Reader
A reader, April 16, 2002 - 9:24 am UTC
Tom,
Could you elaborate on this:
"
In any case -- you can always recover manually regardless
(you do not lose the ability to recover, we just might not be able to supply you
the filenames)
"
It implies as though, you can recover the database, even if you lose the recovery catalog, and the target database, as long as you know the names of the data sets that have
backed up using RMAN.
Thanks
April 16, 2002 - 12:13 pm UTC
yes.
bits are bits and bytes are bytes after all. if you can access the backups....
Reader
A reader, April 16, 2002 - 12:46 pm UTC
Tom,
I am still a bit unclear
The backup sets are in RMAN special format. Therefore we
have to use RMAN command to either *restore* and recover
OR *recover*. The command line syntax does not specify
the location of backup set or name of backup set. How
to supply the information to RMAN command
Thanks
Reader
A reader, April 17, 2002 - 10:06 am UTC
Tom,
This is an absolute extremely useful Recovery scenario.
If we can know the process involved in accomplishing this,
we can test this scenario before bringing a Production
database go alive. Could you outline in a high level,
steps involved to perform this.
Thanks
A reader, April 22, 2002 - 2:04 pm UTC
April 22, 2002 - 9:00 pm UTC
That link -- if you wait for the page to load -- will take you to:
....
Restoring the Control File from a Backup Set Without Using RMAN
You must use a non-standard procedure to restore a control file from an RMAN backup set in the following situations:
* You are using a pre-8.0.5 version of RMAN to restore a database when more than one database with the same name is registered in the recovery catalog (see "Restoring When Multiple Databases Share the Same Name" for a discussion of this problem).
* You are not using a recovery catalog, and your only control file backup is in an RMAN backup set.
............
you are the second * there.
Reader
A reader, May 02, 2002 - 10:09 am UTC
Tom,
I was able to try out this scenario successfully
One caveat though:
If you backup using RMAN (ex. backup database ), the
controlfile is backed up with system datafile (default).
The control file from this backup set, when restored, is
unaware of the names of the backup sets
Probable solutions:
1. Subsequent to RMAN backup of database and archivelogs,
backup the control file in its own backup piece
2. Use the logfile from the RMAN to determine the names
of the constituent file names of backup set.
Use the PROCEDURES "restoredatafileto" and
"restoreArchivedlog" to restore datafiles and
archive logs
In either case, recover is conventional recover
SQL>recover database using backup controlfile until ....
SQL>alter database open resetlogs;
Your comments please
Thanks
Review my script....
Sikandar Hayat Awan, February 17, 2003 - 1:03 am UTC
I am using the following scripts to take full backup with RMAN on Oracle 8.1.7 (Win 2000). Please suggest me if I am doing any thing wrong or any improvement?
***I am using control file instead of catalog.***
--------RMAN_Backup.bat
--delete the previous backup files as it is moved to dat.
!del F:\HotBackup\*.* /q
rman target internal/oracle nocatalog @d:\cglmis\sikandar\rman\full_back.txt log=d:\cglmis\sikandar\rman\of_back.log
sqlplus internal/oracle @d:\cglmis\sikandar\rman\ctrlfile.sql
net send sikandar 'ORAPT backup Completed......'
echo on
Exit
--------Full_backup.txt
run {
allocate channel c1 type disk;
backup database format 'f:\hotbackup\oraptdb_%d_%t_%u.bus';
}
------ctrlfile.sql
alter system archive log current;
alter database backup controlfile to trace;
alter database backup controlfile to 'f:\hotbackup\ctrl_OF.ora';
Exit;
+++++++
I am taking full backup on daily basis please also suggest me that incremental/commulative will be the good option as the data is increasing. Any drawback of commulative backups at the time of restore (time is not my consideration at the moment) as I want to make sure 100% recovery of data.
For reference I am using online manuals and Oracle RMAN Pocket Reference by Darl Kuhn (O'REILLY).
how to read a Control file
Mir, May 11, 2004 - 1:07 pm UTC
Tom,
We had a situation when one of the harddrives failed and our database wont startup becoz it was not able to read the control file on that harddrive.
Do Oracle have or supply a utility which can read Control Files and display them in Text format. Much like "ALTER DATABASE BACKUP CONTROLFILE TO TRACE".
as always thank you very much.
May 11, 2004 - 3:04 pm UTC
if the control file is garbage, the control file is garbage (if the database cannot read it because it is on a failed harddisk, how would you read it?).
take one from your backup and use it to create a new control file from it. documented in the backup and recovery guide.
how to read a Control file
Mir, May 11, 2004 - 6:14 pm UTC
Sorry i wasn't very clear on my previous post. what i am tryin to do is figure out the directories for all my files associated with the database. In our case one harddrive failed so i was trying to figure out the list of files on various driectories. since this information is stored in the Control file i was wondering if oracle has some utility which can read the control file found on other drives, in order for us to figure out the layout of files for our DB.
May 11, 2004 - 9:02 pm UTC
if you have good control files -- I'm not seeing what your problem is????
just comment out the bad one from your init.ora and startup.
if you have at least one good control file -- that is all you need. not understanding what the problem is?
re: control file
Mark A. Williams, May 11, 2004 - 10:11 pm UTC
If you can successfully start the database using at least one good control file, you could do an 'alter database backup controlfile to trace' to get a 'text' version of the file that you could then read. Is that what you are wanting?
