does oracle recovery manager work well in HACMP?

[rs6000er]

my production oracle databse works in HACMP environment.the cluster is made of tow nodes aix servers.
the hacmp is running with cascade, that means our production databse running on the primary aix server can be failovered to the standby server.
at the same time, our developing oracle databse also runs in a aix box which is the anther aix server in the cluster.
both the oracle databases use the rman as the backup tool with rman catalog exclusively.
i don't know if the failover happens, how can i backup the production oracle server on the standby node.
so.we want to integrate the rman which is only used to backup the production oracle databse into HACMP.
i wonder what should i do or where i can get some useful manual to do this.
i think maybe the two ways can handle this situation.
1. add the rman as a new resource into HACMP
2. update the the rman which is only used to backups the developing oracle database to backup our current production oracle databse.
i really don't know if the two ways work or not.
if one or both work and how to do it.

 

[DSMARWAY]

Hi ,
if i understand you correctly , you have a 3 node cluster ?

box A runs production dbase box B is standby and Box C is development ?

Do you back your production dbase using rman direct to tape or to disk ?

if to tape do you have a tape library or tapes connected to each server ?

 

[rs6000er]

actually, i have 2 nodes cluster.
box A runs production dbase, box b is both standy and development.
i used the rman in box A backup the production.
i also used another rman in box B backup the development.
both those two rmans send the backup to my tape library and the backup server is Tivoli.

 

[DSMARWAY]

Hi ,
is oracle installed on the local disks on both servers?
or is oracle on the production server installed on the shared disks ?

if production fails over to the standby , do you still require the development databases up and running ?

sorry about the questions , need to understand your setup .

 

[stefanhei]

Hi
In addition to DSMARWAYs questions we'd need to know:
* where does the catalog reside? (is the production catalog reachable after failover?)
* if oracle-binaries are installed on local disks, is oracle-version same or higher than production?
If catalog for production is reachable and you use same or higher version it would theoretically be enough to make the rman-script you wrote to backup the production DB available on the standby-node.
If the script is stored in the catalog nothing has to be done.
If the script is a file you may put it on a shared volume.

Stefan

 

[rs6000er]

thank you for following this topic.
my production oracle resided the local disks at the prodcution machine and when the production nodes failed over to the standby node, we still need the developing accessible.
the rman's catalog resided the local disks on the prodcution machine, if the production machine fails over to the standby, i think i can't get the rman backup the prodction database as normal.
both oracle database have the same version.


so, i think the way i can use is get rid of the rman working with the developing and resore another rman to backup the production databse and leave the developing along for a while.