Temp tspace is one of the few places where you will concurrent *write* activity (database files all being done with one or just a few db writers) so maybe the read time is suffering from concurrent write activity ? You could validate this to some degree by having lots of temporary tablespaces (and hence of temp files) and round-robin-ing them among users to spread the load. hth connor On 6/9/05, Peter Dixon <peterdixon001@xxxxxxxxxxx> wrote: > I have performed a healthcheck on our production database , I am running > 8.1.7.4 on hp11.11 with the database files running on a san (raid 5). I w= as > surprised to see the average read time for the temporary tablespace as 82= 6ms > where all the other tablespaces in the database are around 10ms. Would it= be > a wise to move the temporary tablespace data files onto another sort of d= isk > storage? >=20 > Peter Dixon > 8i OCP >=20 >=20 > -- > //www.freelists.org/webpage/oracle-l >=20 --=20 Connor McDonald =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D email: connor_mcdonald@xxxxxxxxx web: http://www.oracledba.co.uk "Semper in excremento, sole profundum qui variat" -- //www.freelists.org/webpage/oracle-l