http://otn.oracle.com/oramag/webcolumns/2003/techarticles/scalzo_asm.html
<http://otn.oracle.com/oramag/webcolumns/2003/techarticles/scalzo_asm.html>
This had most of what you need to get a understanding of ASM and how it
compares to other configurations.
I want to keep you all from the 10g dead end that I ran into. Oracle install
docs for Linux would have you think that you must use the ASM drivers to
have any ASM instances to work. To make things worse the drivers are only
usuable on certain kernels of RedHat ES and UL. Well I'm here to tell you
that's nonsense and you don't need it!
I have used DBCA and manually setup both the +ASM (base ASM DB) and 2-n db's
that use the diskgroups served up by the +ASM instance in Suse 8.2 and
RedHat ES 3.0. Using DBCA is a pretty easy way of doing this but it's
obviously good to know what's going on behind the scenes. There is also some
pre-work that needs to be done in RedHat and Suse before the +ASM instance
can use a disk.
The partitions must be setup to be type 83 (Linux who would have guessed).
After this the person who has root must do a chown on /dev/raw/raw* to
oracle:dba so that Oracle can see and use them.
INSTANCE_TYPE=ASM
ASM_DISKGROUPS ='dgroup1'
ASM_DISKSTRING ='/dev/raw/*'
shared_pool_size=64000000
large_pool_size=12m
NOTE: Pay special attention to the ASM_DISKGROUPS setting. The +ASM instance
will only mount diskgroups that are listed in here on startup.
Change your ORACLE_SID environment variable to +ASM and off you go...
sqlplus /nolog
connect / as sysdba
Startup nomount
exit;
INSTANCE_TYPE = RDBMS
DB_CREATE_FILE_DEST = '+dgroup1'
control_files='/u01/apps/oracle/admin/ASMDB/asmdb1.ctl'
db_name=ASMDB
###########################################
###########################################
undo_management=AUTO
undo_retention=900
java_pool_size=100
large_pool_size=83886080
shared_pool_size=83886080
job_queue_processes=1
processes=50
remote_login_passwordfile=none
Change your ORACLE_SID environment variable to ASMDB
Sqlplus /nolog
Startup nomount
As I said you can use DBCA to do this but I like to know how to do this by
hand.