RMAN Confusion

From: Ed Hoeffner <hoeff001_at_umn.edu>
Date: Mon, 1 Jun 2015 21:36:41 -0500
Message-ID: <060801d09cdc$f71a0480$e54e0d80$_at_umn.edu>


I'm trying to develop DR plans for an application we run that's built on top of Oracle. According to what I've read, RMAN is the recommendation. After going through the RMAN docs a couple of times, I still feel I don't know enough to actually make a plan to depend on, so I'm hoping someone can put me on the right path without too much bother.  

The app is not very busy by any standards that might come up here. It supports a liquid handling system that mixes a bunch of solutions together and then periodically over the next month or so takes pictures. The DB is used to keep track of all that from entry to exit from the system. The pictures are stored as a filename, so the picture files themselves also have to be backed up. At the moment, the plan is run RMAN incrementals daily from the AT queue to a USB disk. The picture files should be discoverable through some SQL, though I can envision a way to find them with a shell script (ugh!).  

I think I want to use a catalog because the changes are kept in the recovery area. I'm confused between archive-redo logs, control files, the catalog, snapshot files, the RMAN copy command, and what constitutes all the info necessary for the backup to be able to be restored. The manual feels like it assumes I know way more than I do, so I don't feel I know enough to make the decisions here. Would someone please let me know what is really necessary? Is a catalog really the best way to go?  

The manual talks about RMAN not being able to work across platforms, but I can't seem to find the definition of a platform. Since this is on Server 2003, it's going to be wiped and reloaded (probably 2008 - 32 bit), so I'm hoping I can use this to restore the data once that happens. Will that work?  

In trying to make the initial backup in preparation for turning the archive log on, the shutdown immediate command (wouldn't transactional be better?) works great, but startup mount (as told in the manual) fails with error ORA-12505 TNS: Listener could not resolve SID given in connect descriptor, so it takes a reboot to get everything back in line. Somewhere in the myriad readings I ran into this, but I can no longer find the answer. How do I bring it up correctly without having to reboot?  

Is there any way to create something to test with? I only have one shot at this and would really like to get some experience under my belt before I need it.  

Many thanks!!!  

Ed Hoeffner

Received on Tue Jun 02 2015 - 04:36:41 CEST

Original text of this message