MySQL Point

Introduction This document provides a brief description of how not to update a Database (duplicate a database) are taken from the backup database using RMAN output tapes of other environments. I. Introduction This document describes the process of updating the databases of test database production, RMAN backups, tapes or discs. II Getting started preparing, we must first ensure that the database does not run in fail-safe environment, and the disc is released using the old database so that we can adapt the new database. Here are some steps that must be taken in the car before the start of the updating of databases: 1 If there are required to provide certain data or accounts (schedule and other necessities) of the old environment , export the data before the start of the first update of the production. 2. If the databases are running in Fail-safe, you end with Fail Safe Manager. Even shutting down the listener and the intelligent agent will be executed for each database. 3. Change TNSNAMES. Ora, the INIT. How often (and receiver. Ora if necessary) files, to ensure that the database can be managed independently from each other, with the local public. Try this out using the telephone service and service database manually via the screen on the 2000 computer services Win, and from databases via SQL * Plus. 4. Continue to delete the database using the Failsafe Manager, all files and database, except those directories init Admin (eg Ora). This is necessary to fit the space on the hard drive clean in the new database. If we have enough space for the restoration is done, then we move the existing files to another directory or mount point. III. Preparing RMAN Duplication script once we've finished with the steps above, we can go to the next step to create scripts to restore the database. An example of this script are given below: connect catalog rman / password @ catalog database <RMAN name>; to connect target sys / password @ <Target database> in conjunction with auxiliary system / password @ <Auxiliary Database> ENV parameter;; assign run (auxiliary channel ch1 type 'sbt_tape' = (c = TDPO_OPTFILE: clustertdpo. opt) '; until scn <SCN no> or all <Date>; set newname for datafile 1 to "new path to restore ";…………… twice the target database <new database aids <name logfile group 1 (< rental of Log-Datei> '(the log-Datei> <location') SIZE 100M, group 2 (( '<Standort the Log-Datei>' ( 'the Log-Datei> <location' ) SIZE 100M, group 3 (( '<location the Log-Datei>' ( 'the <location Log-Datei>') SIZE 100M;) The description of the scenario is as follows: The first part deals connecting to databases required: 1 Catalog database of stored RMAN catalog, 2nd target database, the database that we want to clone, and 3. auxiliary database to establish that we are trying to 4. Please note that later in the execution of this script, both databases and target catalog must be open during the process, while the auxiliary database is generally provided nomount. Next in the script, the allocation of channels to access the file system by the TSM. Please note that this measure is that we must change the configuration of TSM is (opt-dsm.) nodename parameter file to appear on node node production. Next in the script to define SNA / date command which indicates how the database will be duplicated. If the SCR UP / date is not mentioned, RMAN will try not to be the last archived log that failure if the paper is to find the cause of tape drive (for example, it is always hard on the production server). Following is the list of set newname for orders if necessary Datafile Structure the new disc production disk structure is different (although all our systems). All files database should (this list is not specified as necessary) for temporary files. The list of file data can be reached. Finally, the command "Duplicate" there to make the duplication of databases real. dba_data_files by querying the dictionary view data IV Running Duplication Database Database Replication We can prepare to run a script or batch command to start it. She may look like this: Script> RMAN rman cmdfile <die msglog name search for the file <a logs> Before starting the RMAN script, the following things must be maintained: 1 Make sure the catalog database RMAN is open. Make sure that the database will be open during the complete duplication, for example, when it usually falls to disable the backup procedures backup. If the connection to the database will be lost During duplication, the process does not need to rest. 2 Make sure the target database is open. Make sure that the database will be completed in duplicate to open, as RMAN database. Check 3. that Oracle services are candidates for the auxiliary database and the database is in state nomount. If the RMAN script is successful, it will receive all files in the system files, place them in appropriate places, as in the Scriptures, and restore the database. It will also change the database ID, and then restart the database. This is the best, however, If no script duplication, try, where appropriate, relax and failure. V. restore failure if RMAN fails duplication, we found the database using RMAN backup. The database can be restored or supplication, because some of the reasons not to: 1 RMAN catalog database down to the 2nd backup logs archived (not the file system when it was established not SNA) in the specified script. In these cases, you can try to recover the following steps, first run the command switch from RMAN clone () after connecting to the target, catalog and auxiliary databases: run (switch all Datafile clone; creates) and then try the control file. RMAN first creates a control file but does not specify all the data files (as it later). The best way to control the backup file to the database target data track, and modify the script in the auxiliary database. Changes to the script are generally: use as file name, the situation may have changed, set a name database and use clause RESETLOGS. Once the control file is created and executed the complete restoration of the database until SCN specified, the RMAN script, you look like this: Assign run (Auxiliary Channel ENV parameter sbt_tape ch1 type ' '' = (c = TDPO_OPTFILE: clustertdpo. OPT), created scn 6899135273; check recover clone database in read-only output channel CH1;) This step gives you all the necessary archived logs from the file system and turns to the database. After restoration is complete, you can open the database: Alter RESETLOGS database open that recovery would be complete. Please note that if again this way, database ID or the same as for production, so that you can not use RMAN to back up the database (if) they use different catalog. Therefore, one should always try the database is properly thanks to RMAN without having failures. VI. Post steps after upgrading the database is replicated There are some steps that are needed May 1 may be. It may be necessary in some quarters, to be managed NOARCHIVELOG fashion mode production, particularly in ARCHIVELOG mode 2. Add files to temporary tablespaces. If the database is restored, you will all files and tablespaces, but no temporary files are created. We need to add temporary tablespaces tempfiles 3. Remove all connections to database and create them developed appropriate environment. After duplicating new database will have the same links as the basis of production data, then pointed the basis of production data. All database links should be deleted, and created new ones to point to the new environment. 4 tracks, if the new database in which Fail-safe environment, you have the new password (the file on the another node, which was not used) to copy the database. If this does not occur, the database does not start on this node and all fail-safe group will be moved to other nodes. amendments Appeal 5e TNSNAMES. ora (and the listener. ora if necessary) to ensure the database can start in fail-safe. 6 extended to modifying the configuration file TSM () OPT DSM. 7. Close to arrest database for the listeners and the local database and restart the listener and database failures. 8 Make sure that the database can be over-covered properly to another node by fail-group strong hand.