How
Oracle Manages Diagnostic Traces and Statistics?
Oracle
traces various types of diagnostic and statistical information useful
to the users and support groups.
01.
For the Database Oracle maintains Alert Log File with an ability to the
user to set which events are to be traced and which are not to. To identify
critical errors and tracing the errors the user is enabled to define destinations
for depositing the trace files by Oracle. They are generally called dump
destinations.
Background
dump destination
User dump destination
Core dump destination
02.
Network diagnostic trace files are generated at the default destinations
under ORACLE_HOME/network/log
Setting additional events to track the errors and diagnose the reasons
user has the ability to set events at session, system and database level.
Depending upon the nature of the event and the level of trace some events
can be set at the session level, some at system level and some at database
level.
The
Initialization Parameters that impact the Oracle trace generation
(1)
ORACLE_TRACE_COLLECTION_SIZE
ORACLE_TRACE_COLLECTION_SIZE
specifies (in bytes) the maximum size of the Oracle Trace collection file
(.dat). Once the collection file reaches this maximum, the collection
is disabled. A value of 0 means that the file has no size limit.
(2)
ORACLE_TRACE_COLLECTION_PATH
ORACLE_TRACE_COLLECTION_PATH
specifies the directory pathname where the Oracle Trace collection definition
(.cdf) and data collection (.dat) files are located. If you accept the
default, the Oracle Trace .cdf and .dat files will be located in ORACLE_HOME/otrace/admin/cdf
on UNIX Platforms and ORACLE_HOME\otrace\admin\cdf ON Windows platforms
(3)
TRACEFILE_IDENTIFIER
TRACEFILE_IDENTIFIER
specifies a custom identifier that becomes part of the Oracle Trace file
name. Such a custom identifier is used to identify a trace file simply
from its name and without having to open it or view its contents.
Each
time this parameter is dynamically modified, the next trace dump will
be written to a trace file which has the new parameter value embedded
in its name. Trace file continuity information is automatically added
to both the old and new trace files to indicate that these trace files
belong to the same process.
This
parameter can only be used to change the name of the foreground process'
trace file; the background processes continue to have their trace files
named in the regular format. For foreground processes, the TRACEID column
of the V$PROCESS view contains the current value of the TRACEFILE_IDENTIFIER
parameter. When this parameter value is set, the trace file name has the
following format:
sid_ora_pid_traceid.trc
In this example, sid is the oracle instance ID, pid is the process ID,
and traceid is the value of the TRACEFILE_IDENTIFIER parameter.
(4)
TIMED_OS_STATISTICS
TIMED_OS_STATISTICS
specifies the interval (in seconds) at which Oracle collects operating
system statistics when a request is made from the client to the server
or when a request completes.
On
dedicated servers, Oracle collects operating system statistics at user
logon and after each subsequent client invocation through the OCI into
the Oracle server as a remote procedure call message.
On
shared servers, Oracle collects statistics when client calls to Oracle
are processed.
A
value of zero specifies that operating system statistics are not gathered.
To collect statistics, set a value meaningful for your application and
site needs.
Gathering
operating system statistics is very expensive. Oracle Corporation recommends
that you set this parameter in an ALTER SYSTEM statement rather than in
the initialization parameter file, and that you reset the value to zero
as soon as the needed statistics have been gathered.
(5)
TRACE_ENABLED
TRACE_ENABLED
controls tracing of the execution history, or code path, of Oracle. Oracle
Support Services uses this information for debugging.
When
TRACE_ENABLED is set to true, Oracle records information in specific files
when errors occur. See Oracle9i Real Application Clusters Deployment and
Performance for the types of files and the default destination directories
in which Oracle records the execution history.
Oracle
records this information for all instances, even if only one instance
terminates. This allows Oracle to retain diagnostics for the entire cluster.
Although
the overhead incurred from this processing is not excessive, you can improve
performance by setting TRACE_ENABLED to false. You might do this, for
example, to meet high-end benchmark requirements. However, if you leave
this parameter set to false, you may lose valuable diagnostic information.
Therefore, always set TRACE_ENABLED to true to trace system problems and
to reduce diagnostic efforts in the event of unexplained instance failures.
(6)
CORE_DUMP_DEST
This
parameter was made dynamic via ALTER SYSTEM in Oracle 8.1.
