分布式数据库的数据备份/恢复，这些你一定要了解 [数据库教程]
IP
Hostname
SequoiaDB用户
备注
192.168.1.3
sdb03
sdbadmin
192.168.1.4
sdb04
sdbadmin
192.168.1.5
sdb05
sdbadmin
备份磁盘(挂载目录/sdbdata/backup）
[[email protected] ~]# cat /etc/exports/sdbdata/backup *(insecure,rw,sync,no_wdelay,insecure_locks,no_root_squash)
[[email protected] ~]# mount -t nfs -o rw,bg,hard,nointr,tcp 192.168.1.5:/sdbdata/backup /sdbdata/backup
[[email protected] ~]# chown -R sdbadmin:sdbadmin_group /sdbdata/backup
[[email protected] ~]# mount -t nfs -o rw,bg,hard,nointr,tcp 192.168.1.5:/sdbdata/backup /sdbdata/backup
[[email protected] ~]# chown -R sdbadmin:sdbadmin_group /sdbdata/backup
[[email protected] ~]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/centos-root 15G 4.4G 11G 29% /
/dev/vdb 985G 129G 806G 14% /sdbdata
/dev/vda1 497M 172M 326M 35% /boot
/dev/vda3 50G 13G 38G 26% /opt
192.168.1.5:/sdbdata/backup 985G 129G 807G 14% /sdbdata/backu
[[email protected] ~]#cat backup_full.sh#!/bin/bash
date
/opt/sequoiasql/mysql/bin/mysqldump --login-path=mysql -A -d > /sdbdata/backup/full/create`date +%y%m%d%H`.sql
/opt/sequoidb/bin/sdblist -l -m list > /sdbdata/backup/full/sdblist`date +%y%m%d%H`.sql
/opt/sequoiadb/bin/sdb ‘db=new Sdb()‘
/opt/sequoiadb/bin/sdb ‘db.backup ( { Name : "cluster_backup", Path : "/sdbdata/backup/full/%g", Overwrite : true, Description : "full_backup" } ) ;‘
date
db.backup()备份常用参数说明
Name:备份名称，缺省则以当前时间格式命名，如“2016-01-01-15:00:00”，格式为“YYYY-MM-DD-HH:mm:ss”。
Description:备份用户描述信息。
Path:本次备份的指定路径，缺省为配置参数“bkuppath”中指定的路径。
EnsureInc:备份方式，true 表示增量备份，false 表示全量备份，缺省为 false。
OverWrite:对于同名备份是否覆盖，true 表示覆盖，false 表示不覆盖，如果同名则报错；缺省为 false。
GroupName:对指定组进行备份，缺省为对全系统备份，当需要对多个组进行备份可以指定为数组类型，如：["datagroup1", "datagroup2"]。
[[email protected] ~]#cat backup_incre.sh#!/bin/bash
date
/opt/sequoiasql/mysql/bin/mysqldump --login-path=mysql -A -d > /sdbdata/backup/full/create`date +%y%m%d%H`.sql
/opt/sequoidb/bin/sdblist -l -m list > /sdbdata/backup/full/sdblist`date +%y%m%d%H`.sql
/opt/sequoiadb/bin/sdb ‘db=new Sdb();‘
/opt/sequoiadb/bin/sdb ‘db.backup ( { Name : "cluster_backup", Path : "/sdbdata/backup/full/%g", EnsureInc : true } );‘
date
(1)配置单向ssh 免密服务[[email protected] ~]#ssh-keygen -t rsa
[[email protected] ~]#ssh-copy-id [email protected]
[[email protected] ~]#ssh-copy-id [email protected]
[[email protected] ~]#ssh-copy-id [email protected]
(2)恢复脚本
[[email protected] ~]#cat restore_full.sh
#!/bin/bash
#停止集群
for hostname in {sdb03,sdb04,sdb05}
do
ssh $hostname /opt/sequoiadb/bin/sdbstop -t all
done
#全量恢复一份完整的副本数据
for groupname in `cat /sdbdata/backup/full/sdblist20200717.sql | awk ‘{print $8}‘ | grep -Evi "GroupName|SYSCoord" `
do
/opt/sequoiadb/bin/sdbrestore -p /sdbdata/backup/full/$groupname/ -n cluster_backup -b 0 -i 0
done
#删除另外2台服务器之前的副本数据，scp拷贝全量恢复的副本数据
for hostname in {sdb03,sdb04}
do
for dbpath in `cat /sdbdata/backup/full/sdblist20200717.sql | awk ‘{print $10}‘| grep -Evi "dbpath|*11810"`
