Database Reference
In-Depth Information
streams_pool_size = 0
db_cache_size=48M
cpu_count = 2
filesystemio_options=setall
parallel_max_servers=0
recyclebin=off
pga_aggregate_target=1G
workarea_size_policy=auto
...
There are some further options you can set that have been omitted here for brevity. Starting the database with
this parameter file allowed the creation of a really small SGA, as you can see from the query below:
SQL> select component,current_size/power(1024,2) mb
2 from v$sga_dynamic_Components
3 where current_size <> 0
4 /
COMPONENT MB
---------------------------------------------------------------- ----------
shared pool 600
large pool 16
DEFAULT buffer cache 48
According to the documentation, the limitation of the
cpu_count
parameter does not influence the benchmark
kit to drive I/O. Running SLOB sessions with an increasing number of sessions, you might notice that the throughput
increases up to a point at which the CPUs are completely saturated and form the bottleneck. When reviewing the AWR
reports, you should pay attention to the I/O latencies and the throughput. The most interesting sections are the
“Load Profile,” “Top 10 Foreground Events by Total Wait Time,” and the “IO profile”:
IO Profile Read+Write/Second Read/Second Write/Second
~~~~~~~~~~ ----------------- --------------- ---------------
Total Requests: 22,138.9 22,132.9 6.0
Database Requests: 22,132.8 22,128.0 4.8
Optimized Requests: 0.0 0.0 0.0
Redo Requests: 0.9 0.0 0.9
Total (MB): 173.0 173.0 0.1
Database (MB): 172.9 172.9 0.0
Optimized Total (MB): 0.0 0.0 0.0
Redo (MB): 0.0 0.0 0.0
Database (blocks): 22,133.8 22,128.3 5.5
Via Buffer Cache (blocks): 22,133.8 22,128.3 5.5
Direct (blocks): 0.0 0.0 0.0
Search WWH ::
Custom Search