE, -> round(DATA_LENGTH/1024/1024+ INDEX_LENGTH/1024/1024) total_mb,TABLE_ROWS, -> round(DATA_LENGTH/1024/1024) data_mb, round(INDEX_LENGTH/1024/1024) index_mb, round(DATA_FREE/1024/1024) free_mb, round(DATA_FREE/DATA_LENGTH*100,2) free_ratio -> from information_schema.TABLES where TABLE_SCHEMA= ’test’ -> and TABLE_NAME= ’user’;+--------------+------------+--------+----------+------------+---------+----------+---------+------------+| table_schema | table_name | ENGINE | total_mb | TABLE_ROWS | data_mb | index_mb | free_mb | free_ratio |+--------------+------------+--------+----------+------------+---------+----------+---------+------------+| test | user | InnoDB | 4 | 50000 | 4 | 0 | 6 | 149.42 |+--------------+------------+--------+----------+------------+---------+----------+---------+------------+1 row in set (0.00 sec)

其中data_free是分配了未使用的字節(jié)數(shù)，并不能說(shuō)明完全是碎片空間。

碎片的回收

對(duì)于InnoDB的表，可以通過(guò)以下命令來(lái)回收碎片，釋放空間，這個(gè)是隨機(jī)讀IO操作，會(huì)比較耗時(shí)，也會(huì)阻塞表上正常的DML運(yùn)行，同時(shí)需要占用額外更多的磁盤空間，對(duì)于RDS來(lái)說(shuō)，可能會(huì)導(dǎo)致磁盤空間瞬間爆滿，實(shí)例瞬間被鎖定，應(yīng)用無(wú)法做DML操作，所以禁止在線上環(huán)境去執(zhí)行。

#執(zhí)行InnoDB的碎片回收mysql> alter table user engine=InnoDB;Query OK, 0 rows affected (9.00 sec)Records: 0 Duplicates: 0 Warnings: 0##執(zhí)行完之后，數(shù)據(jù)文件大小從14MB降低到10M。# ls -lh /data2/mysql/test/user1.ibd -rw-r----- 1 mysql mysql 10M Nov 6 16:18 /data2/mysql/test/user.ibd

mysql> select table_schema, table_name,ENGINE, round(DATA_LENGTH/1024/1024+ INDEX_LENGTH/1024/1024) total_mb,TABLE_ROWS, round(DATA_LENGTH/1024/1024) data_mb, round(INDEX_LENGTH/1024/1024) index_mb, round(DATA_FREE/1024/1024) free_mb, round(DATA_FREE/DATA_LENGTH*100,2) free_ratio from information_schema.TABLES where TABLE_SCHEMA= ’test’ and TABLE_NAME= ’user’;+--------------+------------+--------+----------+------------+---------+----------+---------+------------+| table_schema | table_name | ENGINE | total_mb | TABLE_ROWS | data_mb | index_mb | free_mb | free_ratio |+--------------+------------+--------+----------+------------+---------+----------+---------+------------+| test | user | InnoDB | 5 | 50000 | 5 | 0 | 2 | 44.29 |+--------------+------------+--------+----------+------------+---------+----------+---------+------------+1 row in set (0.00 sec)delete對(duì)SQL的影響未刪除前的SQL執(zhí)行情況

