Oracle without RAC and HACMP Smart Assist

[scoogie73]

We are planning on having two P5 Series AIX machines clustered with HACMP. Our database solution is an Oracle 10g Rel 1 database. What sort of failover does HACMP provide for Oracle? The only information I have found concerns using RAC (pricey). I've also read about Smart Assist for Oracle, but it only seems to be used with Oracle Application Server and not the database. Would it be as simple as creating startup scripts to start the Oracle database on the failover node when/if a failure occurs?

Thanks in advance for advice, help, etc...

 

[RMGBELGIUM]

We are running AIX 5.2 with HACMP and oracle 9.2.0.6 /oracle IAS 10g.We simply created startup and stop scripts doing
1. adjusting the /etc/oratab : we have an oratab for when both nodes are on their own machines, and an oratab when takeover has occured ( all db's must be listed in the oratab for startup then )
2. dbstart and dbshut commands for our databases
3. shutdown and startup of the IAS
4. setting routes,etc ...

rgds,


R.

 

[scoogie73]

The AS10g is what's throwing me. What purpose does the Application Server piece of Oracle provide? It's a bit misleading--it sounds like you can only cluster the Application Server piece of Oracle...I would rather cluster the actual database instance and fail it over in the event the primary node fails.

If HACMP provides "seamless" database instance failover and the only thing that needs to be done is to ensure the dbstart and dbshut commands are called, then this solution is very appealing--more appealing than having to license for 12+ CPUs for a RAC solution on top of the Enterprise Edition licenses for the remaining number of CPUs that we have not yet licensed for.

Thanks for your input! Any other advice you can provide would be great. Is Data Guard used in your organization?

 

[7280]

Hi,
I have two experiences with HACMP/Oracle.
1) Oracle 9.2.0.7 running aix 5.2
2) Oracle 10.2.0.1 running aix 5.3
What HACMP does is to move oracle resources from active node to passive node.
So let's say on node A (active) you have running Oracle and datafiles are in volume group oraVG. When node A crashed HACMP moves oraVG to node B (the passive one) and starts the database and listener.
When it's well configured and tested it does it work fine. My primary nodes crashed a couple of time and I did nothing to bring oracle up on the passive node: hacmp does the work and I had my database up and running without manual intervention.

Hope this helps,
Tarek

 

[khalidaaa]

I have the same configuration as 7280 & RMGBELGIUM except i have 8i LPARs as well!

Regards,
Khalid