Tablespace begin backup mode,

Whether redo-logfiles get created during tablespace begin backup mode.? is that really possible?

a9362c69-5635-4bf2-a548-4a13a7cba11b wrote:
Whether redo-logfiles get created during tablespace begin backup mode.? is that really possible?
Database Concepts
Continuing to flood the forum with documentation questions -- especially after repeated warnings -- will continue to result in nothing more than links to the documentation you should be reading.

Similar Messages

Update & select query when tablespace in backup mode

Hi,
I have a query related to the DML statement when tablespace is in backup mode. Suppose when put the tablespace in begin backup mode then during the time one user run an update statement against this tablespace and commit the transaction and other user run the select query on the same data then which image the user will get and from where the data will be taken to display.
steps:
alter tablespace users begin backup;
update emp set sal=4000 where empno=101;
commit;
select * from emp where empno=101;
Now what result user will get and from where the data will fetch.
Thanks,
Gulshan

Hi,
alter tablespace users begin backup;
update emp set sal=4000 where empno=101;
commit;
select * from emp where empno=101;AFAIK, the database would work normally and you will not find the difference from transactions perspective. If you check from Oracle architecture and process, fuzzy bit will be setted at up and additional redo vectors would be saved up in the online redo logs - due to which size of redo would higher when compared to normal transactions. Secondly, you must be worried about the backup - which might not be consistent. So better opt for RMAN instead of User managed Backups.
- Pavan Kumar N

Demerits of alter tablespace begin backup on 10.2.0.4

Hi,
What are the demerits of putting a tablespace in hot backup mode. I know that it generates additional data in redo logs. But, does DBWR and LGWR performance is also impacted? Does database writer throughput reduce leading to free buffer waits? Are these wait events common - free buffer waits and log file sync waits when tablespace is in hot backup mode? Please clarify.
Thanks,
Madhav

Those wait events would be symptoms of the increased I/O and higher I/O Wait/Response times.
There is obviously more Redo, so LGWR has more to write.
DBWR performance is probably being constrained by the I/O.
Note that it isn't just the ALTER TABLESPACE BEGIN BACKUP that is the cause. It is also the backup itself -- whether you are using cp or cpio or tar or any other method to copy datafiles to disk or to tape or to a remote server / tape library. The backup is reading from the same datafiles that DBWR is writing to and, if it is writing the backup to local disks, it is also introducing additional write I/O.

RMAN and Begin backup mode

Hi guys I have a couple of questiopns...
I want to know whether data is written to datafiles by DBWR when the datafile is in begin backup mode.
What is the difference in archive log creation during normal operation of a database and operation of a database during RMAN backup?

