Recovery in archivelog mode

hi experts around the world.
i have oracle 9i release 2 on server 2003. we make partitions on windows
like
C (windows os)
D (oracle)
E (oracle physical datafiles)
F (oracle physical datafiles)
G (oracle physical datafiles)
H (ARCHIVE logs)
we have similar configuration standby server. we down this standby database every sunday and make cold backup of all the datafiles.
one more thing that we take controlfile from production database which is running
with the command
ALTER DATABASE BACKUP CONTROLFILE TO 'C:\control001.ctl'
like this.
now i wantt to ask that i have cold back up of all the drives and all the archive logs after that sundays backup.
i restore this cold backup on the new server .
tell me
1. thing that from D drive (here oracle software installed) which files and folders (bcz too many folders) i will restore on new server. i search this thing but not satisfactore results
2.will i apply the archives with RECOVER DATABASE command or some other command
3.which controlfile will be restored which we took from production database or some other one.
bcz many critical problems with controlfile during recovery.
plz reply these 3 questions seperately in detail.
regards
rehan
fsd pakistan
Edited by: user10204771 on Apr 3, 2009 4:26 AM

as i told we have production and standby configuration.
datablock got corrupted on primary and similarly this archivelog went on standby and applied there. i think this is the
problem.
No you didnt told me exactly the problem you are facing,you asked for moving OS cold backup from one node to another,i replied exactly.Now you are tellling the problem with yours block corruption then this method for coping from one node to another node will no longer be helpful,its block corruption which requires BMR (Block media recovery).
You can achiev the same by restoring FULL backup which you might have taken before the block corruptiuon.If you dont have FULL backup then you can recover block using RMAN by identifing corrupted block.
You can recover that corrupted block provided you should have redo for that block.
RMAN>blockrecover datafile <datafilenumber> block <blocknumber>;This is what the beauty of RMAN which backup and recover at block level.
now the problem is both primary and standby db not working.
What do you mean by not working ,in what sense you say db does not work ,yours database should be available ,you have corrupted block not system datafile.
i want to restore this cold backup on new server with the same relevant drives I WILL in C drive it will be new OS and recover by applying archive logs.
This will not work the problem will also be shifted.
how to proceed further
Check yours alert logfile for datablock which you got corrupted ans paste it here.
Khurram

Similar Messages

Help required - Recovery of a datafile that is lost in ARCHIVELOG mode

Hi:
This is related to Oracle 10G on windows 2003 Enterprise, and it is running on ARCHIVELOG mode.
The datafile 'C:\oracle\product\10.2.0\orcl\example01.dbf is lost. Most probably when I took a manual backup i missed it.
I am not using rman rather manual backup and restore.
I mounted the database (obviously it cannot be open since it is not finding the datafile) and took the datafile offline.
I cannot take a tablespace "example" offline in mount stage.
I tried "ALTER DATABASE CREATE DATAFILE <old path / old name> AS <old path / old name>.
Oracle is not allowing saying that the file is read-only.
I altere the database open and took the tablespace offline immediate and then tried the "ALTER DATABASE CREATE DATAFILE <old path / old name> AS <old path / old name>" command. But no luck.
_Please provide some steps so that I can bring it back._ I am looking many sites but everywhere is RMAN.
Thanks in advance - Caesar

>
I am looking many sites but everywhere is RMAN.
>
Yes and for a good reason. For example, RMAN would not have forgotten to backup this datafile :-)
Regardless, you messed it up. Look at DBA_TABLES for the owner & tables that resided in that example tablespace. Probably, it was just our demo schemas. You can recreate them without having to recover then. Look at
http://download.oracle.com/docs/cd/E11882_01/server.112/e10831/toc.htm
for a guide how to do that.
Also, rethink your refuse of RMAN, I would suggest
Kind regards &
Don't forget to backup the datafile :-)
Uwe Hesse
http://uhesse.wordpress.com

