Flashback database question

I had a thought the other day and you want to run it by other professionals. Currently, my organization runs a performance test weekly with one of our database. What typcally happens, it is that we back up the database, run the load test, and restore the database. My thought was rather than go through all that, use the flashback database and flash of the database until before the load test started. The database is on Linux 32-bit and 10.2.0.4. Am I think it properly? It should work? Let me know your opinion. Thanks in advance

Published by: kirkladb on 30 Sep 2012 15:45

Flashback Logs introduce overhead. Overhead costs were particularly bad in point 10.2 (bug 6168063) but are supposed to be lower at 11.2

Hemant K Collette

Tags: Database

Similar Questions

  • Basic question about Flashback Database


    Hello

    I have a very generic question on the use of Flashback Database.

    On my systems tests, for use in testing and simulation performance, I want to create a guaranteed restore point so I can test impact on lot when releases of code changes are made before production deployment.  My confusion is in what regards newspapers of restoration, which are summarized in the following questions:

    1. is it possible to change the redo log files, when a guaranteed restore point has been configured?

    2. If so, is the Flashback to restore, point also change the size of the logs again?

    I found nothing in the documentation on this subject... Where my question...

    Appreciate your time taken to answer these questions...

    Kind regards.

    Hello

    donneskold wrote:

    Hello

    I have a very generic question on the use of Flashback Database.

    On my systems tests, for use in testing and simulation performance, I want to create a guaranteed restore point so I can test impact on lot when releases of code changes are made before production deployment.  My confusion is in what regards newspapers of restoration, which are summarized in the following questions:

    1. is it possible to change the redo log files, when a guaranteed restore point has been configured?

    (1) Yes it is possible.

    2. If so, is the Flashback to restore, point also change the size of the logs again?

    (2) it will not change the size of the redo log file.  Cannot resize a redo log file...

    I found nothing in the documentation on this subject... Where my question...

    Appreciate your time taken to answer these questions...

    Kind regards.

    Thank you

  • A new question on how to enable Flashback Database on the database pending?

    Best practical side says:

    Activate the Flashback Database on the database shall minimize downtime for logical corruption.

    I would like to know where to find the document or everybody already had experience in this topic?

    Also, I would like to know what is the best practice for backup/restore of the database of physical relief?

    Thanks in advance.

    Hello;

    Just an opinion, but I like the book 'Oracle Data Guard 11g Handbook' as the document and the answers to the second question.

    Chapter 12 covers the A to Z of RMAN with Data Guard.

    This book sold me since Flashback Flashback allows a former primary database to be reinstated after a failover with ongoing restoration operation (provided that enough newspaper flashback is available)

    My own experience is I tried a test failover without it and trash my primary test.

    I did a test of simple failover a few weeks ago, here are my notes.

    http://www.Visi.com/~mseberg/Data_Guard_Failover_Test_using_SQL.PDF

    Best regards

    mseberg

  • Flashback database syntax question

    EE Oracle 10.2.0.4 on OEL 5

    This should be a matter of syntax simple, right?

    Practice some scenarios to return back and looking at the syntax to
    http://download.Oracle.com/docs/CD/B19306_01/backup.102/b14194/rcmsynta032.htm#sthref515

    In the examples, they show a
    RMAN> FLASHBACK DATABASE TO TIMESTAMP
       TO_TIMESTAMP('2002-03-11 16:00:00', 'YYYY-MM-DD HH24:MI:SS');
    but the real syntax diagram does not show an option "TIMESTAMP", just "on TIME" and "AHEAD of TIME".

