Oracle Tips by Burleson
Since the focus of this book is using the
Wait Interface and Wait Events to resolve performance problems, it
is important to review the methodology that ADDM uses for
identifying resolutions to problems. Looking at the ADDM
information for tuning allows it to be viewed from several different
Top SQL statements
While there may be times when other
perspectives have merit, the focus will be on the CPU and wait time
approach. Wait information is grouped by wait classes. These
classes are groupings of waits by areas of commonality. For
example, db file scattered read and db file sequential read are both
part of the I/O class.
The goal of the analysis that ADDM performs
is to minimize the metric called DB time. DB time is the total
amount of time spent by the database server on user requests. It
includes wait time and CPU time of all non-idle user sessions. DB
time is displayed in the v$sess_time_model and v$sys_time_model
views containing data for individual sessions and system rollup
This metric is cumulative, that is it
continues to grow as long as the session is connected for
v$sess_time_model; or as long as the instance has been running for
v$sys_time_model. This means that a particular value for DB time is
irrelevant unless used in comparison to an earlier or later value.
ADDM gathers data and issues alerts about
performance issues. The following sections introduce several
components that make use of the data gathered by the AWR.
Data Dictionary Objects
The data dictionary objects will be examined
first. They will be briefly introduced with regard to their
purpose, and then their use will be demonstrated.
The dba_advisor_objects view displays
information about all database objects currently referenced by the
advisors. Each row in this view pertains to a specific advisor
tuning object. Figure 10.1 shows the structure of this view.
The above book excerpt is from:
Oracle Wait Event Tuning
High Performance with Wait
Event Iinterface Analysis