Recovery in NOARCHIVELOG mode. A Sybex Study GUide quibble

Hi.
Actually, this is much more than a quibble as its a crucial point vis a vis the exam.
To perform a recovery in noarchivelog mode, we have to restore the db from the last backup.
I was under the impression that a backup of the database should NOT include the online redo logs.
But, to quote this Sybex book (p291 for Exam 032):
"To perform the database restore, you will need to copy all the data files, online redo logs, and controlfiles from the last backup and perform a full database restore" .
I understand that if we do a closed backup (as we would do if the db is in noarchivelog mode) then the redo logs are not available to be backed up and therefore are not included.
I just need this detail ironed out once and for all!
Thanks.
DA

When you restore the redo log file, you are basically defining the most recent time to which the restore can occur.
In a noarchivelog restore (of a proper clean backup), the most recent possible time is the time when the database was shut down and thus there are no redo entries to roll forward. Therefore, functionally, it really does not matter whether the redo log is restored or not. However, if the redo log file is not restored, you have two situations that must be managed: redo is much newer than the rest of the database, or redo is totally missing. Since you are in learning mode, I'll leave it to you to describe how to handle those two cases.
In an archivelog environment, restoring the redo log will severely limit the ability to restore to point-in-time as you can not move forward from the redo log's time. Therefore the recommendation is to not bother backing up the redo log file at all, so it can not be accidently restore. RMAN enforces this.
Personal 'best practice' is to make a safety copy of the redo log files before any restore. That way if an old set of redo is accidently copied or if the restore/recovery needs to be restarted, the environment can be reset to the start of the restore.

Problem in Converting Database into Archivelog mode (Oracle 10G)

Hi Team,
I come across a strange problem in the ORACLE 10G Server.
I am converting database mode from NoArchivelog to Archivelog mode through RMAN in 10G.
Now When I execute these following commands through RMAN prompt it works properly as shown below?
C:\>rman
Recovery Manager: Release 10.2.0.1.0 - Production on Thu Nov 30 18:01:08 2006
Copyright (c) 1982, 2005, Oracle. All rights reserved.
RMAN> connect target /
connected to target database: RAVI (DBID=4025722893, not open)
RMAN> shutdown immediate;
using target database control file instead of recovery catalog
database dismounted
Oracle instance shut down
RMAN> STARTUP MOUNT;
connected to target database (not started)
Oracle instance started
database mounted
Total System Global Area 167772160 bytes
Fixed Size 1247876 bytes
Variable Size 79693180 bytes
Database Buffers 79691776 bytes
Redo Buffers 7139328 bytes
RMAN> SQL 'ALTER DATABASE ARCHIVELOG';
sql statement: ALTER DATABASE ARCHIVELOG
RMAN> ALTER DATABASE OPEN;
database opened
RMAN>
But this same script when i writes in the backup.ora file & pass to Rman prompt it fails,
File backup.ora contains...
run
SHUTDOWN IMMEDIATE;
STARTUP MOUNT;
SQL 'ALTER DATABASE ARCHIVELOG';
ALTER DATABASE OPEN;
passed to RMAN as follows...
C:\OracleCode\BACKUP>"C:\oracle\product\10.2.0\db_1\bin\RMAN.EXE" target /"connect target SYSTEM/sreedhar@RAVI" log="C:\ORACLE~2\LOGS\backup_log.log" append cmdfile="C:\ORACLE~2\BACKUP\backup.ora"
RMAN> 2> 3> 4> 5> 6> 7> 8>
then it fails giving the following errors...
using target database control file instead of recovery catalog
database closed
database dismounted
Oracle instance shut down
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of startup command at 11/30/2006 18:05:59
ORA-12514: TNS:listener does not currently know of service requested in connect descriptor
Recovery Manager complete.
The same thing is working in the Oracle 9i but not in the Oracle 10G.
Can Anybody plz help me in this?
Regards,
S.Tiwari
.

export ORACLE_SID=<SID>
rman target /cmdfile="C:\ORACLE~2\BACKUP\backup.ora"
it will connect to the default SID, there's no such thing as a default SID, what do you mean?
But what if there are more that one SID available & I
want to connect to SID other than the default SID.just specify the desired SID prior starting rman.
more over the same string is working with 9i but not
with 10G.maybe due to a bug?
to summarize, you have two options it you would like to start up the instance:
either you specify the SID prior starting rman and use os authentication
or
you register the instance statically and use oracle authentication.
regards,
-ap