CORE_DUMP_DEST
should point to the directory where core dumps from the Oracle server
will be placed. The above "Related" parameters determine the
size of the core dumps. A core dump is a memory image of the Oracle shadow
process produced when an unexpected , unrecoverable or invalid condition
occurs.
Note that Oracle should always try to write a trace file before producing
a core dump. Always check <Parameter:USER_DUMP_DEST> and <Parameter:BACKGROUND_DUMP_DEST>
locations for tracefiles first.
In
some cases this parameter may not have been read when a core file is produced.
In this case the core dump is typically written to the default location
($ORACLE_HOME/dbs).
CORE_DUMP_DEST
should have WRITE permission for the Oracle user.
(7)
USER_DUMP_DEST
USER_DUMP_DEST
specifies the pathname for a directory where the server will write debugging
trace files on behalf of a user process.
(8)
BACKGROUND_DUMP_DEST
BACKGROUND_DUMP_DEST
specifies the pathname (directory or disc) where debugging trace files
for the background processes (LGWR, DBWn, and so on) are written during
Oracle operations.
An
alert file in the directory specified by BACKGROUND_DUMP_DEST logs significant
database events and messages. Anything that affects the database instance
or global database is recorded here. The alert file is a normal text file.
Its filename is operating system-dependent. For platforms that support
multiple instances, it takes the form alert_sid.log, where sid is the
system identifier. This file grows slowly, but without limit, so you might
want to delete it periodically. You can delete the file even when the
database is running.
(9) DB_BLOCK_CHECKSUM
DB_BLOCK_CHECKSUM
determines whether DBWn and the direct loader will calculate a checksum
(a number calculated from all the bytes stored in the block) and store
it in the cache header of every data block when writing it to disk. Checksums
are verified when a block is read only if this parameter is true and the
last write of the block stored a checksum. In addition, Oracle gives every
log block a checksum before writing it to the current log.
If
this parameter is set to false, DBWn calculates checksums only for the
SYSTEM tablespace, but not for user tablespaces.
Checksums
allow Oracle to detect corruption caused by underlying disks, storage
systems, or I/O systems. Turning on this feature typically causes only
an additional 1% to 2% overhead. Therefore, Oracle Corporation recommends
that you set DB_BLOCK_CHECKSUM to true
(10)
DB_BLOCK_CHECKING
DB_BLOCK_CHECKING
controls whether Oracle performs block checking for data blocks. When
this parameter is set to true, Oracle performs block checking for all
data blocks. When it is set to false, Oracle does not perform block checking
for blocks in the user tablespaces. However, block checking for the SYSTEM
tablespace is always turned on.
Oracle
checks a block by going through the data on the block, making sure it
is self-consistent. Block checking can often prevent memory and data corruption.
Block checking typically causes 1% to 10% overhead, depending on workload.
The more updates or inserts in a workload, the more expensive it is to
turn on block checking. You should set DB_BLOCK_CHECKING to true if the
performance overhead is acceptable.
(11) EVENT
EVENT is a parameter used only to debug the system. Do not alter the value
of this parameter except under the supervision of Oracle Corporation Worldwide
Support staff.
There
are events that can be set at session level, system level and Database
level.
To
set at the database level set the event in the init.ora file
To set at system level which is not persistent use ‘alter system’
command
To set at a session level which is limited to that session use ‘alter
session’.
(12) ORACLE_TRACE_COLLECTION_NAME
A
collection is data collected for events that occurred while an instrumented
product was running. ORACLE_TRACE_COLLECTION_NAME specifies the Oracle
Trace collection name for this instance. Oracle also uses this parameter
in the output file names (collection definition file .cdf and data collection
file .dat). If you set ORACLE_TRACE_ENABLE to true, setting this value
to a non-null string will start a default Oracle Trace collection that
will run until this value is set to null again. Valid collection name
up to 16 characters long (except for platforms that enforce 8-character
file names)
(13)
ORACLE_TRACE_FACILITY_PATH
ORACLE_TRACE_FACILITY_PATH
specifies the directory pathname where Oracle Trace facility definition
files are located. On Solaris, the default path is ORACLE_HOME/otrace/admin/fdf/.
On Windows, the default path is ORACLE_HOME\otrace\admin\fdf
(14)
ORACLE_TRACE_FACILITY_NAME
ORACLE_TRACE_FACILITY_NAME
specifies the event set that Oracle Trace collects. The value of this
parameter, followed by the .fdf extension, is the name of the Oracle Trace
product definition file. That file must be located in the directory specified
by the ORACLE_TRACE_FACILITY_PATH parameter. The product definition file
contains definition information for all the events and data items that
can be collected for products that use the Oracle Trace data collection
API.