    Based on, let's try a simple test
    RMAN> flashback database to time = '18-MAY-2010 12:50:00';
    
    Starting flashback at 18-MAY-2010 13:10:00
    using channel ORA_DISK_1
    
    RMAN-00571: ===========================================================
    RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
    RMAN-00571: ===========================================================
    RMAN-03002: failure of flashback command at 05/18/2010 13:10:00
    ORA-01830: date format picture ends before converting entire input string
    Hmm, okay, then what format does? In fact, he reported in two different formats:
    Starting flashback at 18-MAY-2010 13:10:00
    and
    RMAN-03002: failure of flashback command at 05/18/2010 13:10:00
    If pressing on some others, let's try that should be obvious:
    RMAN> flashback database to time = to_date('05/18/2010 12:50:00','mm/dd/yyyy hh24:mi:ss')
    
    RMAN-00571: ===========================================================
    RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
    RMAN-00571: ===========================================================
    RMAN-00558: error encountered while parsing input commands
    RMAN-01009: syntax error: found "identifier": expecting one of: "double-quoted-string, single-quoted-string"
    RMAN-01008: the bad identifier was: to_date
    RMAN-01007: at line 1 column 30 file: standard input
    
    
    RMAN-00571: ===========================================================
    RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
    RMAN-00571: ===========================================================
    RMAN-00558: error encountered while parsing input commands
    RMAN-01009: syntax error: found "(": expecting one of: "allocate, alter, backup, beginline, blockrecover, catalog, change, connect, copy, convert, create, crosscheck, configure, duplicate, debug, delete, drop, exit, endinline, flashback, host, {, library, list, mount, open, print, quit, recover, register, release, replace, report, renormalize, reset, restore, resync, rman, run, rpctest, set, setlimit, sql, switch, spool, startup, shutdown, send, show, test, transport, upgrade, unregister, validate"
    RMAN-01007: at line 1 column 37 file: standard input
    
    RMAN>
    Built on the docs and other resources on the web, I find several variations on the theme, but all end up throwing one of the errors above.

    I can't believe, I've been reduced to what should be a matter of doc, but in this case, it's the documentation examples which throw errors. I'm going blind? The brain cell was killed by that I had last week of beer criticism? ;-)

    From:

    RMAN> flashback database to time = to_date('05/18/2010 12:50:00','mm/dd/yyyy hh24:mi:ss')
    
    RMAN-00571: ===========================================================
    RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
    RMAN-00571: ===========================================================
    RMAN-00558: error encountered while parsing input commands
    RMAN-01009: syntax error: found "identifier": expecting one of: "double-quoted-string, single-quoted-string"
    RMAN-01008: the bad identifier was: to_date
    RMAN-01007: at line 1 column 30 file: standard input
    

    Could you please try adding double quotes around the expression date?

     flashback database to time = "to_date('05/18/2010 12:50:00','mm/dd/yyyy hh24:mi:ss')";
    
  • Files without a NAME during the Flashback Database

    Hi all

    Why, in flashback, database, retrieve a tablespace has dropped the datafile (s) consisting of within table get renamed to UNNAMEDnnn? I was going through some examples on the internet but did not find any concrete reason as to why this happens during the flashback recovery. If the tablespace can be recreated during the process, why not data files that compose?

    Could someone help me please in my understanding.

    Thank you.

    Prithwish Adhikari wrote:

    Hello

    I was talking about the flashback database to retrieve a tablespace has fallen. Please find below the link

    http://www.DBA-Oracle.com/t_rman_159_restore_dropped_tablespace.htm

    My question is

    "

    flashback

    database to the SNA 460217.

    database of flashback to the SNA 460217

    *

    ERROR on line 1:

    ORA-38795: WARNING: flashback has managed but open resetlogs would get error

    below

    ORA-01245: offline file 5 will be lost if resetlogs is done

    ORA-01111: name for the data file 5 is unknown - rename it to the correct file

    ORA-01110: data file 5: ' / u01/oracle/product/10.2.0/db_1/dbs/UNNAMED00005'

    "

    Why, in flashback, database, the data file associated with the storage space gets renamed to unnamed?

    Thanks for the link and new info.

    The reason is the same as in these quotes from doc that I have provided. As ORA-01111 indicates the name of the actual file is unknown so Oracle creates a unique name, then stops recovery so that you can provide the name and location.

    Why, during the flashback database, retrieve a tablespace has fallen...

    This formulation and the wording of the article, is probably what is mislead you. You cannot recover a tablespace using FLASHBACK DATABASE:

    Here is the doc for FLASHBACK DATABASE

    http://docs.Oracle.com/CD/B28359_01/backup.111/b28273/rcmsynta023.htm

    Notes A Flashback Database use operation applies to the entire base. You cannot flash back, individual storage spaces. A Flashback Database operation is similar to a point-in-time data recovery (DBPITR) with RECOVER , but RMAN uses flashback logs to undo changes to a point before the time runs out or SNA

    See that second sentence? You NEED to recover the entire database and, as this quote from doc says, made it like CAMILLE by RECOVER.

    In this example, the storage space AND data files were dropped. So please check the doc to DROP TABLESPACE

    https://docs.Oracle.com/CD/B28359_01/server.111/b28286/statements_9004.htm

    Oracle database deletes all metadata about the space of storage data files and all the data dictionary tempfiles in the tablespace. The database also automatically removes the operating system all data files managed by Oracle and tempfiles in the tablespace. Other data files and tempfiles are not removed from the operating system unless you specify INCLUDING CONTENTS AND DATAFILES .

    See? No additional data in the dictionary for the tablespace or data files. Real datafiles is more if they were managed by Oracle. But there MAY be data files of the operating system if you do NOT use this last clause.

    If Oracle recreates the files using unique names and then allows you to CONFIRM that the name and location should really be.

    You have a problem with that? Or just try to understand?

  • ORA-38754: FLASHBACK DATABASE not started; required redo log is not available

    Hello

    I am getting below error when executing the activities of flashback. Currently we have backups, so when I tried the command flashback in RMAN my question solved.

    But if I want to do the same thing using SQLPLUS what steps should I take?

    ERROR:

    ORA-38754: FLASHBACK DATABASE not started; required redo log is not available

    ORA-38762: redo logs needed for SNA to SNA 20062842936 20062842926

    ORA-38761: redo log sequence 4 in thread 2, 65 incarnation cannot

    consulted

    DATABASE:11.2.0.3.0 (UAT)

    RMAN will automatically restore logs archived redo required for the operation of blowback when they are not present on the drive, while (as meyssoun said) when using sqlplus required archived redo logs must be available on the disc.

    Concerning

    AJ

  • How to use Flash recovery only for flashback database feature box

    I would use the area of flash recovery only for flashback database feature, which means, the flash recovery area will store only the flashback logs.

    I don't want to use this file system for all other files, such as files of archive log, backups, copies of control files, multiplex redo log, etc.

    is this possible? I want to use this function only for the flashback database option and therefore, I don't want to disturb any existing configuration in my database

    including backup jobs. It will also store the Archives of redo log files in the flash recovery area to use the flashback database feature?

    I can't certainly allow me to have a copy of the entire file system database as my DB size is more than the system recovery flash I have zone files.

    Thank you

    Delphine

    user13312943 wrote:

    Aman,

    I think I was not clear in my question. I'll try again

    My only requirement is to use the flashback database feature (or be prepared to use if necessary). I don't want to create a large file system to use the CRF to store all those files. Out of all the files are stored here a copy of the data files seems to consume more space. I see that these files are classified as transitional and permanent files. I agree with permanent files. I'm afraid that this file system can use more space if I store files of backup to disk. I want backups of databases to go to bands, not on the drive of FRA. Is this possible?

    Thanks for your response

    Okay, well in that case you can use RMAN and push the backups to tape drives. FLB newspapers would be stored on the FRA.

    Aman...

  • FLASHBACK database alternatives

    We are poised to move to 11 GR 2, and so I thought I would ask the question just in case there was something again...

    Are there alternatives to FLASHBACK DATABASE to restore DDL? I was hoping that there is a new feature that I'm not aware, allowing for the same functionality at a lower level, at the level of the schema. I guess that we will not here yet based on the research I have done this morning, but I thought I'd ask.

    Thank you

    -= Chuck

    chuckers wrote:

    We are poised to move to 11 GR 2, and so I thought I would ask the question just in case there was something again...

    Are there alternatives to FLASHBACK DATABASE to restore DDL? I was hoping that there is a new feature that I'm not aware, allowing for the same functionality at a lower level, at the level of the schema. I guess that we will not here yet based on the research I have done this morning, but I thought I'd ask.