#插入100W數(shù)據(jù)mysql> call insert_user_data(1000000);Query OK, 0 rows affected (35.99 sec)#添加相關(guān)索引mysql> alter table user add index idx_name(name), add index idx_phone(phone);Query OK, 0 rows affected (6.00 sec)Records: 0 Duplicates: 0 Warnings: 0#表上索引統(tǒng)計(jì)信息mysql> show index from user;+-------+------------+-----------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+| Table | Non_unique | Key_name | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | Index_comment |+-------+------------+-----------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+| user | 0 | PRIMARY | 1 | id | A | 996757 | NULL | NULL | | BTREE | ||| user | 1 | idx_name | 1 | name | A | 996757 | NULL | NULL | | BTREE | ||| user | 1 | idx_phone | 1 | phone | A | 2 | NULL | NULL | | BTREE | ||+-------+------------+-----------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+3 rows in set (0.00 sec)#重置狀態(tài)變量計(jì)數(shù)mysql> flush status;Query OK, 0 rows affected (0.00 sec)#執(zhí)行SQL語(yǔ)句mysql> select id, age ,phone from user where name like ’lyn12%’;+--------+-----+-------------+| id | age | phone |+--------+-----+-------------+| 124 | 3 | 15240540354 || 1231 | 30 | 15240540354 || 12301 | 60 | 15240540354 |.............................| 129998 | 37 | 15240540354 || 129999 | 38 | 15240540354 || 130000 | 39 | 15240540354 |+--------+-----+-------------+11111 rows in set (0.03 sec)mysql> explain select id, age ,phone from user where name like ’lyn12%’;+----+-------------+-------+-------+---------------+----------+---------+------+-------+-----------------------+| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |+----+-------------+-------+-------+---------------+----------+---------+------+-------+-----------------------+| 1 | SIMPLE | user | range | idx_name | idx_name | 82 | NULL | 22226 | Using index condition |+----+-------------+-------+-------+---------------+----------+---------+------+-------+-----------------------+1 row in set (0.00 sec)#查看相關(guān)狀態(tài)呢變量mysql> select * from information_schema.session_status where variable_name in(’Last_query_cost’,’Handler_read_next’,’Innodb_pages_read’,’Innodb_data_reads’,’Innodb_pages_read’);+-------------------+----------------+| VARIABLE_NAME | VARIABLE_VALUE |+-------------------+----------------+| HANDLER_READ_NEXT | 11111 | #請(qǐng)求讀的行數(shù)| INNODB_DATA_READS | 7868409 | #數(shù)據(jù)物理讀的總數(shù)| INNODB_PAGES_READ | 7855239 | #邏輯讀的總數(shù)| LAST_QUERY_COST | 10.499000 | #SQL語(yǔ)句的成本COST，主要包括IO_COST和CPU_COST。+-------------------+----------------+4 rows in set (0.00 sec)刪除后的SQL執(zhí)行情況

#刪除50w數(shù)據(jù)mysql> delete from user limit 500000;Query OK, 500000 rows affected (3.70 sec)#分析表統(tǒng)計(jì)信息mysql> analyze table user;+-----------+---------+----------+----------+| Table | Op | Msg_type | Msg_text |+-----------+---------+----------+----------+| test.user | analyze | status | OK |+-----------+---------+----------+----------+1 row in set (0.01 sec)#重置狀態(tài)變量計(jì)數(shù)mysql> flush status;Query OK, 0 rows affected (0.01 sec)mysql> select id, age ,phone from user where name like ’lyn12%’;Empty set (0.05 sec)mysql> explain select id, age ,phone from user where name like ’lyn12%’;+----+-------------+-------+-------+---------------+----------+---------+------+-------+-----------------------+| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |+----+-------------+-------+-------+---------------+----------+---------+------+-------+-----------------------+| 1 | SIMPLE | user | range | idx_name | idx_name | 82 | NULL | 22226 | Using index condition |+----+-------------+-------+-------+---------------+----------+---------+------+-------+-----------------------+1 row in set (0.00 sec)mysql> select * from information_schema.session_status where variable_name in(’Last_query_cost’,’Handler_read_next’,’Innodb_pages_read’,’Innodb_data_reads’,’Innodb_pages_read’);+-------------------+----------------+| VARIABLE_NAME | VARIABLE_VALUE |+-------------------+----------------+| HANDLER_READ_NEXT | 0 || INNODB_DATA_READS | 7868409 || INNODB_PAGES_READ | 7855239 || LAST_QUERY_COST | 10.499000 |+-------------------+----------------+4 rows in set (0.00 sec)

結(jié)果統(tǒng)計(jì)分析

操作 COST 物理讀次數(shù) 邏輯讀次數(shù) 掃描行數(shù) 返回行數(shù) 執(zhí)行時(shí)間 初始化插入100W 10.499000 7868409 7855239 22226 11111 30ms 100W隨機(jī)刪除50W 10.499000 7868409 7855239 22226 0 50ms

