RMAN Duplicate + incremental backups + upgrade to 12 c

Similar Questions

• RMAN DUPLICATE: Use backup or CURRENT database

  Hi all

  I have a database of 32 TB, and about what I've experienced so far, using backup to duplicate the database is similar to the use of the command to RESTORE and RECOVERY manually.

  If I use the option of DATABASE ACTIVE, then I should be concerned about the missing log files?

  Hereby, I mean, text source and target are in ARCHIVELOG MODE anyway.

  Thank you

  To answer your original question, yes you are worried for your missing log file.

  RMAN automatically copies the file to the server in the destination host settings, start the auxiliary instance with server settings file, copy all database files needed and archived redo logs over the network to the destination host and recovers the database.

  If your copy of 32 TB on the network would take more than 1 day (what is your strategy of retention according to your other thread), so what, you won't be able to recover the database.

  And you must mark this thread as a response and continue in your original thread.

• RMAN size incremental backups have increased.

  In a database of CARS of two nodes, the size of the file incremental RMAN is usually around 400-500 MB and RMAN complete file size takes up approximately 10 GB.
  Suddenly, the RMAN incremental file size has increased to about 16-17 GB and there are no extension on the block of data file size data of size/volume/DB etc.
  
  What could be the reason for the sudden increase in the size of the RMAN incremental file?
  
  Thank you
  
  Database:
  BANNER
  --------------------------------------------------------------------------------
  Oracle Database 11 g Release 11.1.0.6.0 - 64 bit Production
  PL/SQL release 11.1.0.6.0 - Production
  CORE 11.1.0.6.0 Production
  AMT for Linux: Version 11.1.0.6.0 - Production
  NLSRTL Version 11.1.0.6.0 - Production
  
  OPERATING SYSTEM:
  Linux 2.6.18 - 92.1.22.el5 #1 SMP node1 kills Dec 16 11:57:43 UTC 2008 x86_64 x86_64 x86_64 GNU/Linux

  Have you checked the abnormal generation of redo. East - the number of the generated archivelogs identical or similar to the period you are comparing?
  Would it be because you save including archivelogs topic archivelogs.

  ------------------------------------------
  Coskan GUNDOGAR

  Oracle DBA
  Blog: http://coskan.wordpress.com
  Twitter: http://www.twitter.com/coskan
  LinkedIn: http://uk.linkedin.com/in/coskan
  ------------------------------------------

  Edited by: coskan on May 27, 2010 22:56

• Duplicate the database and incremental backups

  Hello
  
  I reviewing options to maintain a test server that is "relatively" updated compared to the production server.
  This is my first contact with RMAN and I went through the books, but as I have not had a chance to play with it I could have missed something.
  
  It is an Oracle devoted to a Redhat Enterprise Linux (not sure of the complete version number) 11.2.
  
  The test server must be writable and must be synchronized with the server in production about once or twice a month. Due to the size of the database (~ 600 GB) we are studying solutions that could make using incremental backups that are already taken on the production database.
  
  While launching the test using RMAN DUPLICATE server seems to be honest, I'm not sure how the "sync" can be done as effectively as possible - especially because we are changing the data on the test server.
  
  After the RMAN BACK command description, is not clear to me if this could serve to 'recover' the database of test based on the incremental backups to the 'current state' of the production database, even if the test database has been changed. After reading a few posts here, it's not seems to be the case.
  
  A restore point would make such a possible incremental restore? Creating a (name?) point of restoration immediately after duplication, and once the tests are finished making back to the restore point and then apply the incremental backups that have accumulated since then.
  
  
  Another of the options seems to be to create a database of the previous day. Then when running the tests, we turn to the backup database in one instantaneous standby database open read/write and run tests. Once we have finished, we he switch back to a physical standby. This switchback to rewind the changes and recovery logs accumulated are applied as "incremental" catching up with production - which is basically what we want.
  
  Is there another solution to this? Or use a watch of the snapshot, the most effective solution? If possible, we would like to avoid transferring the 600GB whenever we want to re - synchronize the database to test.
  
  Thanks in advance
  Thomas

  Hi Tomas. I think you're right. Ensures the most effective solution in your situation of snapshot.

• RMAN incremental backup

  Hi all

  I have a question, I am on RHEL-6 / Oracle 11 GR 2.

  After taking the backup of level 0 (RMAN), if I plan tale level1 backup then how oracle knows that they are the backup piece is considered a level 1 backup.

  I mean, that's how oracle knows that these are the blocks that had to be backeup.