    Thank you

    -= Chuck

    The alternative to 'Flashback' is just 'recover' If you want to have an option "move forward." But for "replacement" Flashback Database, I'm not sure I remember something else than he.

    HTH

    Aman...

  • FLASHBACK Database: rman vs. sql command differences

    I am new to Oracle and want to understand the difference and that is the correct way of rear turn signals to the database?


    RMAN command vs SQL command on the use of FLASHBACK DATABASE
     SQL> FLASHBACK DATABASE TO RESTORE POINT BEFORE_ONLINE_TEST;  {CODE} Vs 
    RMAN > run {}
    flashback database to restore the BEFORE_TEST_320130313 point;
    SQL ' alter database open resetlogs; }
     
    Referring this article on http://docs.oracle.com/cd/E11882_01/backup.112/e10642/rcmflash.htm#BGBDCAFA it says we can use both RMAN and SQL command for database flashback.
    
    Questions:
    
    1. Difference between using RMAN and SQL command on database flashback?
    
    2. When to use RMAN or SQL on the  database flashback?
    
    3. Which is the Oracle recommended or best way ?
    
    Edited by: Sivaprasad S on Mar 15, 2013 7:26 PM
    
    Edited by: Sivaprasad S on Mar 15, 2013 7:26 PM
    
    Edited by: Sivaprasad S on Mar 15, 2013 7:27 PM
    
    Edited by: Sivaprasad S on Mar 15, 2013 7:27 PM
    
    Edited by: Sivaprasad S on Mar 15, 2013 7:28 PM
    
    Edited by: Sivaprasad S on Mar 15, 2013 7:29 PM                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            

    So when you run your SQLPlus flashback, you get the following error:

    ORA-38761: redo log sequence 1 in thread 1, incarnation 3 could not be found

    This means so that the sequence 1 has probably been removed from the disk and is necessary.

    When you run the RMAN command look at the log what RMAN:

    * journal archived for thread 1 sequence 2 is already on the disc in the file +DG_PERF_FLASH_02/TestDB/archivelog/2013_03_14/thread_1_seq_2.3612.810054205*
    * journal archived for thread 1 with sequence 3 is already on the disc in the file +DG_PERF_FLASH_02/TestDB/archivelog/2013_03_14/thread_1_seq_3.1588.810054205*

    So sequence 2 and 3 are already on record according to the above message.

    Then you get this message:

    channel dev_2: restoration of archives Journal
    Archives journal thread = 1 sequence = 1

    So RMAN knows he needs 1 thus sequence and restores this newspaper archive she needs for recovery.

    In a Word while running SQLPlus she communicate with RMAN and restore the log file for you, however with RMAN, it won't.

    You can restore the newspaper archive manually, and then run in SQLPlus that will then work.

    Published by: Freddie Essex on March 18, 2013 12:46

  • Flashback Database Architecture

    I went to the Architecture of database of Flashback in which he is mentioned process RVWR writes buffer of Flashback Flashback data to newspapers Flashback Database, which is used circularly.

    My question to all the experts here is-
    (a) Flashbacking the database depends on the content of database of Flashback logs?
    (b) given that the flashback logs is used circularly and there's no archives generated, if the content is overwritten, can't go us back to the point
    who's already crushed?

    -Neel

    Hi Neel,

    I guess I see where the confusion is. In the line,.

    This new background process sequentially written Flashback Database data from the buffer of flashback in the Flashback Database logs, that are reused circularly.

    Note the comma after "Newspaper Flashback Database" so, 'who' does not point in newspapers Flashback Database, but in the buffer of blowback.

    Kind regards
    Thierry

  • CSALTER and Flashback Database

    In 10.2.0.4, is it possible to use Flashback Database to cancel a character occupies change that was made through CSALTER? I think it fits very well, especially with CSALTER, it's just changes to the dictionary and the remapping is not the characters, but I wonder if anyone is aware of any questions/limitation in such a scenario.

    Yes. Flashback Database literally back the database to a previous point in time.

    That said... I've never tried it with CSALTER so my advice would be to validate on a test area first.

