Jan 15, 2024 No Comments Oracle RMAN全量+增量备份脚本--生产环境可用 ``` su - oracle mkdir /backup/rmanbak ``` **1\. 全量备份** 如果要用RMAN做迁移，要打开块跟踪，可以提升增量备份的速度。打开和关闭块跟踪的方法见第五部分。 根据cpu核数，设置备份通道数。 ``` vi /backup/rmanbak/RMAN_L0_instance_bak.sh #! /bin/bash export NLS_LANG='american_america.AL32UTF8' source /home/oracle/.bash_profile current_day=`date +%Y%m%d` if [ ! -d /backup/rmanbak/$current_day ] then mkdir -p /backup/rmanbak/$current_day rman target / <<EOF run{ allocate channel d1 type disk; allocate channel d2 type disk; allocate channel d3 type disk; allocate channel d4 type disk; allocate channel d5 type disk; allocate channel d6 type disk; allocate channel d7 type disk; allocate channel d8 type disk; backup as compressed backupset incremental level 0 database format '/backup/rmanbak/$current_day/ORCL_L0_instance_%s_%d_%T_%u.bkp'; sql 'alter system archive log current'; sql 'alter system archive log current'; sql 'alter system archive log current'; backup archivelog all format '/backup/rmanbak/$current_day/ORCL_arclog_instance_%s_%d_%T_%u.bkp' not backed up 1 times; backup spfile format '/backup/rmanbak/$current_day/ORCL_spfile_instance_%s_%d_%T_%u.bkp'; backup current controlfile format '/backup/rmanbak/$current_day/ORCL_controlfile_instance_%s_%d_%T_%u.bkp'; delete noprompt ARCHIVELOG until time 'SYSDATE-30'; crosscheck archivelog all; delete noprompt expired archivelog all; crosscheck backup; delete expired backup; release channel d1; release channel d2; release channel d3; release channel d4; release channel d5; release channel d6; release channel d7; release channel d8; } EOF else echo "Database full backup on $current_day has been completed;" fi cd /backup/rmanbak tar -zcvf $current_day.tar.gz $current_day rm -rf $current_day ``` **2\. 增量备份** ``` vi /backup/rmanbak/RMAN_L1_instance_bak.sh #! /bin/bash export NLS_LANG='american_america.AL32UTF8' source /home/oracle/.bash_profile current_day=`date +%Y%m%d` if [ ! -d/backup/rmanbak/$current_day ] then mkdir -p/backup/rmanbak/$current_day fi rman target / <<EOF run{ allocate channel d1 type disk; allocate channel d2 type disk; allocate channel d3 type disk; allocate channel d4 type disk; allocate channel d5 type disk; allocate channel d6 type disk; allocate channel d7 type disk; allocate channel d8 type disk; backup as compressed backupset incremental level 1 cumulative database format '/backup/rmanbak/$current_day/ORCL_L1_instance_%s_%d_%T_%u.bkp'; sql 'alter system archive log current'; sql 'alter system archive log current'; sql 'alter system archive log current'; backup archivelog all format '/backup/rmanbak/$current_day/ORCL_arclog_instance_%s_%d_%T_%u.bkp' not backed up 1 times; backup spfile format '/backup/rmanbak/$current_day/ORCL_spfile_instance_%s_%d_%T_%u.bkp'; backup current controlfile format '/backup/rmanbak/$current_day/ORCL_controlfile_instance_%s_%d_%T_%u.bkp'; delete noprompt ARCHIVELOG until time 'SYSDATE-30'; crosscheck archivelog all; delete noprompt expired archivelog all; crosscheck backup; delete expired backup; release channel d1; release channel d2; release channel d3; release channel d4; release channel d5; release channel d6; release channel d7; release channel d8; } EOF else echo "Database incremental backup on $current_day has been completed;" fi cd /backup/rmanbak tar -zcvf $current_day.tar.gz $current_day rm -rf $current_day ``` **3\. 