user13087302 wrote:
Normally, Oracle should write only changes into redolog files. But if tablespace is in backup mode, then entire block will be put into redolog as soon as this change it is 1st change occured on that block after enabling backup mode. That is why, we face massive redo log generation if the tablespace is in backup mode. So, logically you should be able to recover to the last SCN you have in hands.What do you mean by last SCN.Is it the SCN before puttting the tablespace in backup mode?Does that mean data present in other files will also be lost even though they are written in datafiles..Pls, look:
SQL> create tablespace myts1 datafile 'C:\app\oracle\oradata\testdb\myts1.dbf' size 10m;
Tablespace created.
SQL> create tablespace myts2 datafile 'C:\app\oracle\oradata\testdb\myts2.dbf' size 10m;
Tablespace created.
SQL> create table myt1 tablespace myts1 as select * from dba_objects where rownum<10001;
Table created.
SQL> create table myt2 tablespace myts2 as select * from dba_objects where rownum<10001;
Table created.
SQL> alter tablespace myts1 begin backup;
Tablespace altered.
SQL> create table myt1_1 tablespace myts1 as select * from dba_objects where rownum<10001;
Table created.
SQL> create table myt2_1 tablespace myts2 as select * from dba_objects where rownum<10001;
Table created.
SQL> create table myt1_3 tablespace myts1 as select * from dba_objects where rownum<10001;
Table created.Backup database:
C:\Documents and Settings\Administrator>rman target /
Recovery Manager: Release 11.2.0.1.0 - Production on Wed Nov 3 15:46:55 2010
Copyright (c) 1982, 2009, Oracle and/or its affiliates. All rights reserved.
connected to target database: TESTDB (DBID=2517263760)
RMAN> backup database;
Starting backup at 03-NOV-10
using target database control file instead of recovery catalog
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=44 device type=DISK
channel ORA_DISK_1: starting full datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
input datafile file number=00001 name=C:\APP\ORACLE\ORADATA\TESTDB\SYSTEM01.DBF
input datafile file number=00002 name=C:\APP\ORACLE\ORADATA\TESTDB\SYSAUX01.DBF
input datafile file number=00003 name=C:\APP\ORACLE\ORADATA\TESTDB\UNDOTBS01.DBF
input datafile file number=00005 name=C:\APP\ORACLE\ORADATA\TESTDB\EXAMPLE01.DBF
input datafile file number=00006 name=C:\APP\ORACLE\ORADATA\TESTDB\MYTS1.DBF
input datafile file number=00007 name=C:\APP\ORACLE\ORADATA\TESTDB\MYTS2.DBF
input datafile file number=00004 name=C:\APP\ORACLE\ORADATA\TESTDB\USERS01.DBF
channel ORA_DISK_1: starting piece 1 at 03-NOV-10
channel ORA_DISK_1: finished piece 1 at 03-NOV-10
piece handle=C:\APP\ORACLE\FLASH_RECOVERY_AREA\PRIM_DB\BACKUPSET\2010_11_03\O1_MF_NNNDF_TAG20101103T154701_6F2LXQ5C_.BKP tag=TAG20101103T154701 comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:01:35
channel ORA_DISK_1: starting full datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
including current control file in backup set
including current SPFILE in backup set
channel ORA_DISK_1: starting piece 1 at 03-NOV-10
channel ORA_DISK_1: finished piece 1 at 03-NOV-10
piece handle=C:\APP\ORACLE\FLASH_RECOVERY_AREA\PRIM_DB\BACKUPSET\2010_11_03\O1_MF_NCSNF_TAG20101103T154701_6F2M0PJ3_.BKP tag=TAG20101103T154701 comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:01
Finished backup at 03-NOV-10
Recovery Manager complete.
SQL> select current_scn from v$database;
CURRENT_SCN
    1466019
SQL> drop table myt1_3 purge;
Table dropped.So, we dropped table MYT1_3 and now trying incomplete recovery.
SQL> shutdown abort;
ORACLE instance shut down.
SQL> startup mount;
ORACLE instance started.
Total System Global Area 431038464 bytes
Fixed Size                  1375088 bytes
Variable Size             322962576 bytes
Database Buffers          100663296 bytes
Redo Buffers                6037504 bytes
Database mounted. C:\Documents and Settings\Administrator>rman target /
Recovery Manager: Release 11.2.0.1.0 - Production on Wed Nov 3 15:51:51 2010
Copyright (c) 1982, 2009, Oracle and/or its affiliates. All rights reserved.
connected to target database: TESTDB (DBID=2517263760, not open)
RMAN> run
2> {
3> set until scn 1466019;
4> restore database;
5> recover database;
6> sql 'alter database open resetlogs';
7> }
executing command: SET until clause
Starting restore at 03-NOV-10
using target database control file instead of recovery catalog
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=20 device type=DISK
channel ORA_DISK_1: starting datafile backup set restore
channel ORA_DISK_1: specifying datafile(s) to restore from backup set
channel ORA_DISK_1: restoring datafile 00001 to C:\APP\ORACLE\ORADATA\TESTDB\SYS
TEM01.DBF
channel ORA_DISK_1: restoring datafile 00002 to C:\APP\ORACLE\ORADATA\TESTDB\SYSAUX01.DBF
channel ORA_DISK_1: restoring datafile 00003 to C:\APP\ORACLE\ORADATA\TESTDB\UNDOTBS01.DBF
channel ORA_DISK_1: restoring datafile 00004 to C:\APP\ORACLE\ORADATA\TESTDB\USERS01.DBF
channel ORA_DISK_1: restoring datafile 00005 to C:\APP\ORACLE\ORADATA\TESTDB\EXAMPLE01.DBF
channel ORA_DISK_1: restoring datafile 00006 to C:\APP\ORACLE\ORADATA\TESTDB\MYTS1.DBF
channel ORA_DISK_1: restoring datafile 00007 to C:\APP\ORACLE\ORADATA\TESTDB\MYTS2.DBF
channel ORA_DISK_1: reading from backup piece C:\APP\ORACLE\FLASH_RECOVERY_AREA\
PRIM_DB\BACKUPSET\2010_11_03\O1_MF_NNNDF_TAG20101103T154701_6F2LXQ5C_.BKP
channel ORA_DISK_1: piece handle=C:\APP\ORACLE\FLASH_RECOVERY_AREA\PRIM_DB\BACKUPSET\2010_11_03\O1_MF_NNNDF_TAG20101103T154701_6F2LXQ5C_.BKP tag=TAG20101103T154701
channel ORA_DISK_1: restored backup piece 1
channel ORA_DISK_1: restore complete, elapsed time: 00:01:45
Finished restore at 03-NOV-10
Starting recover at 03-NOV-10
using channel ORA_DISK_1
starting media recovery
media recovery complete, elapsed time: 00:00:03
Finished recover at 03-NOV-10
sql statement: alter database open resetlogs
RMAN>Now, check tables:
SQL> select count(*) from myt1;
COUNT(*)
     10000
SQL> select count(*) from myt2;
   COUNT(*)
     10000
SQL> select count(*) from myt1_3;
COUNT(*)
     10000
SQL> select count(*) from myt2_1;
COUNT(*)
     10000So, it does not matter what is the status of any tablespace: in backup mode or not. Incomplete recovery uses archivelogs + redologs.

