Database Reference
In-Depth Information
Additionally you often find a multiplexed online redo log per thread and group as well as a multiplexed
controlfile in it. The use of the FRA is so simple and the benefit of having a standard location for important backup
files so big that there is actually no reason not to use it. All you need to do is define the size of the FRA and the
location, shown here for ASM:
SQL> alter system set db_recovery_file_dest_size = 100G sid='*' scope='both';
System altered.
SQL> alter system set db_recovery_file_dest = '+RECO' sid='*' scope='both';
System altered.
For a FRA on a file system you simply point to the directory you set aside for it as shown here:
SQL> alter system set db_recovery_file_dest = '/u01/fast_recovery_area' scope='both';
System altered.
It is important that you do not include an identifier for the database. All files in the FRA are Oracle Managed
Files, and a database identifier is appended automatically for you. Complaints from the system administrator that a
common location or mount point for more than one database are a risk can be countered that every database has a
set quota on that mount point. Unlike the diagnostic destination which—if shared—can become a problem if a single
database suddenly fills the file system to 100% by core dumps or other unforeseen problems you specifically tell
Oracle how much space is available. Querying the view shows the administrator explicitly how much space is used,
and how much is reclaimable, leading to the next great feature of the FRA.
If your system experiences space pressure in the archive log destination and you are not using the FRA then there
might be a point where there is no more space left, leaving the archiver stuck which is a common yet very undesirable
situation. When using the FRA this problem can partly be mitigated by using the automatic space management
features. To understand this feature it is important to understand the type of file in the FRA. Files which are part of the
active database, such as online redo log members and a multiplexed controlfile are termed “permanent” files in the
FRA. The other files in the FRA are called transient. Only transient files are eligible for automatic removal, and only
under certain conditions. You need not worry that permanent files are deleted; they are not touched by the clean-up
mechanism. When considering which files will be deleted, the retention policy plays a very important role. The RMAN
retention policy defines the number of backups you need to keep or alternatively a time period over which backups
are required before being purged.
Note
you can learn more about the retention policy later in this chapter.
Backups that are no longer needed may automatically be removed. Information about reclaimable space is
managed in the v$flash_recovery_area_usage view, as shown here:
SYS@CDB$ROOT> select FILE_TYPE,PERCENT_SPACE_USED,PERCENT_SPACE_RECLAIMABLE
2 from v$flash_recovery_area_usage
3 where PERCENT_SPACE_RECLAIMABLE <> 0;
FILE_TYPE PERCENT_SPACE_USED PERCENT_SPACE_RECLAIMABLE
----------------------- ------------------ -------------------------
ARCHIVED LOG 4.86 4.86
BACKUP PIECE 10.37 .42
 
Search WWH ::




Custom Search