Startup in Archivelog mode

Hi.
I only will do a simply question:
If I have a database in archivelog mode, and ocurr a unexpected shutdown
is There some process that use these logs archived if it´s necesary automatically when I do a normal startup, or only will use the redo logs?
Thanks

Alex Tutor wrote:
Hi.
I only will do a simply question:
If I have a database in archivelog mode, and ocurr a unexpected shutdown
is There some process that use these logs archived if it´s necesary automatically when I do a normal startup, or only will use the redo logs?
ThanksNo ,in this case oracle(SMON) will use only online redo log and will perform instance recovery then database will open.Of course if there will not happen media failure.

Recover datafile in no archivelog mode

Hi, i have always done media recovery using RMAN and this works nicely.
However im doing some testing on a database that does not use RMAN and is not in archivelog mode and im confused.(11.2 - aix)
My testing is as follows;
With the database shutdown cleanly i have taken a copy of dbf file. I then open the database and delete the dbf file.
(to slightly digress how can i still view and update the table from this datafile when the dbf file is dropped?)
In scenario 1 i shutdown the database after dropping the file. I then put the dbf file back and try and startup the database.
I thought that as i have a complete consistent backup i would be able to open the database normally without recovery, but this produced an error saying the file needed media recovery so i recovered the datafile and put online and the the db opened normally.
In scenario 2 i again took a clean backup and removed the dbf file. I then opened my database and switched redo logs 3 times ( i have 3 redo logs). After the 3rd switch the database crashed (which makes sense).
I have tried opening it normally again butagain it says the file needs media recovery. i then tried to recover the file but now i the following error:
ORA-00279: change 6667823 generated at 12/17/2013 11:35:20 needed for thread 1
ORA-00289: suggestion :
/test/oracle/TSTA/TSTA/archivelog/2013_12_17/o1_mf_1_78_%u_.arc
ORA-00280: change 6667823 for thread 1 is in sequence #78
Specify log: {<RET>=suggested | filename | AUTO | CANCEL}
Any help would be greatly appreciated

885842 wrote:
Hi, i have always done media recovery using RMAN and this works nicely.
However im doing some testing on a database that does not use RMAN and is not in archivelog mode and im confused.(11.2 - aix)
My testing is as follows;
With the database shutdown cleanly i have taken a copy of dbf file. I then open the database and delete the dbf file.
(to slightly digress how can i still view and update the table from this datafile when the dbf file is dropped?)
In all *nix systems, the file is still there and open by any processes that had it open when you issued the 'rm' command. It is not really deleted until all processes that have it open have terminated.
In scenario 1 i shutdown the database after dropping the file. I then put the dbf file back and try and startup the database.
I thought that as i have a complete consistent backup i would be able to open the database normally without recovery, but this produced an error saying the file needed media recovery so i recovered the datafile and put online and the the db opened normally.
The backup may have been consistent, but if you restored only that one file, it was several SCN's behind the other files that continued to be updated after you took the consistent backup. "Consistent" means not only consistent interally to the individual file, but consistent with all other data files, plus the control files, plus the redologs.
In scenario 2 i again took a clean backup and removed the dbf file. I then opened my database and switched redo logs 3 times ( i have 3 redo logs). After the 3rd switch the database crashed (which makes sense).
I have tried opening it normally again butagain it says the file needs media recovery. i then tried to recover the file but now i the following error:
ORA-00279: change 6667823 generated at 12/17/2013 11:35:20 needed for thread 1
ORA-00289: suggestion :
/test/oracle/TSTA/TSTA/archivelog/2013_12_17/o1_mf_1_78_%u_.arc
ORA-00280: change 6667823 for thread 1 is in sequence #78
Specify log: {<RET>=suggested | filename | AUTO | CANCEL}
Any help would be greatly appreciated
You said you are not in archivelog mode. So where did this reference to an archivelog come from?

