从全备中恢复单库或单表，小心有坑！

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"前言："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"MySQL 逻辑备份工具最常用的就是 mysqldump 了，一般我们都是备份整个实例或部分业务库。不清楚你有没有做过恢复，恢复场景可能就比较多了，比如我想恢复某个库或某个表等。那么如何从全备中恢复单库或单表，这其中又有哪些隐藏的坑呢？这篇文章我们一起来看下。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","marks":[{"type":"bgcolor","attrs":{"color":"#FC8F99","name":"red"}}],"text":" "},{"type":"text","text":" 1.如何恢复单库或单表"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https://xie.infoq.cn/article/cb36e11d7c4088ad60065fdcb","title":""},"content":[{"type":"text","text":"前面文章"}]},{"type":"text","text":"有介绍过 MySQL 的备份与恢复。可能我们每个数据库实例中都不止一个库，一般备份都是备份整个实例，但恢复需求又是多种多样的，比如说我想只恢复某个库或某张表，这个时候应该怎么操作呢？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"如果你的实例数据量不大，可以在另外一个环境恢复出整个实例，然后再单独备份出所需库或表用来恢复。不过这种方法不够灵活，并且只适用数据量比较少的情况。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"其实从全备中恢复单库还是比较方便的，有个 "},{"type":"codeinline","content":[{"type":"text","text":"--one-database"}]},{"type":"text","text":" 参数可以指定单库恢复，下面来具体演示下："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"sql"},"content":[{"type":"text","text":"# 查看及备份所有库\nmysql> show databases;\n+--------------------+\n| Database |\n+--------------------+\n| information_schema |\n| mysql |\n| performance_schema |\n| sbtest |\n| sys |\n| testdb |\n| testdb2 |\n+--------------------+\n\nmysqldump -uroot -pxxxx -R -E --single-transaction --all-databases > all_db.sql\n\n# 删除testdb库 并进行单库恢复\nmysql> drop database testdb;\nQuery OK, 36 rows affected (2.06 sec)\n\n# 貌似恢复前 testdb库不存在的话要手动新建\nmysql -uroot -pxxxx --one-database testdb < all_db.sql"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"除了上述方法外，恢复单库或单表还可以采用手动筛选的方法。这个时候 Linux 下大名鼎鼎的 "},{"type":"codeinline","content":[{"type":"text","text":"sed"}]},{"type":"text","text":" 和 "},{"type":"codeinline","content":[{"type":"text","text":"grep"}]},{"type":"text","text":" 命令就派上用场了，我们可以利用这两个命令从全备中筛选出单库或单表的语句，筛选方法如下："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"sql"},"content":[{"type":"text","text":"# 从全备中恢复单库\nsed -n '/^-- Current Database: `testdb`/,/^-- Current Database: `/p' all_db.sql > testdb.sql\n\n# 筛选出单表语句\ncat all_db.sql | sed -e '/./{H;$!d;}' -e 'x;/CREATE TABLE `test_tb`/!d;q' > /tmp/test_tb_info.sql \ncat all_db.sql | grep --ignore-case 'insert into `test_tb`' > /tmp/test_tb_data.sql"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","marks":[{"type":"bgcolor","attrs":{"color":"#FC8F99","name":"red"}}],"text":" "},{"type":"text","text":" 2.小心有坑"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"对于上述手动筛选来恢复单库或单表的方法，看起来简单方便，其实隐藏着一个小坑，下面我们来具体演示下："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"sql"},"content":[{"type":"text","text":"# 备份整个实例\nmysqldump -uroot -pxxxx -R -E --single-transaction --all-databases > all_db.sql\n\n# 手动备份下test_tb 然后删除test_tb\nmysql> create table test_tb_bak like test_tb;\nQuery OK, 0 rows affected (0.03 sec)\n\nmysql> insert into test_tb_bak select * from test_tb;\nQuery OK, 4 rows affected (0.02 sec)\nRecords: 4 Duplicates: 0 Warnings: 0\n\nmysql> drop table test_tb;\nQuery OK, 0 rows affected (0.02 sec)\n\n# 从全备中筛选test_db建表及插数据语句\ncat all_db.sql | sed -e '/./{H;$!d;}' -e 'x;/CREATE TABLE `test_tb`/!d;q' > test_tb_info.sql \ncat all_db.sql | grep --ignore-case 'insert into `test_tb`' > test_tb_data.sql\n\n# 查看得到的语句 貌似没问题\ncat test_tb_info.sql\n\nDROP TABLE IF EXISTS `test_tb`;\n/*!40101 SET @saved_cs_client = @@character_set_client */;\n/*!40101 SET character_set_client = utf8 */;\nCREATE TABLE `test_tb` (\n `inc_id` int(11) NOT NULL AUTO_INCREMENT COMMENT '自增主键',\n `col1` int(11) NOT NULL,\n `col2` varchar(20) DEFAULT NULL,\n `col_dt` datetime DEFAULT NULL,\n `create_time` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP COMMENT '创建时间',\n `update_time` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP COMMENT '修改时间',\n PRIMARY KEY (`inc_id`)\n) ENGINE=InnoDB AUTO_INCREMENT=5 DEFAULT CHARSET=utf8 COMMENT='测试表';\n/*!40101 SET character_set_client = @saved_cs_client */;\n\ncat test_tb_data.sql\n\nINSERT INTO `test_tb` VALUES (1,1001,'dsfs','2020-08-04 12:12:36','2020-09-17 06:19:27','2020-09-17 06:19:27'),\n(2,1002,'vfsfs','2020-09-04 12:12:36','2020-09-17 06:19:27','2020-09-17 06:19:27'),\n(3,1003,'adsfsf',NULL,'2020-09-17 06:19:27','2020-09-17 06:19:27'),\n(4,1004,'walfd','2020-09-17 14:19:27','2020-09-17 06:19:27','2020-09-18 07:52:13');\n\n# 执行恢复单表操作\nmysql -uroot -pxxxx testdb < test_tb_info.sql\nmysql -uroot -pxxxx testdb < test_tb_data.sql\n\n# 查看恢复数据 并和备份表比对\nmysql> select * from test_tb;\n+--------+------+--------+---------------------+---------------------+---------------------+\n| inc_id | col1 | col2 | col_dt | create_time | update_time |\n+--------+------+--------+---------------------+---------------------+---------------------+\n| 1 | 1001 | dsfs | 2020-08-04 12:12:36 | 2020-09-17 06:19:27 | 2020-09-17 06:19:27 |\n| 2 | 1002 | vfsfs | 2020-09-04 12:12:36 | 2020-09-17 06:19:27 | 2020-09-17 06:19:27 |\n| 3 | 1003 | adsfsf | NULL | 2020-09-17 06:19:27 | 2020-09-17 06:19:27 |\n| 4 | 1004 | walfd | 2020-09-17 14:19:27 | 2020-09-17 06:19:27 | 2020-09-18 07:52:13 |\n+--------+------+--------+---------------------+---------------------+---------------------+\n4 rows in set (0.00 sec)\n\nmysql> select * from test_tb_bak;\n+--------+------+--------+---------------------+---------------------+---------------------+\n| inc_id | col1 | col2 | col_dt | create_time | update_time |\n+--------+------+--------+---------------------+---------------------+---------------------+\n| 1 | 1001 | dsfs | 2020-08-04 12:12:36 | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 |\n| 2 | 1002 | vfsfs | 2020-09-04 12:12:36 | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 |\n| 3 | 1003 | adsfsf | NULL | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 |\n| 4 | 1004 | walfd | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 | 2020-09-18 15:52:13 |\n+--------+------+--------+---------------------+---------------------+---------------------+\n4 rows in set (0.00 sec)"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"如果你仔细观察的话，会发现恢复出来的数据有问题，貌似时间不太对，你再仔细看看，是不是有的时间差了8小时！详细探究下来，我们发现 timestamp 类型字段的时间数据恢复有问题，准确来讲备份文件中记录的是0时区，而我们系统一般采用东八区，所以才会出现误差8小时的问题。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"那么你会问了，为什么全部恢复不会出问题呢？问的好，我们看下备份文件就知道了。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"sql"},"content":[{"type":"text","text":"# 备份文件开头\n-- MySQL dump 10.13 Distrib 5.7.23, for Linux (x86_64)\n--\n-- Host: localhost Database:\n-- ------------------------------------------------------\n-- Server version 5.7.23-log\n\n/*!40101 SET @OLD_CHARACTER_SET_CLIENT=@@CHARACTER_SET_CLIENT */;\n/*!40101 SET @OLD_CHARACTER_SET_RESULTS=@@CHARACTER_SET_RESULTS */;\n/*!40101 SET @OLD_COLLATION_CONNECTION=@@COLLATION_CONNECTION */;\n/*!40101 SET NAMES utf8 */;\n/*!40103 SET @OLD_TIME_ZONE=@@TIME_ZONE */;\n/*!40103 SET TIME_ZONE='+00:00' */; \n注意上面两行\n/*!40014 SET @OLD_UNIQUE_CHECKS=@@UNIQUE_CHECKS, UNIQUE_CHECKS=0 */;\n/*!40014 SET @OLD_FOREIGN_KEY_CHECKS=@@FOREIGN_KEY_CHECKS, FOREIGN_KEY_CHECKS=0 */;\n/*!40101 SET @OLD_SQL_MODE=@@SQL_MODE, SQL_MODE='NO_AUTO_VALUE_ON_ZERO' */;\n/*!40111 SET @OLD_SQL_NOTES=@@SQL_NOTES, SQL_NOTES=0 */;\n\n\n# 备份文件结尾\n/*!40103 SET TIME_ZONE=@OLD_TIME_ZONE */;\n/*!40101 SET SQL_MODE=@OLD_SQL_MODE */;\n/*!40014 SET FOREIGN_KEY_CHECKS=@OLD_FOREIGN_KEY_CHECKS */;\n/*!40014 SET UNIQUE_CHECKS=@OLD_UNIQUE_CHECKS */;\n/*!40101 SET CHARACTER_SET_CLIENT=@OLD_CHARACTER_SET_CLIENT */;\n/*!40101 SET CHARACTER_SET_RESULTS=@OLD_CHARACTER_SET_RESULTS */;\n/*!40101 SET COLLATION_CONNECTION=@OLD_COLLATION_CONNECTION */;\n/*!40111 SET SQL_NOTES=@OLD_SQL_NOTES */;\n\n-- Dump completed on 2020-09-18 15:56:40"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"仔细看备份文件，你会发现 mysqldump 备份出来的文件中，首先会将会话时区改为0，结尾处再改回原时区。这就代表着，备份文件中记录的时间戳数据都是以0时区为基础的。如果直接执行筛选出的SQL，就会造成0时区的时间戳插入的东八区的系统中，显然会造成时间相差8小时的问题。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"看到这里，不知道你是否看懂了呢，可能有过备份恢复经验的同学好理解些。解决上述问题的方法也很简单，那就是在执行SQL文件前，更改当前会话时区为0，再次来演示下："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":"sql"},"content":[{"type":"text","text":"# 清空test_db表数据\nmysql> truncate table test_tb;\nQuery OK, 0 rows affected (0.02 sec)\n\n# 文件开头增加时区声明\nvim test_tb_data.sql\nset session TIME_ZONE='+00:00';\nINSERT INTO `test_tb` VALUES (1,1001,'dsfs','2020-08-04 12:12:36','2020-09-17 06:19:27','2020-09-17 06:19:27'),\n(2,1002,'vfsfs','2020-09-04 12:12:36','2020-09-17 06:19:27','2020-09-17 06:19:27'),\n(3,1003,'adsfsf',NULL,'2020-09-17 06:19:27','2020-09-17 06:19:27'),\n(4,1004,'walfd','2020-09-17 14:19:27','2020-09-17 06:19:27','2020-09-18 07:52:13');\n\n# 执行恢复并比对 发现数据正确\nmysql> select * from test_tb;\n+--------+------+--------+---------------------+---------------------+---------------------+\n| inc_id | col1 | col2 | col_dt | create_time | update_time |\n+--------+------+--------+---------------------+---------------------+---------------------+\n| 1 | 1001 | dsfs | 2020-08-04 12:12:36 | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 |\n| 2 | 1002 | vfsfs | 2020-09-04 12:12:36 | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 |\n| 3 | 1003 | adsfsf | NULL | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 |\n| 4 | 1004 | walfd | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 | 2020-09-18 15:52:13 |\n+--------+------+--------+---------------------+---------------------+---------------------+\n4 rows in set (0.00 sec)\n\nmysql> select * from test_tb_bak;\n+--------+------+--------+---------------------+---------------------+---------------------+\n| inc_id | col1 | col2 | col_dt | create_time | update_time |\n+--------+------+--------+---------------------+---------------------+---------------------+\n| 1 | 1001 | dsfs | 2020-08-04 12:12:36 | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 |\n| 2 | 1002 | vfsfs | 2020-09-04 12:12:36 | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 |\n| 3 | 1003 | adsfsf | NULL | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 |\n| 4 | 1004 | walfd | 2020-09-17 14:19:27 | 2020-09-17 14:19:27 | 2020-09-18 15:52:13 |\n+--------+------+--------+---------------------+---------------------+---------------------+\n4 rows in set (0.00 sec)"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"总结："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我们在网络中很容易搜索出恢复单库或单表的方法，大多都有提到上述利用 sed 、grep 命令来手动筛选的方法。但大部分文章都未提及可能出现的问题，如果你的表字段有timestamp 类型，用这种方法要格外注意。无论面对哪种恢复需求，我们都要格外小心，不要造成越恢复越糟糕的情况，最好有个空实例演练下，然后再进行恢复。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}}]}