I/o problem with tablespace

First , please excuse my writing; i am a french Canadian and i could have made some english errors .

Problem:
 We have an Oracle instance that is composed of these tablespaces on 4 "DSSI rf73" physical disks on a Vax/VMS system :  system tablespace on disk 2
 application_data tablespace on disk 1 ( 50%25 I/O )  application_index tablespace on disk 2  temporary_tablespace on disk 3
 snapshot_log_tablespace on disk 2 ( 30%25 I/O )  rollback tablespace tripped on disk 3, 2 and 1  10 online redo logs on disk 3
 archive redo log on disk 4

 We figured out that we had a disk I/O performance problem because we have only 3 disk on witch we could setup tablespaces and even if we have tune up the instance to minimize disk I/O, we still have a performance problem; disk I/O still seems to be the problem.

 So, we've decided to install a new disk extension that would be setup as a raid0 logical disk; this raid0 is composed of a ram cache, and tripped on 3  physical disks.

 Question :

 What tablespaces should i transfert to this new faster disk ( alter tablespace rename datafile ) so that i reach a maximum disk I/O performance for all the instance ?

 Should i transfert all tablespaces ( because disk I/O is much better on the new disk ), or should i keep index an data, or most access tablespaces ( data and snapshot ) separate in between the old and new disks, even if these old disks ara much slower ?

 What about rollback tablespace?

 If i keep some tablespaces on the old disk ( these are much slower disks ), would'nt they slower down all the instance operations ? Witch tablespaces  should i keep on the old disks?

 Please , i request and answer as soon as possible.

-----------== Posted via Deja News, The Discussion Network ==---------- http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own Received on Tue Dec 08 1998 - 14:25:56 CST