The
Oracle server has multiple event sets and therefore multiple product definition
files:
ORACLE
is the ALL event set
ORACLED is the DEFAULT event set
ORACLEE is the EXPERT event set
ORACLESM is the SUMMARY event set
ORACLEC is the CACHEIO event set
Oracle Corporation recommends that you use the DEFAULT event set for Oracle
server collections, ORACLED.
(15) ORACLE_TRACE_ENABLE
To enable Oracle Trace collections for the server, set ORACLE_TRACE_ENABLE
to true. This setting alone does not start an Oracle Trace collection,
but it allows Oracle Trace to be used for the server.
With
ORACLE_TRACE_ENABLE set to true, you can perform Oracle Trace collection
of server event data in any of the following ways:
By
using Oracle Trace Manager, which is supplied with the Oracle Diagnostic
Pack
By
using the Oracle Trace command line interface, which is supplied with
the server
By
specifying a collection name in the ORACLE_TRACE_COLLECTION_NAME parameter
Tracing the execution paths for Performance Tuning
Explain
Plan
Usage:
Explain
plan for select * from emp;
Explain
Plan is a great way to tune your queries.
As a bonus for using Explain Plan, you will learn more about how the DBMS
works "behind the scenes", enabling you to write efficient queries
the first time around.
Explain
Plan executes your query and records the "plan" that Oracle
devises to execute your query. By examining this plan, you can find out
if Oracle is picking the right indexes and joining your tables in the
most efficient manner. There are a few different ways to utilize Explain
Plan.
The
first thing you will need to do is make sure you have a table called plan_table
available in your schema.
If
this table is not there run this script provided by oracle to create that
table
ORACLE_HOME/rdbms/admin/utlxplan.sql .. for UNIX plat formas and
ORACLE_HOME\rdbms\admin\utlxplan.sql .. for WINDOWS platforms
When
the plan table is populated then use this query to fetch the data formatted.
So thatyou understand how oracle is planning data access paths
select
*
from table(dbms_xplan.display);
Ex:
on windows 2000 Oracle 10.1
SQL>
@d:\oracle\ora10\rdbms\admin\utlxplan.sql
Table
created.
SQL>
explain plan for select * from emp
2 ;
Explained.
SQL> select *
2 from table(dbms_xplan.display);
PLAN_TABLE_OUTPUT
-------------------------------------------------------------------------------------------------
Plan hash value: 4080710170
-------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 14 | 518 | 3 (0) | 00:00:01 |
| 1 | TABLE ACCESS FULL | EMP | 14 | 518 | 3 (0) | 00:00:01 |
--------------------------------------------------------------------------------------------------
8
rows selected.
Transient
Kernel Profiling (TKProf)
From
Oracle Docs
The
SQL Trace Facility and TKPROF
The SQL trace facility and TKPROF are two basic performance diagnostic
tools which can help you monitor and tune applications that run against
the Oracle Server. This chapter covers:
Note: The output of the EXPLAIN PLAN command reflects the behavior of
the Oracle optimizer. Because the optimizer is likely to evolve between
releases of the Oracle Server, output from the EXPLAIN PLAN command will
also evolve. Similarly, the SQL trace facility and TKPROF program are
also subject to change in future releases of the Oracle Server. Such changes
will be documented in future versions of Oracle manuals.
Introduction
The
SQL trace facility and TKPROF enable you to accurately assess the efficiency
of the SQL statements your application runs. For best results, use these
tools together rather than using EXPLAIN PLAN alone. This section covers:
About
the SQL Trace Facility
About TKPROF
How to Use the SQL Trace Facility and TKPROF
About the SQL Trace Facility
The SQL trace facility provides performance information on individual
SQL statements. It generates the following statistics for each statement:
parse,
execute, and fetch counts
CPU and elapsed times
physical reads and logical reads
number of rows processed
misses on the library cache
username under which each parse occurred
each commit and rollback
You can enable the SQL trace facility for a session or for an instance.
When the SQL trace facility is enabled, performance statistics for all
SQL statements executed in a user session or in an instance are placed
into a trace file.
The
additional overhead of running the SQL trace facility against an application
with performance problems is normally insignificant, compared with the
inherent overhead caused by the application's inefficiency.
