Question

专家，我拥有Oracle数据库11g版本11.2.0.3.0 - 使用Real Application Clusters和自动存储管理进行64位生产。

操作系统是Linux Red Hat 2.6.18-348.12.1.el5。

数据库名称：dbname

数据库大小约为92 GB。

使用RMAN执行备份需要大约4小时45分钟才能完成任务，并且会损坏与数据库交互的其他进程的正确执行。

这些表每天都有很多DML（udpate / insert / delete）。

我在rman日志中看到两个评论家的时刻：

Starting backup at 04-04-2017 04:33:59
channel dbname_backup_disk1: starting incremental level 1 datafile backup set
channel dbname_backup_disk1: specifying datafile(s) in backup set
input datafile file number=00017 name=file_name ....                                                                        
...   "here all files"                                                                         
channel dbname_backup_disk1: starting piece 1 at 04-04-2017 04:34:00
channel dbname_backup_disk1: finished piece 1 at 04-04-2017 07:22:47
piece handle=+RECOVERY/dbname/backupset/2017_04_04/nnndn1_dbname_level_0_0.2614.940394043 tag=DBNAME_LEVEL_0 comment=NONE
channel dbname_backup_disk1: backup set complete, ***elapsed time: 02:48:48***

和

 Starting backup at 04-04-2017 07:46:20
 backup will be obsolete on date 04-07-2017 07:46:20
 archived logs required to recover from this backup will expire when this 
 backup expires
 channel dev1: starting compressed full datafile backup set
 channel dev1: starting piece 1 at 04-04-2017 07:46:21
 channel dev1: finished piece 1 at 04-04-2017 09:22:07
 piece handle=/backup/oracle/backup/DBNAME_940405581_6656_1 tag=TAG20170404T074620 comment=NONE
 channel dev1: backup set complete, ***elapsed time: 01:35:46***
 Finished backup at 04-04-2017 09:22:07

请帮我找到减少所用时间的关键。感谢

下面的行是主脚本的一部分，名为backup.pl。每天上午02:30从当地（UTC -3）执行，来自crontab。

RMAN> run {
    show all;
    CONFIGURE RETENTION POLICY TO RECOVERY WINDOW OF 1 DAYS;
    configure controlfile autobackup on;
    CONFIGURE CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO '+RECOVERY/DBTARGET/%F';
    CONFIGURE CHANNEL DEVICE TYPE DISK FORMAT '+RECOVERY/DBTARGET/%d_%t_%s_%p';
    allocate channel dbname_backup_disk1 device type disk;
    REPORT OBSOLETE;
    DELETE OBSOLETE;
    recover copy of database with tag 'dbname_LEVEL_0' until time 'sysdate-1';
    backup incremental level 1 cumulative copies=1 for recover of copy with tag 'dbname_LEVEL_0' database include current controlfile;
    backup archivelog all not backed up format '+RECOVERY/DBTARGET/%d_%t_s%s_s%p';
    backup current controlfile for standby;
    delete archivelog until time='sysdate-3';
    release channel dbtarget_backup_disk1;
    sql "create pfile=''/backup/oracle/backup/ros1or01-initdbtarget1-20170404.ora'' from spfile";
    }
    allocate channel for maintenance type disk;
    delete noprompt obsolete device type disk;
    CROSSCHECK BACKUP;
    CROSSCHECK DATAFILECOPY ALL;
    CHANGE ARCHIVELOG ALL CROSSCHECK;
    DELETE EXPIRED ARCHIVELOG ALL;
    REPORT OBSOLETE;
    DELETE OBSOLETE;
    release channel;
    run
    {
    CONFIGURE CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO '/backup/oracle/backup/%F';
    allocate channel dev1 device type disk format '/backup/oracle/backup/%d_%t_%s_%p';
    backup backupset completed after 'sysdate-3/24';
    backup as compressed backupset datafilecopy all noduplicates KEEP UNTIL TIME 'SYSDATE+3' logs;
    release channel dev1;
    CONFIGURE CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO '+RECOVERY/DBTARGET/%F';
     }
     quit
    using target database control file instead of recovery catalog
    RMAN configuration parameters for database with db_unique_name DBTARGET are:
    CONFIGURE RETENTION POLICY TO RECOVERY WINDOW OF 1 DAYS;
    CONFIGURE BACKUP OPTIMIZATION OFF; # default
    CONFIGURE DEFAULT DEVICE TYPE TO DISK; # default
    CONFIGURE CONTROLFILE AUTOBACKUP ON;
    CONFIGURE CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO '+RECOVERY/DBTARGET/%F';
    CONFIGURE DEVICE TYPE DISK PARALLELISM 1 BACKUP TYPE TO BACKUPSET; # default
    CONFIGURE DATAFILE BACKUP COPIES FOR DEVICE TYPE DISK TO 1; # default
    CONFIGURE ARCHIVELOG BACKUP COPIES FOR DEVICE TYPE DISK TO 1; # default
    CONFIGURE CHANNEL DEVICE TYPE DISK FORMAT  '+RECOVERY/DBTARGET/%d_%t_%s_%p';
    CONFIGURE MAXSETSIZE TO UNLIMITED; # default
    CONFIGURE ENCRYPTION FOR DATABASE OFF; # default
    CONFIGURE ENCRYPTION ALGORITHM 'AES128'; # default
    CONFIGURE COMPRESSION ALGORITHM 'BASIC' AS OF RELEASE 'DEFAULT' OPTIMIZE FOR LOAD TRUE ; # default
    CONFIGURE ARCHIVELOG DELETION POLICY TO NONE; # default
    CONFIGURE SNAPSHOT CONTROLFILE NAME TO '/cots/oracle/11.2.0/dbhome_1/dbs/snapcf_dbtarget1.f'; # default

先谢谢

Answer 1

你能告诉我92GB的快递吗？

是备份大小吗？
这是真正的数据大小吗？
是否是所有数据文件的大小？

我怀疑它是备份的大小或实际数据大小。

我怀疑你的桌子很碎片。这意味着，从磁盘读取许多空块。输出看起来很少，但实际上还有很多工作要做（适用于压缩备份）。

除了你的备份脚本疯了！我建议首先简化它，然后尝试提高性能。

尝试以下脚本。它或多或少与你的相同：

RMAN>
#Do all the CONFIGURE... stuff here if you do not want to rely on the current controlfile config.
# Configure the default channel to disk and configure a format so the output goes to the filesystem not to your FRA (ASM).
CONFIGURE CHANNEL DEVICE TYPE DISK FORMAT  '/backup/DBTARGET/%d_%t_%s_%p';
CONFIGURE DEVICE TYPE DISK PARALLELISM 1 BACKUP TYPE TO COMPRESSED BACKUPSET;

# Place your controlfile autobackup to disk not asm. In case of a recovery szenario you will be glad to have it in your filesystem.
CONFIGURE CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO '/backup/DBTARGET/%F';

BACKUP DATABASE;
CROSSCHECK ARCHIVELOG ALL;
BACKUP ARCHIVELOG ALL DELETE ALL INPUT;

CROSSCHECK BACKUP;
DELETE NOPROMPT EXPIRED BACKUP;
DELETE NOPROMPT EXPIRED ARCHIVELOG ALL;
DELETE NOPROMPT OBSOLETE;
run {
  allocate channel d1 device type disk;
  recover copy of database with tag 'dbname_LEVEL_0';
  backup incremental level 1 for recover of copy with tag 'dbname_LEVEL_0' database;
}

使用RMAN的Oracle备份需要很长时间

1 个答案: