Could be any/all of the following:

*) optimizer stats due to tables being analyzed differently
*) init parameters
*) hardware (slower/older/fewer CPUs, slower disk)
*) swapping for virtual memory
*) RAID configuration (unlikely for READ-ONLY transactions)
*) segment fragmentation (HWM for tables could be different)

To isolate the cause, I'd check the host systems for differences (i.e. CPUs, RAM, disk RAID config, etc.). If there are no differences, I'd check for differences with loads on the host systems (i.e. number of processes, disk shared with other systems, etc). As more of a last resort, you could clone the DB with good performance to the machine with poor performance, including the init parameters. That'd elliminate DB differences (other than typically defaulted values like cpu_count and asynch_io).

Todd Pepling
[Email address protected]
724-766-3492

________________________________

From: Kelly Z. Yi [mailto:oracledba-ezmlmshield-x83060021.[Email address protected]
Sent: Mon 7/2/2007 14:00
To: LazyDBA Discussion
Subject: what is wrong?



I have an application installed in two different database instances.The
databases both configured the same. But the application run in one
database very fast ( 1 min) and the other takes 25 mins to run.
Does anyone know what is possible reason?



Thanks,

Kelly Z. Yi



Oracle LazyDBA home page