do
ssh -t [email protected]$hostname "rm -rf "$dbpath*
scp -pr $dbpath* $hostname:$dbpath &
done
done
#启动集群
/bin/read -p " Do you want to start SequoiaDB(yes or no)? " Do
while [[ "$Do" != "no" ]] && [[ "$Do" != "yes" ]]
do
/bin/read -p "Do you want to start SequoiaDB(yes/no)?" Do
done
if [ "$Do" == "yes" ]
then
for hostname in {sdb03,sdb04,sdb05}
do
ssh $hostname /opt/sequoiadb/bin/sdbstart -t all
done
elif [ "$Do" == "no" ]
then
echo "please start SequoiaDB by hand"
fi
sdbrestore数据恢复常用参数：
--bkpath -p:备份源数据所在路径。
--increaseid -i:需要恢复到第几次增量备份，缺省恢复到最后一次 ( -1 ）。
--beginincreaseid -b:需要从第几次备份开始恢复，缺省由系统自动计算 ( -1 )。
--bkname -n:需要恢复的备份名。
--action -a:恢复行为，“restore”表示恢复，“list”表示查看备份信息，缺省为“restore”。
--diaglevel -v:恢复工具自身的日志级别，缺省为 WARNING ( 3
[[email protected] ~]#cat restore_incre.sh#!/bin/bash
#停止集群
for hostname in {sdb03,sdb04,sdb05}
do
ssh $hostname /opt/sequoiadb/bin/sdbstop -t all
done
#增量恢复副本数据
for groupname in `cat /sdbdata/backup/full/sdblist20200717.sql | awk ‘{print $8}‘ | grep -Evi "GroupName|SYSCoord"`
do
/opt/sequoiadb/bin/sdbrestore -p /sdbdata/backup/full/$groupname/ -n cluster_backup -b -1
done
#启动集群
/bin/read -p " Do you want to start SequoiaDB(yes or no)? " Do
while [[ "$Do" != "no" ]] && [[ "$Do" != "yes" ]]
do
/bin/read -p "Do you want to start SequoiaDB(yes/no)?" Do
done
if [ "$Do" == "yes" ]
then
for hostname in {sdb03,sdb04,sdb05}
do
ssh $hostname /opt/sequoiadb/bin/sdbstart -t all
done
elif [ "$Do" == "no" ]
then
echo "please start SequoiaDB by hand"
fi
(1)导出集合sbtest.sbtest1 的数据sdbexprt --hostname "localhost" --svcname "11810" --user "sdbadmin" --password "sdbadmin"
--type ‘json‘ --csname ‘sbtest‘ --clname ‘sbtestl‘ --file ‘/tmp/sbtest.sbtest1.json‘
(2)导出集合空间sbtest下，所有集合的数据
sdbexprt --hostname "localhost" --svcname "11810" --user "sdbadmin" --password "sdbadmin"
--type ‘json‘ --cscl ‘sbtest‘ --dir ‘/tmp‘
sdbimprt --hosts ‘localhost:11810‘ --user ‘sdbadmin‘ --password ‘sdbadmin‘--csname ‘sbtest‘ --clname ‘sbtest1‘ --insertnum 10000 --jobs 20 --type ‘json‘ --file ‘/tmp/sbtest.sbtest1.json‘
--coord false --ignorenull true --verbose true --force false --errorstop true --sharding true --transaction false --allowkeydup true
(1)备份所有库:mysqldump -h 192.168.3.6 -P 3306 -u root -p root123456 -A > /backup/all.sql
(2)备份几个库：
mysqldump --h 192.168.3.6 -P 3306 -u root -p root123456 -B 库名1 库名2 > /backup/database.sql
(3)备份单个库某几个表（表名用空格隔开即可）
Mysqldump -h 192.168.3.6 -P 3306 -u root -p root123456 库名 表名1 表名2> /backup/table.sql
(4)mysqldump恢复
mysql -h 192.168.3.6 -P 3306 -u root -p root123456 < /backup/all.sql;
或者
MySQL>source /backup/all.sql
$time mydumper -h 192.168.3.6 -P 3306 -u root -p root123456 -t 6 -c -e -B sbtest -o /home/data/$time myloader -h 192.168.3.6 -P 3306 -u root -p root123456 -t 6 -B sbtest -o -d /home/data
分布式数据库的数据备份/恢复，这些你一定要了解
以上是 分布式数据库的数据备份/恢复，这些你一定要了解 [数据库教程] 的全部内容， 来源链接： utcz.com/z/534815.html