Backup in ARCHIVELOG Mode

Dear Sirs,
Actually I've a non archivelog mode RDBMS, but I wan't to make a
hot backup. Could I make a hot backup after I turn the
ARCHIVELOG mode ON, then turn it off again after Back - Up ?
And how do I make the hot backup ?
Thank you for the attention and help.
Regards,
Tonny

You could turn archivelog mode on, do the backup, and then turn
archivelog mode off. This really defeats the purpose of hot
backups though. You would have to know the SCN, archivelog
file, or time to do a point-in-time recovery manually, or if you
are using RMAN knowing the time will help. I think RMAN will
recover until the last archivelog if you don't specify a time or
SCN. Keep in mind that you will lose all the changes since the
last backup if you turn archivelog mode off after the backup and
your database crashes inbetween the next scheduled backup. If
you are going through the trouble for turning archivelog mode on
and then off I would perform a cold backup. You have to
shutdown the database with normal or immediate, mount the
database, turn archvelog mode on/off and then open the
database. So why not just do a cold backup?

Backup and Recovery in different mode

We have a database in noarchivelog mode.
Currently we convert it to archivelog mode before monthly full database RMAN backup.
And after backup,we again back it to noarchivelog mode.
Now my question is that, in case of failure can we recover this database(any how) with that backup?
Please advice and if possible give the recovery steps.
Thanks & Regards
Quazi Abdur Rab

Hi,
It's not good to switch from Archive Log Mode and No Archive Log Mode as your wish.
See, As Khurram had explained you the thing. Point in time recovery before crash is quite not possible but only last the restore point applicable for your DB. Rest of the things can be recovered. Why are the implications you are facing if the DB is in Archive Mode. ?? I Suggest do peform and maintain Archive Log which is good for your DB.
- Pavan Kumar N

Force logging in archivelog mode

Hi !!!
What happen if I have "force_logging" parameter set to true when the database is in archivelog mode ?
Thanks.

rarain wrote:
Hi Juamd,
You should only use this option when it is really required because this option will forcibly generate redo for all Nologging operations that means you might find more archives and you need to setup more space for archive.
Normally we use this option when we need to replicate data changes from one database to another database like in standby configuration, Golden Gate replication etc. I would suggest you to monitor Redo amount generated after enabling this option and accordingly estimate archive space and backup space for archive.
Thanks...Ah, don't agree with that at all. You can compromise your recovery if you happen to want to restore to a point-in-time when there was a NOLOGGING operation going on. Fine, if it's an index, but if it happens to be on a table...
(Yes, been there, done that - with a non-Production database, thankfully)
This is one of the 'must haves', IMO, for Production - set it at the database-level and it overrides any tablespace or object setting.
Archivelogs are generated for a reason. If you have a particular operation that really does massively benefit from NOLOGGING and is something you are sure that you simply re-run/re-create yourself, fine. If not, by default, you really should FORCE LOGGING.