Rumor : the begin backup mode will disappear

I have heard that the "begin backup" mode for tablespaces will disappear in futures versions.
I can't believe it.
Anybody can say something on this rumor ?

I have heard that the "begin backup" mode for tablespaces will disappear in futures versions.
I can't believe it.
Anybody can say something on this rumor ?

Do I need to use standby database in begin backup mode to take the backup?

Hi All,
I have a 10g(10.2.0.4) primary database with standby on a solaris platform. I need to create another standby database for the same primary.
I have stopped the recovery on existing standby database and started copying(os copy) the files from existing standby to 2nd standby server. However I haven't kept the existing standby database(mounted) in begin backup mode. Is it okay?
Thanks,
Arun

Backup mode for a mounted database?
orclz>
orclz> startup mount
ORACLE instance started.
Total System Global Area 788529152 bytes
Fixed Size                  3050600 bytes
Variable Size            373293976 bytes
Database Buffers          406847488 bytes
Redo Buffers                5337088 bytes
Database mounted.
orclz>
orclz> alter database begin backup;
alter database begin backup
ERROR at line 1:
ORA-01109: database not open
orclz>

Db corrupted in begin backup mode

hi all,
This is part of my practice only.
My db is 11.2.0.1.0 on linux. i kept my db in begin backup mode and started copy and forgot to end backup. taken controlfile backup like alter database backup controlfile to '/u01/app/oracle/orcl/control_bkp.ctl'.
Now i deleted all db files, redo files, control files corrupted. (to perform complete db recovery).
I started recovery. started in nomount and copied controlfiles from backup location to it's original location.
and then issued >recover database using backup controlfile until cancel;
given follwoing error.
ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get error below
ORA-01195: online backup of file 1 need more recovery to be consistent
ORA-01110: data file 1: '/u04/oracle/oradata/jt10g/system01.dbf'
any suggestions please.
thanks for you help..

i can see from the link: https://forums.oracle.com/thread/634538 (MaranViswarayar replied) this type can't recovered. is it true.

Begin backup mode

hi,
while the tablespace in begin backup, and oracle want to access the blocks, first time its writes the entire block into redo, any one tell me, why its write the entire block in redo

