We have a database, a rather huge one, with about 10G of
metadata...and almost 400 users...
We have serious performance problems over the past few mnths....and
all sorts of tuning etc...didnt seem to help much..
The System Admins claim that nothing is wrong with the network and its
all the database's fault...Sometimes, the users are just waiting and
waiting till a query returns..
Running 'top' on the system gives this O/P
load averages: 3.42, 3.49, 3.56
291 processes: 287 sleeping, 4 on cpu
CPU states: 59.4% idle, 30.9% user, 9.7% kernel, 0.0% iowait, 0.0%
Memory: 16G real, ...