转储包含所有表的所有数据库，但跳过某些表

在github: https://github.com/rubo77/mysql-backup.sh/blob/master/mysql-backup.sh

#!/bin/bash
# mysql-backup.sh

if [ -z "$1" ] ; then
  echo
  echo "ERROR: root password Parameter missing."
  exit
fi
DB_host=localhost
MYSQL_USER=root
MYSQL_PASS=$1
MYSQL_CONN="-u${MYSQL_USER} -p${MYSQL_PASS}"
#MYSQL_CONN=""

BACKUP_DIR=/backup/mysql/

mkdir $BACKUP_DIR -p

MYSQLPATH=/var/lib/mysql/

IGNORE="database1.table1, database1.table2, database2.table1,"

# strpos $1 $2 [$3]
# strpos haystack needle [optional offset of an input string]
strpos()
{
    local str=${1}
    local offset=${3}
    if [ -n "${offset}" ]; then
        str=`substr "${str}" ${offset}`
    else
        offset=0
    fi
    str=${str/${2}*/}
    if [ "${#str}" -eq "${#1}" ]; then
        return 0
    fi
    echo $((${#str}+${offset}))
}

cd $MYSQLPATH
for i in */; do
    if [ $i != 'performance_schema/' ] ; then 
    DB=`basename "$i"` 
    #echo "backup $DB->$BACKUP_DIR$DB.sql.lzo"
    mysqlcheck "$DB" $MYSQL_CONN --silent --auto-repair >/tmp/tmp_grep_mysql-backup
    grep -E -B1 "note|warning|support|auto_increment|required|locks" /tmp/tmp_grep_mysql-backup>/tmp/tmp_grep_mysql-backup_not
    grep -v "$(cat /tmp/tmp_grep_mysql-backup_not)" /tmp/tmp_grep_mysql-backup

    tbl_count=0
    for t in $(mysql -NBA -h $DB_host $MYSQL_CONN -D $DB -e 'show tables') 
    do
      found=$(strpos "$IGNORE" "$DB"."$t,")
      if [ "$found" == "" ] ; then 
        echo "DUMPING TABLE: $DB.$t"
        mysqldump -h $DB_host $MYSQL_CONN $DB $t --events --skip-lock-tables | lzop -3 -f -o $BACKUP_DIR/$DB.$t.sql.lzo
        tbl_count=$(( tbl_count + 1 ))
      fi
    done
    echo "$tbl_count tables dumped from database '$DB' into dir=$BACKUP_DIR"
    fi
done

在https://stackoverflow.com/a/17016410/1069083的帮助下

它使用更快的lzop，参见:http://pokecraft.first-world.info/wiki/Quick_Benchmark:_Gzip_vs_Bzip2_vs_LZMA_vs_XZ_vs_LZ4_vs_LZO

基于@Brian-Fisher的回答，并回答了一些人对这篇文章的评论，我的数据库中有一堆巨大的(不必要的)表，所以我想在复制时跳过它们的内容，但保持结构:

mysqldump -h <host> -u <username> -p <database> --no-data > db.sql
mysqldump -h <host> -u <username> -p <database> --no-create-info --ignore-table=schema.table1 --ignore-table=schema.table2 >> db.sql

生成的文件在结构上是合理的，但转储的数据现在是~500MB而不是9GB，对我来说要好得多。现在，我可以将文件导入到另一个数据库中进行测试，而不必担心操作9GB的数据或耗尽磁盘空间。

转储包含所有表的所有数据库，但跳过某些表

在github: https://github.com/rubo77/mysql-backup.sh/blob/master/mysql-backup.sh

#!/bin/bash
# mysql-backup.sh

if [ -z "$1" ] ; then
  echo
  echo "ERROR: root password Parameter missing."
  exit
fi
DB_host=localhost
MYSQL_USER=root
MYSQL_PASS=$1
MYSQL_CONN="-u${MYSQL_USER} -p${MYSQL_PASS}"
#MYSQL_CONN=""