定期删除RMAN备份** ``` vi /backup/rmanbak/del_rman_instance_bak.sh #!/bin/bash removedir=/backup/rmanbak dt=`date +%Y%m%d -d "30 day ago"` Rdir=$removedir for subdir in `ls $Rdir`; do subdir=`echo ${subdir//.*/}` #subdir=`echo ${subdir}` if [ "${subdir}" -lt "${dt}" ]; then rm -rf $Rdir/$subdir.tar.gz >/dev/null #echo "the directory $Rdir/$subdir has been removed." fi chmod 775 /backup/rmanbak/RMAN_L0_instance_bak.sh chmod 775 /backup/rmanbak/RMAN_L1_instance_bak.sh chmod 775 /backup/rmanbak/del_rman_instance_bak.sh ``` **4\. 设置自动备份任务** ``` crontab -e 10 0 * * 1 /backup/rmanbak/del_rman_instance_bak.sh >> /backup/rmanbak/del_rman_backup.log 2>&1 30 0 * * 0 /backup/rmanbak/RMAN_L0_instance_bak.sh >> /backup/rmanbak/rman_backup_L0.log 2>&1 30 0 * * 1,2,3,4,5,6 /backup/rmanbak/RMAN_L1_instance_bak.sh >> /backup/rmanbak/rman_backup_L1.log 2>&1 ``` **5\. 相关ORACLE知识** > Block Change Tracking（块跟踪）主要用于RMAN备份的增量备份，将自从上一次备份以来数据块的变化记录到**block change tracking文件中**，相关后台进程CTWR（Change Tracking Writer）, > > 主要目的是改善增量备份性能，RMAN可以不再扫描整个文件以查找变更数据。第一个0级的增量备份扫描整个datafile。 随后的增量备份使用block changetracking file的信息，只扫描自上次备份以 > > 来被标记为change 的block，RMAN 不会对该文件进行备份。。从10g开始提供此功能。 > > **trace 记录保留** > > Oracle 会自动管理change tracking file的大小，只保留最近最近8次blockchange 的信息。 超过8次， 那么最前面的block bitmap 信息会被current change 覆盖。要考虑change tracking file > > 的8次限制与备份策略的关系。 如果我们在0级备份之后又进行了7次不同的增量备份，那么这时block change tracking 就包含了8个bitmaps信息。 如果我们在进行1级的增量备份，那么此时的增量 > > 就不是最优的， 还会扫描整个data file。 因为我们之前0级的bitmap信息被我们当前的bitmap 信息覆盖掉了，这样就没有了参考的依据。所以，两次全量备份的时间间隔要控制在8天以内（包含）。 > > **trace 文件存放** > > oracle 不建议使用RAWdevice来存放block tracking. 如果是RAC一定要放到共享存储上。 > > **trace 文件重置** > > + change tracking file 变为无效。 > + RMAN 进行恢复时。 > > **文件大小** > > + 与数据库中的数据文件总大小相关，与数据量无关。每次递增10M，初始化文件时，300G 数据文件会配置10M ,600G 数据文件会配置20m .1T 及以上数据文件以32M 为标准。 > + 每个datafile，在tracking file里最少需要分配320k的空间，所以如果有大量的小data file， change tracking file 也会相对较大。 **打开和关闭块跟踪** ``` RAC: #查看是否开启了块跟踪 select * from v$block_change_tracking; #打开块跟踪 alter database enable block change tracking using file '+DATA'; #关闭块跟踪 alter database disable block change tracking; 非RAC: #打开块跟踪 alter database enable block change tracking using file '/backup/block.trc'; #关闭块跟踪 alter database disable block change tracking; ``` **【如果提高RMAN备份速度】** ``` 1. 开启多通道 通道数据与cpu核数有关系 2. 设置块跟踪 用于提升增量备份的速度（因为备份时只检查修改过的块） 3. 设置压缩级别为medium 综合看，medium的备份和恢复速度比较好。 CONFIGURE BACKUP OPTIMIZATION ON; CONFIGURE COMPRESSION ALGORITHM 'MEDIUM' AS OF RELEASE 'DEFAULT' OPTIMIZE FOR LOAD TRUE; ``` 最后更新于 2024-01-15 15:22:21 并被添加「」标签，已有 918 位童鞋阅读过。 本站使用「署名 4.0 国际」创作共享协议，可自由转载、引用，但需署名作者且注明文章出处
此处评论已关闭