  Kind regards.

  Please refer to the https://docs.oracle.com/cd/B19306_01/backup.102/b14192/bkup004.htm documentation

  Article: 4.4.1 Alogrithm incremental backup

  -Jonathan Rolland

• Before physical deployment standby with RMAN incremental backups.

  This method is available in 10g:

  http://Web.Stanford.edu/dept/ITSS/docs/Oracle/10gR2/server.102/b14239/scenarios.htm#CIHIAADC

  For 11g:

  http://docs.Oracle.com/CD/E11882_01/server.112/e41134/RMAN.htm#SBYDB00759

  For 12 c:

  http://docs.Oracle.com/database/121/BRADV/rcmadvre.htm#BRADV685

  According to dosc 10 g, the scenario is simple:

  1. create an incremental backup to SNA required on primary and copy it in standby mode

  2 save the backup in sleep mode and perform a recovery

  11 g dosc, we must recreate the standby control file and update the file names before performing the recovery.

  12 c introduces the new method - recovery on network, but again, we must recreate and standby control file update.

  The question is: why we must create new controlfile ensures g 11 and 12 c and can we use a simple method of 10g?

  Thank you.

  Hello

  1 why the incremental backup application does not update standby controlfile?

  -Your database pending is be retrieved using the incremental backup. This does not update the eve controlfile. Your standby controlfile completely ignores the last CNS to the primary database.

  2. what happens if we leave the old standby controlfile? RFS will record new logs of the primary, but will ignore them MRP? If so, can update us standby controlfile by manually applying the following log?

  -Your old standby controlfile is unaware of the last CNS to the primary database.

  As Mseberg said, trying things on a test computer and you will have the answers to your questions.

  Kind regards

  Shivananda

• Case of level 1 RMAN incremental backup failure...

  Hello

  I have a question. If the incremental level 1 RMAN backup fails, the next level 1 backup attempt will have backup all changes since the last backup of level 0 OR level 1 successful?

  Best regards

  Hello

  I have a question. If the incremental level 1 RMAN backup fails, the next level 1 backup attempt will have backup all changes since the last backup of level 0 OR level 1 successful?

  Check the dictionary BACKUP_SET V$ to confirm if the last backup has been affected by the data dictionary. If incremental backup failed is picked up then the last incremental backup will be required in the case of the next LEVEL 1 DIFFERENTIAL INCREMENTIELLE backup.

  But to be on the safe side, it is recommended that you take a LEVEL 01 CUMILATIVE INCREMENTAL backup of level 0 up to this day. For the recovery process can go through even if the previous backup is not valid.

  Thank you &

  Best regards

• RMAN incremental backup continue to make backups "complete"?

  Hello, I am new to RMAN. I used only the full backup by RMAN, and it was OK. Now, due to lack of disk space, I decided to make incremental backups.

  I put 'to recovey retention policy window of 3 days; then made a new backup (by MS) job, looks like:

  $rman_script = ' backup database cumulative device type disk tag '% TAG' extra strength 1;

  backup disc tag '% TAG' type device archivelog all not saved delete all entries;

  But, always visible, size of the backup is the same all the time as a full backup? Where is my mistake?

  This is a backup log:

  RMAN > database backup incremental level 1 cumulative device type disk tag "BACKUP_DB10_00002_070413021501".

  From backup 4 July 13

  allocated channel: ORA_DISK_1

  channel ORA_DISK_1: sid = 1043 devtype = DISK

  channel ORA_DISK_1: additional start-up compressed 1 datafile backupset

  channel ORA_DISK_1: specifying datafile (s) in backupset

  input datafile fno = name=/database/app/oracle/oradata/db/tablespace1.dbf 00007

  input datafile fno = name=/database/app/oracle/oradata/db/tablespace2.dbf 00002

  ...

  channel ORA_DISK_1: starting total, 1 to 4 July 13

  channel ORA_DISK_1: finished piece 1 to 4 July 13

  piece handle=/backup/oracle/flash_recovery_area/db/backupset/2013_07_04/o1_mf_nnnd1_BACKUP_DB10_00002_0_8x9d45f5_.bkp tag = BACKUP_DB10_00002_070413021501 comment = NONE

  channel ORA_DISK_1: complete set of backups, time: 00:41:15

  Backup finished on 4 July 13

  > database backup incremental level 1 cumulative device type disk tag "BACKUP_DB10_00002_070413021501".

  There is a Note of support Oracle ' size of level 1 or level 1 cumulative backup greater than level 0 backups [ID 1534691.1].

  which applies if you use the DUNGEON and change the tag.