Senthil,
What you mean is: "the first time the block is modified the whole block is written to the redo log before the change vectpr for the update is written to the redo log".
This happens because Oracle assumes that the backup mechanism is external to Oracle, so the backup mechanism doesn't know what the database writers (DBWn) might be doing.
If a database writer writes an Oracle block just as the backup process is making a backup copy of that part of the file, the backup may contain a "fuzzy" or "split" block, i.e. a block which is partly a copy of the old version of the block and partly a copy of the new version of the block.
At recovery time, a block in this state would cause a problem so, by copying the current version of the block into the redo log, Oracle knows that as it performs a recovery it can copy the block from the redo log to the data file and have a completely safe copy of the block on the datafile before applying the next redo vector.
In fact the block copy can happen more frequently than just on the first update. If the block is flushed from the cache, then re-read, then updated again, it is copied into the redo log again. And it is copied every single time a change requires it to be re-read.
There is a flag column in x$bh (the buffer header structure) which has a bit set for the status block_has_been_logged (see http://www.jlcomp.demon.co.uk/buf_flag.html for other states - I think this is the bit that says the block has been written to the redo log, but since it's meaning changed by 10.2.0.3 it is possible that it is the redo_since_read bit that matters). If the block is flushed from the buffer and re-read, the buffer header record obviously has to be recreated - which means the bit is cleared - hence the potential for a new block copy to go into the redo log every time the block has to be re-read and modified.
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
http://www.jlcomp.demon.co.uk

How to tell tablespace is in backup mode?

Ive issued the command:
alter tablespace example begin backupMy problem is that I can't find a table/view that shows the tablespace is now in backup mode. Is there somewhere I can find this information?
Regards,
Warren

Hi,
U can check from v$backup about the status of the tablespace which is under begin backup mode.

Hot backup : Rman vs. ALTER TABLESPACE...BEGIN BACKUP

Dear Experts,
I'm currently using the following statements
- ALTER TABLESPACE (...) BEGIN BACKUP
- host ocopy (...)
- ALTER TABLESPACE (...)END BACKUP state
I'm going for rman now but I've got 2 questions for those who did that in the past:
1/ Will rman reduce the size of my hotbackup ?
2/ Does the hot backup run faster using rman ?
Thanks.
Best Regards,
Jerome

1/ Will rman reduce the size of my hotbackup ? rman backup only those blocks which get accessed by oracle, means it doesn't copy the empty blocks as a part of backup as your OS copy command does. So yeah, the size of the backup will be smaller if there are huge datafiles but with comparably less data. Moreover if you are on 10g then you can use compressed backupset feature which will compress the backupsets.
2/ Does the hot backup run faster using rman ?
By using rman, there is no need to put every tablespace in backup mode, you can run the backup in parallel by allocating multiple channels, on 10g you can use block change tracking feature to speed up your incremental backups AND as explained in point 1 there is no need to copy the whole datafile. These all are the benefits of RMAN and I can't make a state statement that by all this the backup will run faster but will definitely go for it.
Daljit Singh

How SCN sync after the tablespace is put back to END Backup mode

When we put the tablespace in begin backup mode, the scn get forzen and it will not be further updated, until it is put back to end backup.
Question is : If at the time of being backup , say the snc forzed at 45. Before we put back the tablespace to end back up , we had genrerated 10 SCN. Then the current scn is 55. If we put the tablespace in end backup , does the datafile will start with scn 56 ( by maintaint it previous scn at 45)
Or after we put back the tsp in end back up , does the datafile will be updated from 46 to 55 , before it is updated with scn 56.
In the above example i am considering by db has one datafile only.
Thanks
Naveen

Are you implying that the "Checkpoint number" is a different series of numbers --
seperate from SCNs ?
Are you implying that when a checkpoint occurs it is this "Checkpoint number"
that is incremented and written to the datafile headers ?
There really is no such thing as a "Checkpoint number".
If you look at the definitions of V$DATABASE , V$DATAFILE and V$DATAFILE_HEADER, you would see that a Checkpoint increments the SCN.
A header can have multiple entries -- a Checkpoint SCN (called CHECKPOINT_CHANGE#) , ArchiveLog SCN (called ARCHIVELOG_CHANGE#),
the Current SCN (called CURRENT_SCN), SCN of last Unrecoverable (Nologging)
(called UNRECOVERABLE_CHANGE#) etc etc. They are all derived from SCNs
only.
The SCN is used as the single point of reference. Thus, we can compare
CHECKPOINT_CHANGE# with CURRENT_SCN with UNRECOVERABLE_CHANGE# etc.