Switched Database to ARCHIVELOG Mode, Then it Hung

Hi everyone,
This db is 9.2 on Windows.
**After switching my database to Archivelog mode and reopening, it hung forever.**
(I had to use Shutdown abort to bring it down - these steps worked to get it back online:
shutdown abort; --startup mount;
alter system archive log all;
alter database noarchivelog;
alter database open;)
The Alert.log shows "Cannot allocate new log" immediately after I switched to Archivelog mode. So it hung at that point. Had I known why it hung, maybe running 'Alter system archive log all' would have fixed everything. But I've never had to run that when I've switched other databases to Archivelog mode.
Does this indicate another problem with this database? I'm leery of it now and feel lucky that I was able to get it up again.
Thanks, John
entries from Alert.log when I switched to Archivelog mode:
ALTER DATABASE MOUNT
Wed Aug 06 11:44:19 2008
Successful mount of redo thread 1, with mount id 1689013583
Wed Aug 06 11:44:19 2008
Database mounted in Exclusive Mode.
Completed: ALTER DATABASE MOUNT
Wed Aug 06 11:44:45 2008
alter database archivelog
Wed Aug 06 11:44:45 2008
Starting control autobackup
Control autobackup written to DISK device
handle 'F:\ORACLE\ORA92\DATABASE\C-1602026732-20080806-00'
Completed: alter database archivelog
Wed Aug 06 11:45:17 2008
alter database open
Wed Aug 06 11:45:17 2008
Thread 1 opened at log sequence 9014
Current log# 1 seq# 9014 mem# 0: F:\ORACLE\ORADATA\PBSMEM\REDO01.LOG
Successful open of redo thread 1
Wed Aug 06 11:45:17 2008
SMON: enabling cache recovery
Wed Aug 06 11:45:17 2008
Undo Segment 1 Onlined
Undo Segment 2 Onlined
Undo Segment 3 Onlined
Undo Segment 4 Onlined
Undo Segment 5 Onlined
Undo Segment 6 Onlined
Undo Segment 7 Onlined
Undo Segment 8 Onlined
Undo Segment 9 Onlined
Undo Segment 10 Onlined
Successfully onlined Undo Tablespace 1.
Wed Aug 06 11:45:17 2008
SMON: enabling tx recovery
Wed Aug 06 11:45:17 2008
Database Characterset is WE8MSWIN1252
replication_dependency_tracking turned off (no async multimaster replication found)
Completed: alter database open
Wed Aug 06 11:59:58 2008
Thread 1 advanced to log sequence 9015
Current log# 2 seq# 9015 mem# 0: F:\ORACLE\ORADATA\PBSMEM\REDO02.LOG
Wed Aug 06 12:01:03 2008
Thread 1 advanced to log sequence 9016
Current log# 3 seq# 9016 mem# 0: F:\ORACLE\ORADATA\PBSMEM\REDO03.LOG
Wed Aug 06 12:09:16 2008
Thread 1 cannot allocate new log, sequence 9017
All online logs needed archiving
Current log# 3 seq# 9016 mem# 0: F:\ORACLE\ORADATA\PBSMEM\REDO03.LOG
**** PrefBid App locked up at this point – hung – users cannot do anything -- Note the gap in log entry time – no further entries until my Shutdown request

Hi,
We can enable automatic archiving after instance startup.
In order to enable automatic archiving without shutting down the current instance, use the SQL command ALTER SYSTEM with the ARCHIVE LOG START parameter.
ALTER SYSTEM ARCHIVE LOG START;
When using this option, the instance does not have to be shut down to enable automatic archiving.
However, if an instance is shutdown and restarted after automatic archiving is enabled using this option, the instance is reinitialized using the settings of the parameter file ("LOG_ARCHIVE_START"), which may or may not enable automatic archiving.
NOTE: On 9i, you can use an SPFILE instead of the standard pfile. The LOG_ARCHIVE_START parmaeter is deprecated with 10g.
alter system set log_archive_start=true scope=spfile;
Did you read the Metalink Note I gave you? These information was there...
Thanks,
Marcello M.

Backups check in no archivelog mode -Question

I am trying to figure either DB is backedup or not, I checked its logfiles, it did'n shutdown from last one month, and running in no-archivelog mode.....I can assume its not being backed up from last month , make me right if i am not...........
I have been assigned to apply Oct2007 CPU on this server, & want to take backups first, any suggestions?

if i have export backup, will i be able to do
recovery of db is something goes wrong while cpus
applying with import.........Export are not backups but still popularly used as backups. Yes, if you have the latest export you will be able to get back your database to that point. Make sure you take FULL database export.
I would suggest you to bring down the database and take a cold backup.

ARCHIVELOG MODE --- Please assist

dear experts,
I have this doubt in my mind and would like you guys to advise on.
Is it true that when a database is in ARCHIVELOG MODE, database recovery using archived logs & hotbackup is possible to the last commit?
probably using the command: recover database using backup controlfile until cancel; (to apply all the arhive logs to last hotbackup)
From my point of view, i feel that the database recovery using the ARCHIVED LOGS and hotbackup is possible only up to the last redo log switch. In order to recover the database to the last commit, I have to recover the database using both the ARCHIVED LOGS as well as the last ACTIVE REDO LOG.
It is because the most recent commited transactions on the current redo log might not be written to archive log (as the active log is not full and no alter system switch logfile performed) at the point of time when the db crashed.
Please advise.
thanks all!

Also, by using RMAN you can approach more to the last commit with RMAN parameters like SCN, etc.
It's quite easy!