  I note that you change the tag every day, but do not use the clause to KEEP this note should not apply.

  Why not explicitly run EXTRA STRENGTH 0 backup once and then monitor subsequent LEVEL 1 backups?

  Hemant K Collette

• While you take the cumulative incremental backup RMAN. Meanwhile users ke

  While you take the cumulative incremental backup RMAN.
  During this time users keep be updated, insert the corresponding schema.
  But my Client doesn't backup up-to-date. Is it possible to get the up-to-date backup?
  (If I have something wrong, sorry about that, that was my interview question)
  
  Thanks for all,
  Kind regards
  DB

  DB wrote:
  While you take the cumulative incremental backup RMAN.
  During this time users keep be updated, insert the corresponding schema.
  But my Client doesn't backup up-to-date. Is it possible to get the up-to-date backup?
  (If I have something wrong, sorry about that, that was my interview question)

  Thanks for all,
  Kind regards
  DB

  RMAN backup is not compatible and easy to follow on-line transactions. If it's something any purpose of migration suggest taking COLD backup.
  If it's no way suggested, then after your entire + cumulative backups, however, that data will be available in archives

• Let the RMAN to know the location of the backup during RMAN DUPLICATE item

  RDBMS Version: 11.2.0.3
  Platform: AIX 6.1
  
  I spent a few discussions lately about the creation of standby DB by using the RMAN DUPLICATE method.
  
  Taking an example from a thread by steve74 yesterday.
  
  The commands below are executed after copying the elements of hot backup of the primary DB Backup on the standby server.
  

  --- Will be running the following DUPLICATE command from standby server 
   
  export ORACLE_SID=orcldg
  SQL> startup nomount pfile='/u01/app/oracle/11.2/db_home1/dbs/initorcldg.ora'
   
   
  rman target sys/mypass@PRIMARY_TNS auxiliary / 
   
  run 
  {
       allocate channel C1 device type disk;
       allocate auxiliary channel C2 device type disk;
       duplicate target database for standby nofilenamecheck;
  }

  The auxiliary instance is just in NOMOUNT. Lets say I copied the primary DB Backup items to the
  following in the standby server location
  

  /u04/backup/RMAN

  I don't see all the above commands which allows to RMAN knows about this place where are stored the backup pieces. For regular DB cloning, we are cataloging of backup items using the CATALOG command. But, you need a control file to do cataloging. The auxiliary instance does still not have control file. So, how this work?

  Hello

  simple. duplicate is a faster and simpler, method because you issue a command.
  After configuration, you issue a command, and if the pfile/spfile is properly configured, RMAN takes care of ALL the necessary measures, even for ASM (assuming the switch of data file also). It restores the CF, he set up the db, it restores, it retrieves, it updates the CF with the correct names, it converts file names if necessary, open the DB ETC.
  And, since only a single command is used, it is less prone to errors.
  And allows you to save a lot of trouble.
  Quite simply, you have a few details such as the location of the backuppieces.

  It works very well also for DEV env refreshes (with a few modifications on the pfile/spfile file).

  BR,
  jpinela.

  Published by: Pinela December 21, 2012 16:02

• When the RMAN brand is backup incremental level1 obsolete?

  Hi all
  
  On a server Oracle 10.2.0.2 particular (with ODG), I implemented a strategy of incremental backup. A level 0 backup is available and all level 1 next backups are getting recovered to this first copy of data file. So far so good.
  
  The problem is that all these incremental backups at level 1 stay in my box of flashback and are not cleaned by RMAN. The retention policy is set on a window of 1 day.
  
  When the RMAN will start scoring these backups to level 1 as being obsolete? I guess that these backups are obsolete right after that they were applied to the copy of the data file? This has something to do with the size of the parameter area flashback, which means that it will not begin to remove level 1 backups when RMAN runs into problems of free space on the disk?
  
  Any help is appreciated!
  
  JM

  Hi JM,

  This is expected behavior, because you have changed the default retention of redundancy strategy 1.
  You can find this on My Oracle Support note + incremental backup strategies: Merge [ID 745798.1] +.

