Re: Relocating Flashback Recovery Area While A Guranteed Restore Point Is Present
- From: Don Seiler <don@xxxxxxxxx>
- To: fmhabash@xxxxxxxxx
- Date: Fri, 11 Apr 2014 10:17:52 -0500
I think you'll find, as with other FRA files (archivelogs, backups), that
the old flashback log files will continue to exist in the old FRA location
where the database will continue to look for them if needed. New files will
be written to the new location. When the old files are obsolete then
they'll be removed from the old location automatically (or you can delete
them via RMAN).
On Fri, Apr 11, 2014 at 9:34 AM, Fred Habash <fmhabash@xxxxxxxxx> wrote:
> The official oracle docs (UG & MOS) asks that flashback is toggled (and
> GRP recreated) if db_recovery_file_dest is changed.
>
> However, my tests on 11.2.0.4.0 shows that this is not needed. I saw that
> Oracle produced the flb files to the new location and alert log did not
> complain.
>
> Have you done this before? If yes, what was your experience?
>
> References
> ----------------------------------------
> 1) How to change Flash Recovery Area to a new location ? (Doc ID 305651.1)
> 2) Space Usage Figures in V$flash_recovery_area_usage Misleading When FRA
> Location is Changed (Doc ID 460145.1)
>
>
> Case study
> ----------------------------------------
>
> Setup FRA to dir dbr1
> #########################
> alter system set db_recovery_file_dest='/u01/ora_stage/fra/dbr1'
> scope=both ;
> System altered.
> Sql+( SYS@canon)> show parameter db_recovery
> NAME TYPE VALUE
> ------------------------------------ -----------
> ------------------------------
> db_recovery_file_dest string /u01/ora_stage/fra/dbr1
> db_recovery_file_dest_size big integer 1G
>
> Enable FB & GRP
> #########################
> Sql+( SYS@canon)> alter database flashback on;
> Database altered.
> Sql+( SYS@canon)> CREATE RESTORE POINT before_upgrade GUARANTEE
> FLASHBACK DATABASE;
> Restore point created.
>
> Verify flb files are generated
> #########################
> dbr1/canon/flashback:
> total 205072
> -rw-r----- 1 oracle oinstall 52436992 Apr 11 09:14 o1_mf_9nhxyr12_.flb
> -rw-r----- 1 oracle oinstall 52436992 Apr 11 09:14 o1_mf_9nhxyv7b_.flb
> -rw-r----- 1 oracle oinstall 52436992 Apr 11 09:14 o1_mf_9nhy6rnf_.flb
> -rw-r----- 1 oracle oinstall 52436992 Apr 11 09:14 o1_mf_9nhy6978_.flb
> myhost:oracle:11.2.0.4:no_sid> ls -ltrR dbr2
>
>
> Repoint FRA to dir dbr2
> #########################
> alter system set db_recovery_file_dest='/u01/ora_stage/fra/dbr2'
> scope=both ;
> System altered.
> Sql+( SYS@canon)> show parameter db_recovery
>
> NAME TYPE VALUE
> ------------------------------------ -----------
> ------------------------------
> db_recovery_file_dest string /u01/ora_stage/fra/dbr2
> db_recovery_file_dest_size big integer 1G
>
> Verify flb files are not created in new dir dbr2
> #########################
> myhost:oracle:11.2.0.4:no_sid> ls -ltrR dbr2
> ...
> dbr2/canon/flashback:
> total 256340
> -rw-r----- 1 oracle oinstall 52436992 Apr 11 09:18 o1_mf_9nhybs0f_.flb
> -rw-r----- 1 oracle oinstall 52436992 Apr 11 09:18 o1_mf_9nhyc89s_.flb
> -rw-r----- 1 oracle oinstall 52436992 Apr 11 09:19 o1_mf_9nhygdd9_.flb
> -rw-r----- 1 oracle oinstall 52436992 Apr 11 09:19 o1_mf_9nhyfzz2_.flb
> -rw-r----- 1 oracle oinstall 52436992 Apr 11 09:19 o1_mf_9nhygrpx_.flb
>
>
> Verify alert log is clean
> #########################
> Fri Apr 11 09:16:13 2014
> ALTER SYSTEM SET db_recovery_file_dest='/u01/ora_stage/fra/dbr2'
> SCOPE=BOTH;
> Fri Apr 11 09:16:50 2014
> Thread 1 cannot allocate new log, sequence 152
> Checkpoint not complete
> Current log# 1 seq# 151 mem# 0: /u01/ora_stage/oradata/canon/redo01.log
> Thread 1 advanced to log sequence 152 (LGWR switch)
> Current log# 2 seq# 152 mem# 0: /u01/ora_stage/oradata/canon/redo02.log
> Fri Apr 11 09:16:52 2014
> Archived Log entry 11 added for thread 1 sequence 151 ID 0x5acbddcc dest 1:
> Thread 1 cannot allocate new log, sequence 153
> Checkpoint not complete
> Current log# 2 seq# 152 mem# 0: /u01/ora_stage/oradata/canon/redo02.log
> Fri Apr 11 09:17:01 2014
> Thread 1 advanced to log sequence 153 (LGWR switch)
> Current log# 3 seq# 153 mem# 0: /u01/ora_stage/oradata/canon/redo03.log
> Fri Apr 11 09:17:01 2014
> Archived Log entry 12 added for thread 1 sequence 152 ID 0x5acbddcc dest 1:
> Thread 1 cannot allocate new log, sequence 154
> Checkpoint not complete
> Current log# 3 seq# 153 mem# 0: /u01/ora_stage/oradata/canon/redo03.log
> Thread 1 advanced to log sequence 154 (LGWR switch)
> Current log# 1 seq# 154 mem# 0: /u01/ora_stage/oradata/canon/redo01.log
>
> --
>
>
> Thank you ...
> ----------------------------------------
> Fred Habash, Database Solutions Architect (Oracle OCP 8i,9i,10g,11g)
>
>
--
Don Seiler
http://www.seiler.us
Other related posts:
