standby in sun solaris machine

[murugappan]

Dear all,

I have two High end SUN solaris servers. I'm planning to implement standby database. Database is now in arch mode.
I have done this for NT. But can any one tell me the know-how of how to map the network in this SUN setup and direct the archives of Production database into Standby database. And other related tips. And how to assign two ip addresses for tnsnames so that when the production db server is down, users can automatically connect to the Standby database server.


Thanx in advance.

[jm]

Originally posted by murugappan


And how to assign two ip addresses for tnsnames so that when the production db server is down, users can automatically connect to the Standby database server.

try this

TEST.WORLD=
(DESCRIPTION_LIST= (DESCRIPTION=
(ADDRESS=(PROTOCOL=TCP)(HOST=85.73.0.250)(PORT=1521))
(CONNECT_DATA=(SID=DBA1)(GLOBAL_NAME=DBA1.WORLD))
)(DESCRIPTION=
(ADDRESS=(PROTOCOL=TCP)(HOST=85.73.0.240)(PORT=1521))
(CONNECT_DATA=(SID=DBA1)(GLOBAL_NAME=DBA1.WORLD))
)
)

Even this works, I am not feel comfortable to let user switch to standby before all "switch" processes done. Normally, I will trun listener off on standby site to stop user login until standby fully available.

[gpsingh]

On Sun Solaris, for 8i now ORacle Standby Data Guard is available , use that for graceul switch over and failover.

[tamilselvan]

The graceful switch over will work only for "SELECT" statement. If a transaction uses INSERT/UPDATE statement on the Primary DB, then it will not be carried over work standby DB when it becomes production. Though the idea of having same service name pointing to primary DB and standby DB, it is of much help ful that every time you need not change the tnsnames.ora file.

[murugappan]

Dear All,

I'm clear with dual-ip that can be set in Oracle8. And my first question have not been answered....?

Can any one say how to map filesystems between two sun solaris servers so that my archives of production get stuffed at standby database directory (filesystem).

How to map the standby server in network and map the appropriate filesystem where arch files are going to be stuffed.


Help me out

Thanx & Regards
murugappan

[Pankajj]

Hi murugappan,

It is easy. If i can do it , anyone can do it

Here are my assumption:
1. You use ssh on Unix boxes
2. You are aware of rsync and scp
3. PROD - Production Box . STBY - Standby Box.
4. Setting up cronjob on unix.

Step1: Setup PROD DB & STBY DB the way u want.
Step2: As u know for setting up STBY , u will need to apply the redo logs generated by Arch process time by time. For that u will need to setup cronjob on PROD box which copy all arch log to STBY box on fixed frequency. To created trusted relationship kind of thing between to unix boxes u will need to copy the public key of oracle user to STBY box in folder .ssh folder as authorized_keys file name. Once u do that it wouldn't ask you for password.
Step3: Setup cronjob on PROD box to rsync ( Sync the arch folder on both machine , that way cronjob will copy only the file which are newly generate ( missing on Standby )
Step4: Setup cronjob on Standby to apply those redo logs.

Another Solution : If u don't want to use , you can setup NFS , and allow oracle user to write on it , and let STANDBY & PROD box to share it. That way you will just need to copy the ARCH logs to NFS folder time by time.

Regards !

Pankaj

PS: Let me know if u want detailed scripts for it !