rac workaround question : might be silly

I am new to rac, hence please forgive if this question is really silly =
!!

Lets assume that we have to move a 64 bit db having a huge sga to 32 bit =
linux rac.
Using 32 bit raises the problem of > 4gb sga allocation.=20

Assuming that we do not use the "hugemem" or "VLM shmfs/ramfs" memory =
options=20
on linux RH AS3, and want to use the maximum amount of physical ram on =
one box,=20
is the following RAC work-around practical from a performance =
perspective ?=20

To workaround the limitation of ~4gb sga per instance, what if we create =
multiple=20
instances on the same physical box .=20

Hence if i have a 2 box RAC cluster, and i create 2 instances on each =
box,=20
( my_total_sga for 1 database =3D 2 box * 2 instances * =
max_sga_per_instance )
This effectively can double the amount of total ram i can allocate to my =
database.

Please advise if this assumption is correct and practical.

thanks & regards
ratnesh

GE-Financial Assurance
Lynchburg, VA
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Other related posts: