Clone From Production To Development
-Make sure preclone scripts already run, if it is not first run preclone on dbTier and appsTier.
-Take the backup of application binaries as well as database backup.
-Mount the backups to the target side, if it is not mounted then do the scp.
-Create blackouts on target side, and send an outage email to business users that this instance is going to clone.
-Shutdown the application services, shutdown the database and startup restrict-mode drop the database, and take backup if required (export-import, pfile, context_file and other config files) sometimes it depends on your project to project)
-On target side untar the binaries of database and prepare the pfile, in pfile keep all the parameters which is there, just add two parameters (db_file_name_convert and log_file_name_convert) which actually convert the filesystem from source to target, and then we do startup nomount (which starts the instance and background processes)
-Now we will run rman duplicate.
-Rman connect auxiliary and run command allocate channels.
-"DUPLICATE" target database to name of the database along with the backup location and then release channels.
what internally rman duplicate will do is?
*it will restore the controlfile from the backup.
*it will mount the database internally (alter database mount).
*it will restore the database.
*it will recover the database.
*and then it will open the database with resetlogs option.
when you open the database with resetlogs, incarnation of the database will be happen, it means
*new logfile sequence number will be assigned and it will be set to 1
*redologfile will be given a new timestamp
*new scn number will be assigned and then your datafile and controlfile will be sync.
-Untar binaries for application side, and then go to common_top/clone/bin and fire perl adcfgclone.pl appsTier.
-Fire perl adcfgclone.pl appsTier context_file=<give cloned instance contextfile>
Post Clone Steps:
* Change the APPS password.
*Cancel the scheduled requests.
*Change the java color and banner.
*Updating workflow mailer status.
*Cleanup the nodes table by running.
EXEC FND_CONC_CLONE.SETUP_CLEAN;.
*Run autoconfig on both the tiers.
*Drop softlinks which are pointing to production
*Do sanity check and release instance for user access.
*Etc.
No comments:
Post a Comment