這也說(shuō)明對(duì)普通的大表，想要通過(guò)delete數(shù)據(jù)來(lái)對(duì)表進(jìn)行瘦身是不現(xiàn)實(shí)的，所以在任何時(shí)候不要用delete去刪除數(shù)據(jù)，應(yīng)該使用優(yōu)雅的標(biāo)記刪除。

delete優(yōu)化建議控制業(yè)務(wù)賬號(hào)權(quán)限

對(duì)于一個(gè)大的系統(tǒng)來(lái)說(shuō)，需要根據(jù)業(yè)務(wù)特點(diǎn)去拆分子系統(tǒng)，每個(gè)子系統(tǒng)可以看做是一個(gè)service，例如美團(tuán)APP，上面有很多服務(wù)，核心的服務(wù)有用戶服務(wù)user-service，搜索服務(wù)search-service，商品product-service，位置服務(wù)location-service，價(jià)格服務(wù)price-service等。每個(gè)服務(wù)對(duì)應(yīng)一個(gè)數(shù)據(jù)庫(kù)，為該數(shù)據(jù)庫(kù)創(chuàng)建單獨(dú)賬號(hào)，同時(shí)只授予DML權(quán)限且沒(méi)有delete權(quán)限，同時(shí)禁止跨庫(kù)訪問(wèn)。

#創(chuàng)建用戶數(shù)據(jù)庫(kù)并授權(quán)create database mt_user charset utf8mb4;grant USAGE, SELECT, INSERT, UPDATE ON mt_user.* to ’w_user’@’%’ identified by ’t$W*g@gaHTGi123456’;flush privileges;delete改為標(biāo)記刪除

在MySQL數(shù)據(jù)庫(kù)建模規(guī)范中有4個(gè)公共字段，基本上每個(gè)表必須有的，同時(shí)在create_time列要?jiǎng)?chuàng)建索引，有兩方面的好處：

一些查詢業(yè)務(wù)場(chǎng)景都會(huì)有一個(gè)默認(rèn)的時(shí)間段，比如7天或者一個(gè)月，都是通過(guò)create_time去過(guò)濾，走索引掃描更快。 一些核心的業(yè)務(wù)表需要以T +1的方式抽取數(shù)據(jù)倉(cāng)庫(kù)中，比如每天晚上00:30抽取前一天的數(shù)據(jù)，都是通過(guò)create_time過(guò)濾的。

`id` bigint(20) NOT NULL AUTO_INCREMENT COMMENT ’主鍵id’,`is_deleted` tinyint(4) NOT NULL DEFAULT ’0’ COMMENT ’是否邏輯刪除：0：未刪除，1：已刪除’,`create_time` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP COMMENT ’創(chuàng)建時(shí)間’,`update_time` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT ’修改時(shí)間’#有了刪除標(biāo)記，業(yè)務(wù)接口的delete操作就可以轉(zhuǎn)換為updateupdate user set is_deleted = 1 where user_id = 1213;#查詢的時(shí)候需要帶上is_deleted過(guò)濾select id, age ,phone from user where is_deleted = 0 and name like ’lyn12%’;數(shù)據(jù)歸檔方式通用數(shù)據(jù)歸檔方法

