Wolfgang, Sorry, that was just my bias. After getting through the 8.1.7 to 9.2 migrations, I didn't want to alter the environment until the 9.2 to 10.1 migrations. that's when I started gathering system statistics. Paul ----------------------------------------------------------------------------------------------------------------------------- "Footnote 2: 5th April 2005 References to penguins in the above article should not be interpreted as being related in any way to Linux." ----------------------------------------------------------------------------------------------------------------------------- On Apr 5, 2005 4:10 PM, Wolfgang Breitling <breitliw@xxxxxxxxxxxxx> wrote: > Paul, > > why only after migrating to 10g? The CBO in Oracle 9i with system stats > ought to do the same - derive the cost of a full scan from the MBRC > system stats value rather than from the dfmrc init parameter - or do you > have evidence to the contrary? > > Paul Drake wrote: > > > > > After you migrate to 10g (and gather system stats), you may be able to > > get away with a larger db_file_multiblock_read_count (without the CBO > > wanting to tablescan/hash_join everything) in which case a larger > > stripe size will be better. > > > -- > Regards > > Wolfgang Breitling > Centrex Consulting Corporation > www.centrexcc.com > -- #/etc/init.d/init.cssd stop # f=ma, divide by 1, convert to moles. -- //www.freelists.org/webpage/oracle-l