- Mark
is my customer lucky
Marvin, July 08, 2004 - 11:57 am UTC
Hi
I have a customer who is using controlfile for his RMAN backups. He has
control_file_record_keep_time integer 7
So one would expect he can only recover his databae using RMAN until 7 days ago however when I run list backup I can actually see RMAN recognizes backups from a month ago. When counting v$archived_log I can actually count logs from February to today.
SQL> select min(FIRST_TIME), min(COMPLETION_TIME), max(FIRST_TIME), max(COMPLETION_TIME) from v$archived_log;
MIN(FIRST MIN(COMPL MAX(FIRST MAX(COMPL
--------- --------- --------- ---------
05-FEB-04 05-FEB-04 08-JUL-04 08-JUL-04
select * from V$CONTROLFILE_RECORD_SECTION
TYPE RECORD_SIZE RECORDS_TOTAL RECORDS_USED FIRST_INDEX LAST_INDEX LAST_RECID
----------------- ----------- ------------- ------------ ----------- ---------- ----------
DATABASE 192 1 1 0 0 0
CKPT PROGRESS 4084 8 0 0 0 0
REDO THREAD 104 8 1 0 0 0
REDO LOG 72 32 6 0 0 9
DATAFILE 180 254 18 0 0 99
FILENAME 524 319 28 0 0 0
TABLESPACE 68 254 16 0 0 25
RESERVED1 56 254 5 0 0 9
RESERVED2 1 1 0 0 0 0
LOG HISTORY 36 2949 2949 113 112 11227
OFFLINE RANGE 56 291 0 0 0 0
ARCHIVED LOG 584 2867 2867 116 115 8883
BACKUP SET 40 612 612 495 494 2126
BACKUP PIECE 736 510 510 87 86 2126
BACKUP DATAFILE 116 563 563 476 475 2164
BACKUP REDOLOG 76 23859 23859 14545 14544 41842
DATAFILE COPY 660 519 2 1 2 2
BACKUP CORRUPTION 44 371 0 0 0 0
COPY CORRUPTION 40 408 0 0 0 0
DELETED OBJECT 20 7351 7351 7115 7114 69189
PROXY COPY 852 575 0 0 0 0
RESERVED4 1 8168 0 0 0 0
Is this a lucky case that controlfile data are not overwritten?
July 08, 2004 - 12:17 pm UTC
7 is the minimum, it can be much larger.
if for example, they had a period of really big activity for 7 days -- the control file would have grown to hold that.
Now, they settle down - the control file stays big and uses things in a circular (fifo) manner so much more than 7 would be there.
they don't even need a big period of activity -- could be that their activity is just slow enough that given the size of their control files, it sticks around.
great
Marvin, July 08, 2004 - 3:55 pm UTC
Hi
Thanks for the explanation, is there a way to shrink a control file?
Cheers
July 08, 2004 - 8:42 pm UTC
you can recreate them...
large controlfile - should I be concerned?
AR, November 23, 2004 - 12:53 pm UTC
Tom,
I just inherited a 500G database where the control_file_record_keep_time has been set to 270days. Consequently, controlfile sections have been expanding every couple of weeks to accomodate this value. The controlfile size is upto 250Mb already (and still growing).
Should I be concerned about this large controlfile? Yes, an RMAN "resync catalog" would be slightly slower with a large controlfile. But are there any other performance implications? (waits??) Is there an immediate need for me to reduce the keep_time parameter and recreate the controlfile?
Thank you for your time.
November 23, 2004 - 2:10 pm UTC
other than rman wanting to exclusively lock it during some backup operations - while getting a copy (250m might take longer than your average control file copy)-- I'm not aware of any issues.
might back off that keep time if you feel you have sufficient history to avoid it growing more.
if you do hit an issue, at least they are very easy to rebuild "empty"
controlfile size
Sam, August 17, 2005 - 10:34 pm UTC
Is there a limit on the size of controlfile? My colleague was saying that on unix/linux platform it is about 20000*oracle block size? He said he heard from someone. I could not find any oracle document discussing about it? Could you share if you have any information on this? Thanks.
August 18, 2005 - 3:45 pm UTC
Controlfile Rebuild
reader, September 28, 2005 - 8:20 am UTC
When we rebuild the Controlfile we will loose the RMAN information stored in the controlfile.Is there a way to keep them and rebuild .
Why I am asking this because I think I read somewhere that with 10gr2 it is possible (but lost the link)
Thanks
abort MAXLOGHISTORY
alan tan, May 09, 2007 - 2:07 am UTC
Hi,Tom:
I create a controlfile as parameter like:
CREATE CONTROLFILE REUSE DATABASE "STUDY" RESETLOGS ARCHIVELOG
2 -- SET STANDBY TO MAXIMIZE PERFORMANCE
3 MAXLOGFILES 20
4 MAXLOGMEMBERS 5
5 MAXDATAFILES 100
6 MAXINSTANCES 1
7 MAXLOGHISTORY 50-- I set this value.
¿¿¿
and then I backup control to trace,I got:
MAXLOGFILES 20
MAXLOGMEMBERS 5
MAXDATAFILES 100
MAXINSTANCES 1
MAXLOGHISTORY 226 -- is differenct !
why?
Thank you !
May 11, 2007 - 9:33 am UTC
some things are rounded up to minimums based on other parameters, your operating system and the like.
sequence# and thread# in RAC
jian huang zheng, April 27, 2009 - 11:07 am UTC
Hi Tom
I want to know sequence# is unique within each thread or globally unique, such as
could this make sense for 2 node RAC:
thread 1 sequence# 20998
thread 2 sequence# 20998
Thanks!
April 27, 2009 - 2:38 pm UTC
you need both to identify a redo log file, yes.