#1. 創(chuàng)建歸檔表，一般在原表名后面添加_bak。CREATE TABLE `ota_order_bak` ( `id` bigint(11) NOT NULL AUTO_INCREMENT COMMENT ’主鍵’, `order_id` varchar(255) DEFAULT NULL COMMENT ’訂單id’, `ota_id` varchar(255) DEFAULT NULL COMMENT ’ota’, `check_in_date` varchar(255) DEFAULT NULL COMMENT ’入住日期’, `check_out_date` varchar(255) DEFAULT NULL COMMENT ’離店日期’, `hotel_id` varchar(255) DEFAULT NULL COMMENT ’酒店ID’, `guest_name` varchar(255) DEFAULT NULL COMMENT ’顧客’, `purcharse_time` timestamp NULL DEFAULT NULL COMMENT ’購(gòu)買時(shí)間’, `create_time` datetime DEFAULT NULL, `update_time` datetime DEFAULT NULL, `create_user` varchar(255) DEFAULT NULL, `update_user` varchar(255) DEFAULT NULL, `status` int(4) DEFAULT ’1’ COMMENT ’狀態(tài) ： 1 正常 ， 0 刪除’, `hotel_name` varchar(255) DEFAULT NULL, `price` decimal(10,0) DEFAULT NULL, `remark` longtext, PRIMARY KEY (`id`), KEY `IDX_order_id` (`order_id`) USING BTREE, KEY `hotel_name` (`hotel_name`) USING BTREE, KEY `ota_id` (`ota_id`) USING BTREE, KEY `IDX_purcharse_time` (`purcharse_time`) USING BTREE, KEY `IDX_create_time` (`create_time`) USING BTREE) ENGINE=InnoDB DEFAULT CHARSET=utf8PARTITION BY RANGE (to_days(create_time)) ( PARTITION p201808 VALUES LESS THAN (to_days(’2018-09-01’)), PARTITION p201809 VALUES LESS THAN (to_days(’2018-10-01’)), PARTITION p201810 VALUES LESS THAN (to_days(’2018-11-01’)), PARTITION p201811 VALUES LESS THAN (to_days(’2018-12-01’)), PARTITION p201812 VALUES LESS THAN (to_days(’2019-01-01’)), PARTITION p201901 VALUES LESS THAN (to_days(’2019-02-01’)), PARTITION p201902 VALUES LESS THAN (to_days(’2019-03-01’)), PARTITION p201903 VALUES LESS THAN (to_days(’2019-04-01’)), PARTITION p201904 VALUES LESS THAN (to_days(’2019-05-01’)), PARTITION p201905 VALUES LESS THAN (to_days(’2019-06-01’)), PARTITION p201906 VALUES LESS THAN (to_days(’2019-07-01’)), PARTITION p201907 VALUES LESS THAN (to_days(’2019-08-01’)), PARTITION p201908 VALUES LESS THAN (to_days(’2019-09-01’)), PARTITION p201909 VALUES LESS THAN (to_days(’2019-10-01’)), PARTITION p201910 VALUES LESS THAN (to_days(’2019-11-01’)), PARTITION p201911 VALUES LESS THAN (to_days(’2019-12-01’)), PARTITION p201912 VALUES LESS THAN (to_days(’2020-01-01’)));#2. 插入原表中無(wú)效的數(shù)據(jù)（需要跟開(kāi)發(fā)同學(xué)確認(rèn)數(shù)據(jù)保留范圍）create table tbl_p201808 as select * from ota_order where create_time between ’2018-08-01 00:00:00’ and ’2018-08-31 23:59:59’;#3. 跟歸檔表分區(qū)做分區(qū)交換alter table ota_order_bak exchange partition p201808 with table tbl_p201808;#4. 刪除原表中已經(jīng)規(guī)范的數(shù)據(jù)delete from ota_order where create_time between ’2018-08-01 00:00:00’ and ’2018-08-31 23:59:59’ limit 3000;優(yōu)化后的歸檔方式