PIT recovery until time when database was in backup mode

Hello,
Is it possible to recover database until some point/time which is within "time window" when database tablespaces were still in online backup mode?
I was basicaly replicatying database from source to target DB by putting DB to online backup mode, then copy all datafiles, after copy of datafiles was finished I started applying redologs generated on source DB afterwards. After I applied tens of redologs with "recover database until cancel using backup controlfile" and then typed "CANCEL" after 10-20 redologs applied I still got an error that opening DB would get an error due to inconsistent datafile. When I ended backup mode on source DB and applied few of redologs after ending backup mode on source DB datafiles were consistent.....
Thx a lot for answer in advance.

Hello Hemant K Chitale ,
thats clear. I probably did not express it clearly, so I will describe it once again:
Teoreticall scenario 1:
I put 'source" database to backup mode and start copy datafiles to "target" db. When copy of all datafiles is succesfully finished (lets say it 5:00AM), I don't issue "end backup" on source DB but I keep tablespaces in backup mode until 7:00AM. On target DB I will start recovery and apply redologs from source system.
Question 1:Is it possible to recover database until time 6:00AM? At this time all datafiles were already succesfully copied, BUT source db was still in backup mode. Or will I be able to get target DB to consistent state only after 7:00AM (end backup issuied on source DB) and not in window 5:00AM-7:00AM?
Teoreticall scenario 2 (small correction comparing to scenario1):
I will put source DB to begin backup mode at 5:00AM on 20.february and let it in this state until 7:00AM on 20.february. On target DB I will restore backup from source DB performed day before (or few days before does not matter) that means on 19.february and start to apply redologs generated after this backup.
Question 2: Will I be able to recover DB until 6:00AM on 20.february and open database after recovery until this time? Or wont I be able to recover and open DB witin window 20.february 5:00AM-7:00AM (source db in backup mode)
Bear in mind this is just teoreticly (don't ask me why would you do it and why you don't do it this way etc...). I just want to understand how it really works ;)
Thx a lot!

Where do changed data values goto if DB is in BACKUP mode for LONG PERIODs

Where does oracle write if put the database in begin backup mode for LONG PERIODs. Lets say I issued a "ALTER DATABASE BEGIN BACKUP" command in a busy database and forgot about it for a long time.
I understand that when the DB IS IN BEGIN BACKUP MODE, "the database copies whole changed data blocks into the redo stream." (Page 503 of 11.1 backup and recovery guide). But the redo stream is limited by the number of online redologs. After some time redologs also wont be sufficient for the changed data values after a begin backup is issued. I understand that there are archived redologs.
Lets say there are 2 redolog groups in this database and Lets say 10 archive log files got generated since the ALTER DATABASE BEGIN BACKUP was issued.
When i finally issue the "ALTER DATABASE END BACKUP" command, will Oracle sync the datafiles with the changed data blocks reading the data from these 10 archived log files ? What happens if i delete these archive redologs from the archive log destination. ??
page 504 of 598 in the backup and recovery guide
Caution : If you fail to take the tablespace out of backup mode,
then Oracle Database continues to write copies of data blocks in
this tablespace to the online redo logs, causing performance
problems. Also, you receive an ORA-01149 error if you try to shut
down the database with the tablespaces still in backup mode.
it just says "performance problems", nothing more than that.*
Any answers ? I am sure this question would have popped to some of you senior DBA people out there.

