This content has been marked as final. Show 3 replies
I don't know of any whitepaper, but I also have not searched for this topic recently.
Still I will try to answer some questions.
* as you have copies of your ASM-Disks on your DR side, you just make these visible to your linux-box and make sure the permissions of the devices are set correct. there is no need (from ASM-side) to fiddle with device-names, but maybe it's easier for admins. In general, you can mimk the setup of your production system without any problems. You can start ASM on your DR side and it should find the disks, you should be able to mount the diskgroups.
* You can start the instances with the same spfile you used on your primary side. (make sure all log-directories etc are really the same)
* I am not sure what you are asking by different name ? You can set any ORACLE_SID you want. If you want to change the DBNAME you can use nid . You can change the DB_UNIQUE_NAME in SPFILE, but this wil have some consequences on new created files in case you are using OFA.
hope this helps,
As promised - I tried it out, it was rather simple once disks were presented.
I needed to use srvctl to add the database, reconfigure tempfiles (we dont mirror those) and then i used nid utility to change both the dbname and dbid and finally make the spfile changes.
All is well. Thanks.
The * alter database open resetlogs; * sqlplus command took a long time to run considering the database was crash consistent and had next to zero update activity to rollback/rollforward.
Not sure why it took so long, perhaps it depends on the physical files of the oneline redo logs (reinit time?). It finished some 10-15 min later so no problems.