Creating Fast Incremental Backup

Creating Fast Incremental Backups : 

We have seen how block change tracking is enabled.

Enable block change tracking on the physical standby database

The motive of Incremental backup is to backup only those blocks which are changed since the last zero or level 1 incremental backup.

Block change tracking provides a boost to this. If we enable block change tracking then Oracle Database tracks the physical location of all database changes in the change tracking file. And while taking backup RMAN takes care of it. RMAN determines changed block from change tracking file so it does not need to scan the whole database for changed block since the last backup.

How is the Block Change Tracking maintained?

A background process called CTWR works for it. When redo data is being written to redo log, CTWR awakes and writes physical location of that changed block in Change Tracking file.

Note: Change Tracking file only holds the physical location of the changed block and does not contain any data.

The size of the Change Tracking File depends on following aspects :

1. Database size in bytes.
2. Number of threads in RAC environment
3. Number of old backups maintained by the change tracking file

The minimum size of change tracking file is 10 MB and when it reaches to 10 MB, new 10 MB will be allocated.

Monitoring Block Change Tracking :

We can check use block change tracking file from v$backup_datafile view with a used_change_tracking column.

Let us understand this with one example

Step 1: Take level zero backup.

RMAN> BACKUP INCREMENTAL LEVEL 0 DATABASE;

Starting backup at 10-MAY-18
using channel ORA_DISK_1
channel ORA_DISK_1: starting incremental level 0 datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
input datafile file number=00001 name=/u01/oracle/oradata/mgr/system01.dbf
input datafile file number=00002 name=/u01/oracle/oradata/mgr/sysaux01.dbf
input datafile file number=00003 name=/u01/oracle/oradata/mgr/undotbs01.dbf
input datafile file number=00004 name=/u01/oracle/oradata/mgr/users01.dbf
channel ORA_DISK_1: starting piece 1 at 10-MAY-18
channel ORA_DISK_1: finished piece 1 at 10-MAY-18
piece handle=/u01/oracle/fast_recovery_area/STD_MGR/backupset/2018_05_10/o1_mf_nnnd0_TAG20180510T224813_fh8zqorx_.bkp tag=TAG20180510T224813 comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:01:05
Finished backup at 10-MAY-18

Starting Control File and SPFILE Autobackup at 10-MAY-18
piece handle=/u01/oracle/fast_recovery_area/STD_MGR/autobackup/2018_05_10/o1_mf_s_975797323_fh8zsonf_.bkp comment=NONE
Finished Control File and SPFILE Autobackup at 10-MAY-18

RMAN>

Step 2: Do some DML operation on the database.

Step 3: Take increment level 1 backup.

RMAN> backup incremental level 1 database;

Starting backup at 10-MAY-18
using channel ORA_DISK_1
channel ORA_DISK_1: starting incremental level 1 datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
input datafile file number=00001 name=/u01/oracle/oradata/mgr/system01.dbf
input datafile file number=00002 name=/u01/oracle/oradata/mgr/sysaux01.dbf
input datafile file number=00003 name=/u01/oracle/oradata/mgr/undotbs01.dbf
input datafile file number=00004 name=/u01/oracle/oradata/mgr/users01.dbf
channel ORA_DISK_1: starting piece 1 at 10-MAY-18
channel ORA_DISK_1: finished piece 1 at 10-MAY-18
piece handle=/u01/oracle/fast_recovery_area/STD_MGR/backupset/2018_05_10/o1_mf_nnnd1_TAG20180510T231238_fh916dc0_.bkp tag=TAG20180510T231238 comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:03
Finished backup at 10-MAY-18

Starting Control File and SPFILE Autobackup at 10-MAY-18
piece handle=/u01/oracle/fast_recovery_area/STD_MGR/autobackup/2018_05_10/o1_mf_s_975798818_fh916g29_.bkp comment=NONE
Finished Control File and SPFILE Autobackup at 10-MAY-18

RMAN>

I have a small database, but we can level zero backup took 35 seconds and level 1 backup has just taken 3 seconds.

Step 4 : Check v$backup_datafile view with used_change_tracking=’YES’

SQL> select file#,avg(blocks),avg(blocks_read/datafile_blocks)*100 as pct_for_backup from v$backup_datafile where used_change_tracking='YES' group by file#;

     FILE# AVG(BLOCKS) PCT_FOR_BACKUP
---------- ----------- --------------
	 1	 27075	    32.629523
	 2	 16519	   32.1580189
	 4  58.6666667	   12.0833333
	 3  970.666667	   30.2430556

SQL> 

We can see here PCT_FOR_BACKUP is 32 % for specific files.

Stay tuned for More articles on Oracle DataGuard

Thank you for giving your valuable time to read the above information.

If you want to be updated with all our articles send us the Invitation or Follow us:

Telegram Channel: https://t.me/helporacle

Skant Gupta’s LinkedIn: www.linkedin.com/in/skantali/

Joel Perez’s LinkedIn: Joel Perez’s Profile

LinkedIn Group: Oracle Cloud DBAAS

Facebook Page: OracleHelp