user13076519 wrote:
Where does oracle write if put the database in begin backup mode for LONG PERIODs. Lets say I issued a "ALTER DATABASE BEGIN BACKUP" command in a busy database and forgot about it for a long time. It writes just like it always does, plus it puts some extra in the redo log the [url http://oraclenz.com/2008/07/11/logging-or-nologging-that-is-the-question-part-ii/]first time a block is changed.
>
>
I understand that when the DB IS IN BEGIN BACKUP MODE, "the database copies whole changed data blocks into the redo stream." (Page 503 of 11.1 backup and recovery guide). But the redo stream is limited by the number of online redologs. After some time redologs also wont be sufficient for the changed data values after a begin backup is issued. I understand that there are archived redologs. This appears to be a typo (incompleteness, really) in the backup and recovery guide.
The redo stream is not limited by the number of of redo logs, only the volume of data. When a log fills up, it gets archived. If all the logs get filled up before the first one is finished archiving, the db will stall until the next redo becomes available. The only limit to archiving is disk space (and bandwidth if that is an issue, which it can be in some configurations).
>
Lets say there are 2 redolog groups in this database and Lets say 10 archive log files got generated since the ALTER DATABASE BEGIN BACKUP was issued.
When i finally issue the "ALTER DATABASE END BACKUP" command, will Oracle sync the datafiles with the changed data blocks reading the data from these 10 archived log files ? What happens if i delete these archive redologs from the archive log destination. ??Archived logs are archived, Oracle only reads them in recovery. You do not understand archive logs, read the concepts manual.
>
page 504 of 598 in the backup and recovery guide
Caution : If you fail to take the tablespace out of backup mode,
then Oracle Database continues to write copies of data blocks in
this tablespace to the online redo logs, causing performance
problems. Also, you receive an ORA-01149 error if you try to shut
down the database with the tablespaces still in backup mode.
it just says "performance problems", nothing more than that.*Because it is overgeneralizing.
>
Any answers ? I am sure this question would have popped to some of you senior DBA people out there.Oh, you want to send me a gift for showing where in oracle it's documented? See [url http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:271815712711]here for something over a decade old.

Alter database begin backup; missing keyword error

Hi all,
I am trying to put my database in backup mode using the command "alter database begin backup", but it is giving error shown below:
$ sqlplus '/as sysdba'
SQL*Plus: Release 9.2.0.5.0 - Production on Wed Nov 25 16:52:58 2009
Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.
Connected to:
Oracle9i Enterprise Edition Release 9.2.0.5.0 - 64bit Production
With the Partitioning, OLAP and Oracle Data Mining options
JServer Release 9.2.0.5.0 - Production
SQL> select name,open_mode from v$database;
NAME OPEN_MODE
N05PCS07 READ WRITE
SQL> alter database begin backup;
alter database begin backup
ERROR at line 1:
ORA-00905: missing keyword
NE one can tell why the command is failing??
pls...
Thanks in advance.
SHAILESH

Hello,
All the former posts are right, alter database begin backup; doesn't exists in 9.2 althought
you already have in this release the alter database end backup;.
So as to put all the tablespaces in backup mode in one command I suggest you to create a
stored procedure as follows:
CREATE OR REPLACE PROCEDURE begin_backup IS
TYPE RefCurTyp IS REF CURSOR;
cv RefCurTyp;
sql_cv VARCHAR2(400);
tbs VARCHAR2(128);
sql_stmt VARCHAR2(200);
BEGIN
-- Mise en BEGIN BACKUP des TABLESPACES
sql_cv := 'SELECT distinct (A.name) FROM sys.v_$tablespace A, sys.v_$datafile B, sys.v_$backup C where A.TS# = B.TS# and B.FILE# = C.FILE# and C.STATUS = ''NOT ACTIVE''';
OPEN cv FOR sql_cv;
LOOP
FETCH cv INTO tbs;
EXIT WHEN cv%NOTFOUND;
sql_stmt := 'ALTER TABLESPACE '||tbs||' BEGIN BACKUP';
EXECUTE IMMEDIATE sql_stmt;
END LOOP;
CLOSE cv;
END;
Then, I just have to call the Procedure like that:
execute <schema>.begin_backup; and all the tablespaces are in backup mode.
This procedure must be created on a User/Schema with the following privileges:
alter tablespace
select on sys.v_$tablespace
select on sys.v_$datafile
select on sys.v_$backupHope it can help.
Best regards,
Jean-Valentin
Edited by: Lubiez Jean-Valentin on Nov 25, 2009 9:44 PM

What exactly happens when datafile is put in backup mode ?

Dear Friend,
i want to know what exactly happens when a datafile is put in backup mode ?
that is what happens between Begin Backup ........................... end backup
with regards.

When u start hot backup via 'alter tablespace begin backup', the tablespace headers are marked as such. The dirty data buffers that belong to the database files are written out to the files and the datafiles are check pointed. The headers are updated with SCN @ the time the backup commences. It is the data file headers that were not updated until the 'alter tablespace end backup' is issued.
The database files still continue to be read and updated. Recovery will commence from the scn written at the commencement of the backup and it is the changes written to the redo\archive logs during the backup that are rolled forward during recovery
Cheers,
Bin

Tablespace begin backup mode,

Similar Messages

Maybe you are looking for