Migration R12.1.3 to new server

[satish]

Hi erman,

We are planning to move our environment to new server.new machine with different hostname and ip adress

Source

R12.1.3
Db 11.2.0.4
Rhel 5.5

Target

R12.1.3
11.2.0.4
Rhel 7.9---changed

We are planning to tar the oracle home and copy to target server.
Can we use rman to copy the data to new server?
If we use rman,do we need to go for reset logs?


Dataloss should be zero

Can you provide high level steps with less less downtime involved.
We are not having dataguard or golden gate.Our database size is 300GB

Thank you,
Satish

[ErmanArslansOracleBlog]

You can do a RMAN duplicate to target.
Then clear the db config and run autoconfig (after building xml for the new db environment of course)
you can copy oracle home, that's okay. But if you change its patch, you got to relink it.(-relink all- actually)
For the apps tier, you got to do a postclone after making db up& running with a proper config.. That's all.

[satish]

Hi erman,

Is reset logs mandatory?We are following below steps.

1)tar the db oracle home in source and copy to target

2)untar the oracle home in target and run adcfgclone dbtechstack

3)shutdown the source and take in Mount state.Take rman cold backup and just copy the backup to new server

4)once step 2 completed(dbtechstack) we will start the instance in nomount and then issue below commands to restore and recover the database from backup

a)restore database

b)recover database(it's not required since using cold backup)

c)Since it is prod,will resetlogs have any impact?

Then post steps follows...like fnd_nodes cleanup,autoconfig etc..

Please correct me if i am wrong

[ErmanArslansOracleBlog]

Just copy your oracle home from PROD to CLONE. Relink it on CLONE. if -- if necessary. if you directory structure changes..
Just use RMAN duplicate to clone your database from PROD to CLONE. RMAN will open the database in CLONE with resetlogs, but this resetlogs operation will be on newly created clone (namely CLONE) and it will not affect your production (namely PROD)

There are different duplicate options (duplicate from backup, from your active db and so on..)
See RMAN Documentation for this.

Here is document if you want to do a backup based RMAN duplicate -> Perform Backup Based RMAN DUPLICATE Without Connecting To Target Database For Both Disk & Tape Backups (Doc ID 1375864.1)

[satish]

Hi erman,

We are migrating to new hardware.Can we still follow the same approach and open resetlogs in prod

Thanks,
Satish

[ErmanArslansOracleBlog]

RESETLOGS is required after incomplete recovery.
It depends on the migration method, and the capabilities you have.

For instance; RMAN Duplicate uses OPEN RESETLOGS, as it is the way it works;

When running an RMAN duplicate in Oracle 11.2 and higher, RMAN will perform the following steps:

1. It restores a controlfile from the target into the auxiliary
2. It resets the db_name to the target so the controlfile can be mounted.
3. It restores the datafiles setting the auxiliary datafiles in the proper location using either db_file_name_convert or 'set newname', if specified.
4. It recovers the auxiliary datafiles to the 'until' clause specified or derived.
5. It recreates the controlfile with the system datafile (datafile #1)
6. It catalogs the auxiliary datafiles into the newly created datafiles.
7. Executes a 'switch datafile to copy' of all the datafiles.
8. It changes the DBID and opens the database with resetlogs.

See?

For RMAN , we have following document for instance as well ->

How to clone the database using RMAN without new incarnation (with NORESETLOGS) (Doc ID 1608236.1)

Basically, if you do an incomplete recovery , then you have to open your database with resetlogs option.
The incomplete recovery is -> restore a backup, apply some redo (not whole) and recover the database according to the redo that you apply, to a specific time or SCN.

At the recovery process, we should apply the online redo logs when the time comes, but if we didn't apply them, or if the recovery process applies them as archived redo logs (when we use backup controlfile, recovery process treats redologs as archived redologs), then we need to reset our redologs and that's why we have to use resetlogs option.

By using a backup controlfile, you actually saying, my control file is a backup controlfile:)
In the following 2 conditions , you use it;

1)If the controlfile is older than any of your datafiles.
2)You do not have a valid current control file to use.

And those mechanisms like RMAN duplicate use backup controlfile for one of these reasons.

[ErmanArslansOracleBlog]

And for your specific question. my answer is yes.