Setting Archivelog mode

In order to use RMAN y need to alter my database into archivelog mode. I found some info abuot how to do that, however, I am not sure about the procedure.
These are the commands I have to write for make the modification:
shutdown immediate;
starup mount;
alter database archivelog;
alter database open;
But it looks very simple and I don´t think I'm in the correct way. There is a parameter that confuse me because the oracle page's examples don't show anythig about it:
USE_DB_RECOVERY_FILE_DEST
Can anybody helpme??

Hi,
You need to set the archive destination in the pfile/spfile before placing the database to archivelog mode.
In your database pfile/spfile:
db_recovery_file_dest_size=2G #space size can be specified according to your requirement
db_recovery_file_dest='<destination where you want to store the archives>'Once done, mount the database using the modified spfile/pfile and place the database in archive mode. Finally open the database.
If you do not want to use FRA (flash recovery area) and just want to place the database in archive mode, then in your pfile/spfile, add:
log_archive_dest='<location where you want to store the archives>'Once done, mount the database using the modified spfile/pfile and place the database in archive mode. Finally open the database.
If you feel that your questions have been answered, then please consider closing the threads by marking it as answered and providing appropriate points.

Enable archivelog mode while datbase is open

Hi everybody
Kindly I need to ask if possible to enable archivelog mode while database in open status for oracle 11g? and if yes .. what is the statment?
Thanks in advance
Edited by: user11254543 on Jun 8, 2011 11:51 PM

You cant enable the archivelog mode when the database is in nomount state because the archive log information is kept in Control file like db in archivelog mode or not,log sequence number,SCNs,logfiles members,etc.. When we start the database in nomount state then only the pfile/spfile is read and sga is allocated and background processes are started but no controlfile is read ..to read the controlfile the database must be started in mount state..Why we require a controlfile is because it stores critical information about the database...If we start the database in mount state then it recognizes whether database is in archivelog mode or not , If it is in archivelog mode then it sees whether an instance recovery is needed or not by verifying the log sequence numbers in controlfile,redologfile & the datafile ..If an instance recovery is required then it performs the instance recovery by calling smon background process or if a datafile header is not up to date then we have to perform the media recovery using the redo logfiles...
For open state you have to post the reply..it will help you to explore yourself...Good Luck!

Set Archivelog Mode

How can I turn on/off the Archivelog Mode for a database by using the OEM, (say, I launched the OEM at http://My_Machine:1158/oem, after I log on where I go next)?
Thanks to help.
Scott

Thanks for the response.
Now I am on the Recovery Settings page:
1) Under Media Recovery, I checked ARCHIVELOG Mode*
2) For the Log Archive Filename Format*, it carries the entry of: ARC%S_%R.%T, what should I replace %S, %R, abd %T with? --Could you give an example?
3)Under Number Archive Log Destination Quota (512B) Status Type, there are 10 lines for entries, so I filled with the following:
1 C:\Oracle\product\archivelog1_dir
2 C:\Oracle\product\archivelog2_dir
3 C:\Oracle\product\archivelog3_dir
9
10 USE_DB_RECOVERY_FILE_DEST n/a VALID Local
Quesion:
3.1) Is the (format of ) the above entry :"C:\Oracle\product\archivelogx_dir" --Correct? Note: archivelogx_dir is the sub-directory name, not the file name.
3.2) It looks like the Quota for each destination is (512B) (byte?). If so, generally, how many such (512 Byte) do I need? I mean on average. And why does it restrict to 512 Byte for each destination?
3.3) Assume I'll use up ALL the 10 (lines) of destinations above, if I understand it correctly, the total and maximum of the archivelog file is 10 x 512 Bytes -- Is this correct?
3.4) What should I do about Line #10 above, which carries the entries of:
USE_DB_RECOVERY_FILE_DEST under Archive Log Destination
n/a under Quota
VALID under Status
Local under Type
--I mean, should I replaced with a C:\Oracle\product\archivelogx_dir? or delete it? or else? What is this Line's default entry for?
4) If arvhice log file ARC%S_%R.%T is specified, will this same file be spread into multiple destination (directories)?
Thanks to help.
Scott

Recovery in archivelog mode

Similar Messages

Maybe you are looking for