  • Can feature FLASHBACK DATABASE undo the loss of control of file/ORL?

    DB version: 11.2.0.4

    Platform: Oracle Linux 6.4

    I personally have not tested the FLASHBACK DATABASE feature. I came across examples of logical errors such as truncate Table being overturned on FLASHBACK DATABASE.

    I guess that FLASHBACK DATABASE feature may undo the loss of the data files. But will be a loss of control or the Group of newspapers online file again defeated using FLASHBACK DATABASE?

    Hello

    "I guess that feature FLASHBACK DATABASE can cancel the data files loss." Do not assume, check it out

    Flashback database cannot cancel the loss of data files, control files and redo log files.

    Flashback Database limitations

    Because Flashback Database works by undoing changes to data files that exist at the present time when you run the command, it has the following limitations:

    • Flashback Database can only undo changes in an Oracle database data file. It cannot be used to fix the failings of the media, or recover the accidental deletion of data files.
    • You cannot use Flashback Database to cancel a file of data reduction operation. However, you can disconnect the wizened file, flash back, the rest of the database, and then later restore and restore the shriveled data file.
    • You cannot use Flashback Database only to retrieve a data file. If you Flash back, a database at a time when a data file exist in the database, that the entry of data file is added to the control file. You can only recover the database using RMAN to fully restore and recover the data file.
    • If the database control file is restored from backup or re-created, all the accumulated flashback journal information is ignored. You cannot use FLASHBACK DATABASE to get back to a point in time before the restoration or re-creation of a control file.
    • When you use Flashback Database with a time of target during a NOLOGGING operation was underway, corruption block is likely in database objects and files of data that is affected by the NOLOGGING operation. For example, if you perform a direct path INSERT operation in NOLOGGING mode and it runs from 09:00 to 09:15 April 3, 2005, and then use you Flashback Database to return to the target time 09:07 on this date, the objects and data files updated by the direct path INSERT have finished block operation Flashback Database corruption. If possible, avoid using Flashback Database with a time or SCN coinciding with a NOLOGGING operation. Also, perform a full or incremental backup of the immediately affected data files backup after each NOLOGGING operation to ensure recovery of points in time after the operation. If you plan to use Flashback Database to return to a point in time in the course of an operation, for example a direct path INSERT , consider performing the operation in LOGGING mode.

    See: https://docs.oracle.com/cd/E11882_01/backup.112/e10642/flashdb.htm#BRADV286

  • ORA-38760: this database instance could not turn on the flashback database

    Hi guys

    Without knowing it, I deleted all the files in the directory "fra" where "flashback" folder was there because of space issues. Now am facing the error.

    It's my practical database.

    SQL > startup

    ORACLE instance started.

    Total System Global Area 608174080 bytes

    Bytes of size 1268920 fixed

    184550216 variable size bytes

    419430400 of database buffers bytes

    Redo buffers 2924544 bytes

    Mounted database.

    ORA-38760: this database instance could not turn on the flashback database

    SQL > alter flashback database;

    Database altered.

    SQL > SHUTDOWN IMMEDIATE;

    SQL > STARTUP MOUNT

    SQL > select flashback_on from database v$.

    FLASHBACK_ON

    ------------------

    RESTORE POINT ONLY

    Am not able to extinguish the flame.

    SQL > select * from v$ restore_point;

    Select * from v$ restore_point

    *

    ERROR on line 1:

    ORA-38701: Flashback database log 1 seq 1 wire 1:

    "/ u01/fra/PROD/flashback/o1_mf_c1l5n3lq_.flb".

    ORA-27037: unable to get file status

    Linux error: 2: no such file or directory

    Additional information: 3

    SQL > select flashback_on from database v$.

    FLASHBACK_ON

    ------------------

    RESTORE POINT ONLY

    I also tried to create a restore point and the flashback there.

    SQL > create restore point guaranteed rspt1 flashback database;

    Restore point created.

