Stop trying to compare the two databases. They are not the same.
Instead try to figure out what is an acceptable time from your database
server for that specific query result to get to your java.
Then examine all the timing differences. i.e. how long from java to
listener - how long for query to execute and so forth.
Once you have determined this and you have a solid idea of what time is
being spent in which component then you can think that perhaps .5 sec is
acceptable for that many physical IOs
"jyou" <member_at_dbforums.com> wrote in message
news:2279957.1039814950_at_dbforums.com...
>
> Thanks,
>
> I set the trace on in sqlplus and redo the query, following is the
> statistics of the query on our production server:
>
> Statistics
> ----------------------------------------------------
> recursive calls
> 54 db block gets
> 2989 consistent gets
> 1596 physical reads
> redo size
> 232 bytes sent via SQL*Net to client
> 316 bytes received via SQL*Net from client
> 2 SQL*Net roundtrips to/from client
> sorts (memory)
> sorts (disk)
> 1 rows processed
>
> While the statistics in my pc is:
> Statistics
> ----------------------------------------------------
> 18 recursive calls
> 57 db block gets
> 3157 consistent gets
> 1 physical reads
> redo size
> 434 bytes sent via SQL*Net to client
> 425 bytes received via SQL*Net from client
> 2 SQL*Net roundtrips to/from client
> 2 sorts (memory)
> sorts (disk)
> 1 rows processed
>
> I noticed the "physical reads" in the production server is pretty big.
> I also checked the "db_block_buffer" is 20,000, db_block_size=8096.
> Since the RAM in that server is 10G, maybe increasing the
> db_block_buffer
> will improve the performance?
>
> --
> Posted via http://dbforums.com
Received on Sat Dec 14 2002 - 08:30:25 CST
