kouchba_at_yahoo.com (kouchba) wrote in message news:<f55a8ec8.0401142353.4ced6d1e_at_posting.google.com>...
> First I would like to thank everyone for your help. Here is some more
> info.
> The disk on which everything is stored is a netapp F810 filesystem. Is
> ia a NAS (Network accessed Storage) disk accessed through NFS and
> recommanded by Oracle.
> Both Instances 7.3.4 and 8.1.7 run on a quadri pro on Solaris 7 and
> following are the initXXX.ora files. No object has ever been analyzed
> in both instances so I changed Oracle817 optimizer from choose to rule
> without any success.
>
> The tkprof output looks like this and is not linked to any particular
> query :
> SELECT * FROM TEST (which produces the same execution plan on both
> machines ) I also checked that TEST is not fragmented on both extents
> (one single extent)
>
> on Oracle 7.3.4
> elapsed disk rows
> 100.10 65000 200000
>
> on Oracle 8.1.7
> elapsed disk rows
> 153.70 43000 180000
>
> which means it takes less time on Oracle 7.3.4 to make a full scan
> which takes more IO than on Oracle 8.1.7..! That problem is
> essentially visible on reads as don't make that many disk
> writes(INSERT/UPDATES). So I suspect that may be Oracle 7.3.4 can read
> asynchronously whereas Oracle 8.1.7 does not..? May be I am wrong but
> can you please help ?
>
>
My gut feeling tells the problem is you installed Netapp out of the
box and you don't know which parameters to set to make sure Oracle
will access it asynchronously. Your suggestion Oracle 8.1.7 can not
read asynchronously is just utter bs.
Not an Oracle problem, consequently.
Sybrand Bakker
Senior Oracle DBA
Received on Thu Jan 15 2004 - 05:02:08 CST