    SQL > select * from v$ restore_point;

    Select * from v$ restore_point

    *

    ERROR on line 1:

    ORA-00600: internal error code, arguments: [krfg_get_keepseq_2], [1], [0], [0,]

    [0], [0], [0], [0]

    SQL > flashback data base to restore the point rspt1;

    flashback database to restore the rspt1 point

    *

    ERROR on line 1:

    ORA-38729: not enough flashback database data log do FLASHBACK.

    Please help me in this regard.

    Alert log file:

    ...

    ...

    ...

    Mount of redo thread 1, with mount id 305866694

    Sun 25 Oct 10:13:30 2015

    Bytes allocated of 3981204 in the shared pool for the generation of flashback buffer

    From RVWR background process

    RVWR started with pid = 16, OS id = 15486

    Sun 25 Oct 10:13:30 2015

    Errors in the /u01/oracle/product/10.2.0/db_1/admin/prod/bdump/prod_rvwr_15486.trc file:

    ORA-38701: seq 11 Flashback database log 11 1 thread: ' / u01/fra/PROD/flashback/o1_mf_c2oqypb0_.flb '.

    ORA-27037: unable to get file status

    Linux error: 2: no such file or directory

    Additional information: 3

    Sun 25 Oct 10:13:30 2015

    Database mounted in exclusive Mode

    Completed: ALTER DATABASE MOUNT

    Sun 25 Oct 10:13:30 2015

    ALTER DATABASE OPEN

    ORA-38760 marked during: ALTER DATABASE OPEN...

    Trace file:

    ...

    ...

    ...

    :() 2015-10-25 10:13:30.087 SERVICE NAME

    SESSION ID: (156.1) 2015-10-25 10:13:30.087

    ORA-38701: seq 11 Flashback database log 11 1 thread: ' / u01/fra/PROD/flashback/o1_mf_c2oqypb0_.flb '.

    ORA-27037: unable to get file status

    Linux error: 2: no such file or directory

    Additional information: 3

    1. Thank you.

    you experience serious problem... because usually it works very well.

    Try the following:

    SQL > shutdown immediate;

    SQL > bootable media.

    SQL > alter flashback database;

    SQL > alter database backup controlfile to trace noresetlogs;

    SQL > shutdown immediate;

    go to the trace file generated from running (alter database backup controlfile to trace noresetlogs ;)

    location of the trace file is located in the query:

    SQL > select * from v$ diag_info;

    It will contain the script with the SQL commands, you can rename the file with the extension ".sql".

    SQL > startup nomount

    SQL > @trace_file. SQL

    SQL > restore database;

    SQL > alter database open;

    So basically you are going to recreate the files of control again...

  • Several Flashback Database to the time prerequisites

    Current time: 17/12/2013-13:22

    Created restore point guaranteed: 10:53 09/12/2013.

    More old Flashback Time: 10:34 09/12/2013


    A Flashback Database timestamp 17/12/2013-06:30 was carried out at the 12:00 17/12/2013. (Done Resetlogs).  It's my understanding that a new incarnation is created when I resetlogs.  Guaranteed restore point was created under the old incarnation, but it seems to keep the archivelogs recovery guaranteed, even if the database has been Databased Flashback.

    Is it possible database flashback by timestamp 10:34 09/12/2013 (oldest flashback time) even if we have already done a recovery (flashback by timestamp database) at 17/12/2013 06:30?  I am concerned about corruption.

    Thanks in advance.

    See this and the ancestral discussion below: scene of Flashback and the database Point-in - time recovery

  • flashback database

    Experts - which would be the overload if configure us the database of blowback in the production db, and what are the different practices for the flashback database (11g r1). How archive flashback data differs from the flashback database?

    user12075620 wrote:

    Aman thanks for your response. It is the light version of the informatica ETL process that extracts data from db2 based on simple business rules. The NFR would be maximum 1-3 minutes for each transaction. Insert operation initially and after that for the most update operation. This situation suggest the flashback database?

    Yet once, I would say that it is something that you must carefully test. Are all tables have equally important you need to keep data undo for each of them? If Yes, then flashback database is correct to be put in place, but if not, I suggest to use the data of Flashback of archives that would be specific to the table of interest.

    HTH

    Aman...

Maybe you are looking for