Standby database

[KapilRaj]

Guys,

Help me to get an idea of creating a standby database (9.2.04) . Does the flow goes as follows ,

a).Take a hot backup of production DB
b).Transfer the backup to remote host
c).Recover the remote DB to the current production
d).Do some magic with oracle configs
e) Ready to go !

Is it that simple ?

Regards,

Kaps

[Calandrello]

friend
if you to make backup with the base in begin and offline are not necessary to apply file them in case that he wants to leave identica to another base. but without off doubts to best make the procedure with the base and opçâo and more easy.

[KapilRaj]

I did not understand anything out of your message. Can you please explain a little more ?

Regards,

Kaps

[Calandrello]

Friend The steps that you describe this correct one above. I only commented on bases in begin backup.

[Patti Johnson]

Kaps,

That's pretty close. Check out Metalink note 180031.1 on setting up a dataguard environment for more details.

The 9i version of data guard offers some improvements over the old standby database setup.

A couple highlights
If your filesystems are not identical on the seconday server you can use log/db_file_name_convert parameters in the init.ora and have oracle automatically convert file names for you. Also new files can be automatically added to the secondary database - something not available with standard recovery.

You do need to ensure all changes are logged in the database - set force logging in the primary instance.

Check out the Data Guard documentation.

Patti

[gthai]

My recommendation is to get a hold of some instruction (Metalink is a good start). As you are building it, this forum can help guide you if you encounter any problems.

step a: you can use RMAN or user-managed backup.
Between b and c. alter database create standby controlfile as '/tmp/standby.ctl';
there is more to step d. some of the highlights are: service_names, lock_name_space, log_archive_dest_##,log_archive_dest_state_##, db_file_name_convert, log_file_name_convert, force_logging.
Consider using recover managed standby, dataguard to manage your standby.
NFS is not recommended.
Of course, last but not least, make sure that you have enough Oracle license to run your server. Good luck.

[gthai]

Patti is absolutely right,
Query select FILE#, UNRECOVERABLE_CHANGE#, UNRECOVERABLE_TIME from v$datafile; and make sure that all the changes on the primary are written to logs.

[Kurt Greener]

Your steps are basically what I do. Once you complete those steps, you could apply your archived logs to the recovered database to keep it current. Of course your database has to be running in archive log mode. That also assumes you do not need to recover other parts of the filesystem. Patches to the RDBMS is one reason you may need those files as well.

[Yogeeraj_1]

hi Kaps!

What you are trying to do is neither simple nor complex.

You need to do everything right! I would prefer that you go through the proper documentation and get your steps are clear and precise possible...

http://download-uk.oracle.com/docs/cd/B10501_01/server.920/a96653/toc.htm


hope this helps!

kind regards
yogeeraj

[KapilRaj]

Sorry .. I took so much time to get back to you. Thanks for all the replies. I will get back if I face some errors.

Regds,

Kaps