  The extract of this says:
  >
  When using Oracle-Suggested policy or a daily incremental backup merged and using a Flash recovery area and then keep the retention policy of RMAN for redundancy 1. This will ensure that the Oracle space management will automatically remove incremental backups at level 1 of the FRA after they have been applied to the image based on the pressure of space in the FRA copy. If you increase the redundancy or set a window of recovery space is is automatically retrieved by Oracle. This works only when the retention policy is set to the default value of redundancy = 1. If the configured retention policy is set to the window of recovery or > 1 redundancy, RMAN cannot satisfy the retention policy if there is only one copy of database and it's at a later point in time. Thus, in order to satisfy political retention Oracle keeps all the incremental backups and archivelogs since the creation of the database because ideally a data file can be re-created and archivelogs can be applied to meet the retention policy (although not followed in the practical case).
  >

  To have made them obsolete automatically, you must add the 'UNTIL TIME 'SYSDATE-2' ' clause to the COPY of RECOVER sentence, for a 1-day recovery window.

  The order having to RETRIEVE it after the BACKUP, you will have the current copy image, and you should only apply the archivelogs if you need to recover instead of use the last day of Level0 backup now level1.

  Kind regards.
  Nelson

• Roll forward standby with RMAN incremental backup when used DG broker

  Hello
  
  I'm about to use the process defined in the note: steps to roll forward a database ensures the RMAN incremental backup using during primaries and Eve are in the file [836986.1 ID] ASM system. I'm on 11.1.0.7.
  
  However, it does not mention the DataGuard broker.
  Did someone played this process when you use the broker.
  The broker he manages properly?
  Does it matter if the broker is upwards or downwards during the physical recovery of standby or replacement of control file?
  Should the broker configuration be re-created?
  
  Thank you.

  user11993454 wrote:
  Hello

  I'm about to use the process defined in the note: steps to roll forward a database ensures the RMAN incremental backup using during primaries and Eve are in the file [836986.1 ID] ASM system. I'm on 11.1.0.7.

  However, it does not mention the DataGuard broker.
  Did someone played this process when you use the broker.
  The broker he manages properly?
  Does it matter if the broker is upwards or downwards during the physical recovery of standby or replacement of control file?
  Should the broker configuration be re-created?

  Thank you.

  (1) disable the broker
  (2) perform the Roll forard SNA.
  (3) select the broker

• Incremental backup RMAN and compression

  RMAN compresses data for additional membership 0 and level 1 backups. Specifically - can I use advanced compression if olny use incremental backups?

  Welcome.

  When in forums try and answer relatively useful and correct responses, helps us and as well as others who later discovers the forums when the answers are marked as having a value compared to the question.

• NOLOGGING with DEC and RMAN incremental backup

  DB Version: 10.2.0.4
  OS : AIX
  
  We use RMAN for our backup scheduled via cron.
  Retention Policy : REDUNDANCY 1
  Type             : Incremental 

  On the 31 of every month in most of our patterns, we recreate a lot of tables using DEC.
  
  Due to the size, the CREATE TABLE instructions take a while. Because we run on the Standard edition, we cannot use the PARALLELISM.
  NOLOGGING is the only option to improve performance. But we are concerned the roll forward not being generated for these tables and the potential loss of data in case of crash.
  
  Currently we take a Level0 backup every Wednesday and Sunday night and LEVEL1 backup on the rest of the days.
  
  I know that all will be saved on Level0. But what happens if 31 falls on a day which occurs the only backup of level 1. We will be safe only until the next Level0 arrives.
  
  No work around for this?

  HI T.Boyd,

  NOLOGGING is the only option to improve performance.

  So no recovery will be generated but the blocks in the data files will change.

  But what happens if 31 falls on a day which occurs the only backup of level 1.

  Only blocks changed since the last level 0 (or level 1 if you use differential differential) will be saved which includes blocks modified by the ETG.

  No work around for this?

  No need for a work around. You can check at any time whether a certain activity nologging will compromise a restoration with < report="" unrecoverable;=""> .
  Kind regards
  Tycho

• Resume failure RMAN incremental backup to disk 0 cold

  Linux x 86-64 10.2.0.4
  
  I take an incremental backup of 0 to disk in a database that is mounted. I'm backup the database on an NFS volume. I lost the NFS mounting. Is it possible to take a backup where it stopped at? He complains "ORA-27038: created file already exists" if I restart it. RMAN can easily return to a restore. I don't have time to redo everything all again. The database is greater than 10 TB. The target is not upward. Is there something else I could do?

  If you have lost the NFS mounting, has not lost what had been saved to this mount? So, don't you create a fresh backup of all the files again?

  How are you "restart" backup? If you issue a fresh BACKUP DATABASE, it must create a new backup, with new backupsets - I hope the backuppiece (i.e. the name of the file) is unique and/or the file with the same name is not existing on the target.

  Hemant K Collette