#1. 創(chuàng)建中間表CREATE TABLE `ota_order_2020` (........) ENGINE=InnoDB DEFAULT CHARSET=utf8PARTITION BY RANGE (to_days(create_time)) ( PARTITION p201808 VALUES LESS THAN (to_days(’2018-09-01’)), PARTITION p201809 VALUES LESS THAN (to_days(’2018-10-01’)), PARTITION p201810 VALUES LESS THAN (to_days(’2018-11-01’)), PARTITION p201811 VALUES LESS THAN (to_days(’2018-12-01’)), PARTITION p201812 VALUES LESS THAN (to_days(’2019-01-01’)), PARTITION p201901 VALUES LESS THAN (to_days(’2019-02-01’)), PARTITION p201902 VALUES LESS THAN (to_days(’2019-03-01’)), PARTITION p201903 VALUES LESS THAN (to_days(’2019-04-01’)), PARTITION p201904 VALUES LESS THAN (to_days(’2019-05-01’)), PARTITION p201905 VALUES LESS THAN (to_days(’2019-06-01’)), PARTITION p201906 VALUES LESS THAN (to_days(’2019-07-01’)), PARTITION p201907 VALUES LESS THAN (to_days(’2019-08-01’)), PARTITION p201908 VALUES LESS THAN (to_days(’2019-09-01’)), PARTITION p201909 VALUES LESS THAN (to_days(’2019-10-01’)), PARTITION p201910 VALUES LESS THAN (to_days(’2019-11-01’)), PARTITION p201911 VALUES LESS THAN (to_days(’2019-12-01’)), PARTITION p201912 VALUES LESS THAN (to_days(’2020-01-01’)));#2. 插入原表中有效的數(shù)據(jù)，如果數(shù)據(jù)量在100W左右可以在業(yè)務(wù)低峰期直接插入，如果比較大，建議采用dataX來(lái)做，可以控制頻率和大小，之前我這邊用Go封裝了dataX可以實(shí)現(xiàn)自動(dòng)生成json文件，自定義大小去執(zhí)行。insert into ota_order_2020 select * from ota_order where create_time between ’2020-08-01 00:00:00’ and ’2020-08-31 23:59:59’;#3. 表重命名alter table ota_order rename to ota_order_bak; alter table ota_order_2020 rename to ota_order;#4. 插入差異數(shù)據(jù)insert into ota_order select * from ota_order_bak a where not exists (select 1 from ota_order b where a.id = b.id);#5. ota_order_bak改造成分區(qū)表，如果表比較大不建議直接改造，可以先創(chuàng)建好分區(qū)表，通過(guò)dataX把導(dǎo)入進(jìn)去即可。#6. 后續(xù)的歸檔方法#創(chuàng)建中間普遍表create table ota_order_mid like ota_order;#交換原表無(wú)效數(shù)據(jù)分區(qū)到普通表alter table ota_order exchange partition p201808 with table ota_order_mid; ##交換普通表數(shù)據(jù)到歸檔表的相應(yīng)分區(qū)alter table ota_order_bak exchange partition p201808 with table ota_order_mid;

這樣原表和歸檔表都是按月的分區(qū)表，只需要?jiǎng)?chuàng)建一個(gè)中間普通表，在業(yè)務(wù)低峰期做兩次分區(qū)交換，既可以刪除無(wú)效數(shù)據(jù)，又能回收空，而且沒(méi)有空間碎片，不會(huì)影響表上的索引及SQL的執(zhí)行計(jì)劃。

總結(jié)

通過(guò)從InnoDB存儲(chǔ)空間分布，delete對(duì)性能的影響可以看到，delete物理刪除既不能釋放磁盤空間，而且會(huì)產(chǎn)生大量的碎片，導(dǎo)致索引頻繁分裂，影響SQL執(zhí)行計(jì)劃的穩(wěn)定性；

同時(shí)在碎片回收時(shí)，會(huì)耗用大量的CPU，磁盤空間，影響表上正常的DML操作。

在業(yè)務(wù)代碼層面，應(yīng)該做邏輯標(biāo)記刪除，避免物理刪除；為了實(shí)現(xiàn)數(shù)據(jù)歸檔需求，可以用采用MySQL分區(qū)表特性來(lái)實(shí)現(xiàn)，都是DDL操作，沒(méi)有碎片產(chǎn)生。

另外一個(gè)比較好的方案采用Clickhouse，對(duì)有生命周期的數(shù)據(jù)表可以使用Clickhouse存儲(chǔ)，利用其TTL特性實(shí)現(xiàn)無(wú)效數(shù)據(jù)自動(dòng)清理。

到此這篇關(guān)于淺談為什么MySQL不建議delete刪除數(shù)據(jù)的文章就介紹到這了,更多相關(guān)MySQL不建議delete刪除內(nèi)容請(qǐng)搜索好吧啦網(wǎng)以前的文章或繼續(xù)瀏覽下面的相關(guān)文章希望大家以后多多支持好吧啦網(wǎng)！