BACKUP_DIR=/backup/mysql/

mkdir $BACKUP_DIR -p

MYSQLPATH=/var/lib/mysql/

IGNORE="database1.table1, database1.table2, database2.table1,"

# strpos $1 $2 [$3]
# strpos haystack needle [optional offset of an input string]
strpos()
{
    local str=${1}
    local offset=${3}
    if [ -n "${offset}" ]; then
        str=`substr "${str}" ${offset}`
    else
        offset=0
    fi
    str=${str/${2}*/}
    if [ "${#str}" -eq "${#1}" ]; then
        return 0
    fi
    echo $((${#str}+${offset}))
}

cd $MYSQLPATH
for i in */; do
    if [ $i != 'performance_schema/' ] ; then 
    DB=`basename "$i"` 
    #echo "backup $DB->$BACKUP_DIR$DB.sql.lzo"
    mysqlcheck "$DB" $MYSQL_CONN --silent --auto-repair >/tmp/tmp_grep_mysql-backup
    grep -E -B1 "note|warning|support|auto_increment|required|locks" /tmp/tmp_grep_mysql-backup>/tmp/tmp_grep_mysql-backup_not
    grep -v "$(cat /tmp/tmp_grep_mysql-backup_not)" /tmp/tmp_grep_mysql-backup

    tbl_count=0
    for t in $(mysql -NBA -h $DB_host $MYSQL_CONN -D $DB -e 'show tables') 
    do
      found=$(strpos "$IGNORE" "$DB"."$t,")
      if [ "$found" == "" ] ; then 
        echo "DUMPING TABLE: $DB.$t"
        mysqldump -h $DB_host $MYSQL_CONN $DB $t --events --skip-lock-tables | lzop -3 -f -o $BACKUP_DIR/$DB.$t.sql.lzo
        tbl_count=$(( tbl_count + 1 ))
      fi
    done
    echo "$tbl_count tables dumped from database '$DB' into dir=$BACKUP_DIR"
    fi
done

在https://stackoverflow.com/a/17016410/1069083的帮助下

它使用更快的lzop，参见:http://pokecraft.first-world.info/wiki/Quick_Benchmark:_Gzip_vs_Bzip2_vs_LZMA_vs_XZ_vs_LZ4_vs_LZO

对于多个数据库:

mysqldump -u user -p --ignore-table=db1.tbl1 --ignore-table=db2.tbl1 --databases db1 db2 ..

另一个忽略多个表的例子

/usr/bin/mysqldump -uUSER -pPASS --ignore-table={db_test.test1,db_test.test3} db_test> db_test.sql

使用——ignore-table创建一个表数组，语法为

--ignore-table={db_test.table1,db_test.table3,db_test.table4}

额外的:

导入数据库

 # if file is .sql
 mysql -uUSER  -pPASS db_test < backup_database.sql
 # if file is .sql.gz
 gzip -dc < backup_database.sql.gz | mysql -uUSER -pPASSWORD db_test

简单的脚本，忽略表和导出在.sql.gz节省空间

#!/bin/bash

#tables to ignore
_TIGNORE=(
my_database.table1
my_database.table2
my_database.tablex
)

#create text for ignore tables
_TDELIMITED="$(IFS=" "; echo "${_TIGNORE[*]/#/--ignore-table=}")"

#don't forget to include user and password
/usr/bin/mysqldump -uUSER -pPASSWORD --events ${_TDELIMITED} --databases my_database | gzip -v > backup_database.sql.gz

有帮助你的信息链接

Mysqldump & gzip Mysqldump到tar.gz 如何加载一个sql.gz文件到我的数据库?

注意:在ubuntu服务器上测试，mysql version 14.14 Distrib 5.5.55

一般来说，当您不想或没有时间处理一个巨大的表时，就需要使用这个特性。如果是这种情况，最好使用mysqldump限制结果集的——where选项。例如:mysqldump -uuser -ppass database——where="1 = 1 LIMIT 500000" > resultset.sql。