About
TKPROF
You can run the TKPROF program to format the contents of the trace file
and place the output into a readable output file. Optionally, TKPROF can
also determine the execution plans of SQL statements create a SQL script
that stores the statistics in the database
Since
TKPROF reports each statement executed with the resources, which it has
consumed, the number of times it was called, and the number of rows, which
it processed, you can easily locate those statements, which are using
the greatest resource. With experience or with baselines available, you
can assess whether the resources used are reasonable given the work done.
How
to Use the SQL Trace Facility and TKPROF
Follow
these steps to use the SQL trace facility and TKPROF.
Set
initialization parameters for trace file management.
Enable
the SQL trace facility for the desired session and run your application.
This step produces a trace file containing statistics for the SQL statements
issued by the application.
Run TKPROF to translate the trace file created ins step 2 into a readable
output file. This step can optionally create a SQL script that stores
the statistics in the database.
Interpret the output file created in step 3.
Optionally, run the SQL script produced in step 3 to store the statistics
in the database.
In the following sections each of these steps is discussed in depth.
Step
1:
Setting
Initialization Parameters for Trace File Management
When
the SQL trace facility is enabled for a session, Oracle generates a trace
file containing statistics for traced SQL statements for that session.
When the SQL trace facility is enabled for an instance, Oracle creates
a separate trace file for each process.
Before
enabling the SQL trace facility, you should:
Check
settings of the TIMED_STATISTICS, USER_DUMP_DEST, and MAX_DUMP_FILE_SIZE
parameters.
Parameter
TIMED_STATISTICS
Notes
This parameter enables and disables the collection of timed statistics,
such as CPU and elapsed times, by the SQL trace facility, as well as the
collection of certain statistics in the dynamic performance tables. The
default value of FALSE disables timing. A value of TRUE enables timing.
Enabling timing causes extra timing calls for low-level operations. This
is a dynamic parameter.
Parameter
MAX_DUMP_FILE_SIZE
Notes
When the SQL trace facility is enabled at the instance level, every call
to the server produces a text line in an operating system file. The maximum
size of these files (in operating system blocks) is limited by the initialization
parameter MAX_DUMP_FILE_SIZE. The default is 500. If you find that your
trace output is truncated, increase the value of this parameter before
generating another trace file.
Parameter
USER_DUMP_DEST
Notes
This parameter must fully specify the destination for the trace file according
to the conventions of your operating system. The default value for this
parameter is the default destination for system dumps on your operating
system.This value can be modified with ALTER SYSTEM SET USER_DUMP_DEST=newdir.
Devise
a way of recognizing the resulting trace file.
Be
sure you know how to distinguish the trace files by name. Oracle writes
them to the user dump destination specified by USER_DUMP_DEST. However,
this directory may soon contain many hundreds of files, usually with generated
names. It may be difficult to match trace files back to the session or
process which created them. You can tag trace files by including in your
programs a statement like SELECT `program name' FROM DUAL. You can then
trace each file back to the process that created it.
If
your operating system retains multiple versions of files, be sure your
version limit is high enough to accommodate the number of trace files
you expect the SQL trace facility to generate.
The
generated trace files may be owned by an operating system user other than
yourself. This user must make the trace files available to you before
you can use TKPROF to format them.
Step
2:
Enabling
the SQL Trace Facility
You can enable the SQL trace facility for a session or for the instance.
This section covers:
Enabling
the SQL Trace Facility for Your Current Session
Enabling the SQL Trace Facility for a Different User Session
Enabling the SQL Trace Facility for an Instance
Attention: Because running the SQL trace facility increases system overhead,
you should enable it only when tuning your SQL statements, and disable
it when you are finished.
Enabling
the SQL Trace Facility for Your Current Session
To enable the SQL trace facility for your current session, enter:
ALTER
SESSION SET SQL_TRACE = TRUE;
Alternatively,
you can enable the SQL trace facility for your session by using the DBMS_SESSION.SET_SQL_TRACE
procedure.
To
disable the SQL trace facility for your session, enter:
ALTER
SESSION SET SQL_TRACE = FALSE;
The
SQL trace facility is automatically disabled for your session when your
application disconnects from Oracle.
Note:
You
may need to modify your application to contain the ALTER SESSION command.
For example, to issue the ALTER SESSION command in Oracle Forms, invoke
Oracle Forms using the -s option, or invoke Oracle Forms (Design) using
the statistics option. For more information on Oracle Forms, see the Oracle
Forms Reference manual.
Enabling
the SQL Trace Facility for a Different User Session
To enable the SQL trace facility for a session other than your current
session, you can call the procedure DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION.
This can be useful for database administrators who are not located near
their users, or who do not have access to the application code to set
SQL trace from within an application.
This
procedure requires the session ID and serial number of the user session
in question, which you can obtain from the V$SESSION view. In the WHERE
clause you can specify sessions by referencing the value of the OSUSER,
USERNAME, or PROGRAM column in V$SESSION. For example, the following Server
Manager line mode session obtains the session ID and serial number for
the operating system user jausten and then enables SQL trace for that
user's session:
SVRMGR>
SELECT sid, serial#, osuser
2> FROM v$session
3> WHERE osuser = 'jausten';
SID
SERIAL# OSUSER
---------- ---------- ---------------
8 12 jausten
1 row selected.
SVRMGR>
EXECUTE dbms_system.set_sql_trace_in_session(8,12,TRUE);
Statement processed.
To
enable SQL trace in stored procedures, use this SQL statement:
DBMS_SESSION.SET_SQL_TRACE
(TRUE);
Enabling
the SQL Trace Facility for an Instance
To enable the SQL trace facility for your instance, set the value of the
SQL_TRACE initialization parameter to TRUE. Statistics will be collected
for all sessions.
Once
the SQL trace facility has been enabled for the instance, you can disable
it for an individual session by entering:
ALTER
SESSION SET SQL_TRACE = FALSE;
Step
3:
Formatting
Trace Files with TKPROF
This section covers:
Sample
TKPROF Output
Syntax of TKPROF
TKPROF Statement Examples
TKPROF accepts as input a trace file produced by the SQL trace facility
and produces a formatted output file. TKPROF can also be used to generate
execution plans.
Once
the SQL trace facility has generated a number of trace files, you can:
run
TKPROF on each individual trace file, producing a number of formatted
output files, one for each session
concatenate the trace files and then run TKPROF on the result to produce
a formatted output file for the entire instance
TKPROF does not report COMMITs and ROLLBACKs that are recorded in the
trace file.
Sample TKPROF Output
Sample
output from TKPROF is as follows:
SELECT
* FROM emp, dept WHERE emp.deptno = dept.deptno;
call count cpu elapsed disk query current rows
---- ------- ------- --------- -------- -------- ------- ------
Parse 1 0.16 0.29 3 13 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.03 0.26 2 2 4 14
Misses in library cache during parse: 1
Parsing user id: (8) SCOTT
Rows
Execution Plan
------- ---------------------------------------------------
14 MERGE JOIN
4 SORT JOIN
4 TABLE ACCESS (FULL) OF 'DEPT'
14 SORT JOIN
14 TABLE ACCESS (FULL) OF 'EMP'
For
this statement, TKPROF output includes the following information:
the
text of the SQL statement
the SQL trace statistics in tabular form
the number of library cache misses for the parsing and execution of the
statement
the user initially parsing the statement
the execution plan generated by EXPLAIN PLAN
TKPROF also provides a summary of user level statements and recursive
SQL calls for the trace file.
Syntax
of TKPROF
Invoke TKPROF using this syntax:
If
you invoke TKPROF with no arguments, online help is displayed.
Use
the following arguments with TKPROF:
TKPROF
Arguments
Argument
Meaning
filename1 Specifies the input file, a trace file containing statistics
produced by the SQL trace facility. This file can be either a trace file
produced for a single session or a file produced by concatenating individual
trace files from multiple sessions.
filename2 Specifies the file to which TKPROF writes its formatted output.
AGGREGATE If you specify AGGREGATE = NO, then TKPROF does not aggregate
multiple users of the same SQL text.
EXPLAIN Determines the execution plan for each SQL statement in the trace
file and writes these execution plans to the output file. TKPROF determines
execution plans by issuing the EXPLAIN PLAN command after connecting to
Oracle with the user and password specified in this parameter. The specified
user must have CREATE SESSION system privileges. TKPROF will take longer
to process a large trace file if the EXPLAIN option is used.
TABLE Specifies the schema and name of the table into which TKPROF temporarily
places execution plans before writing them to the output file. If the
specified table already exists, TKPROF deletes all rows in the table,
uses it for the EXPLAIN PLAN command (which writes more rows into the
table), and then deletes those rows. If this table does not exist, TKPROF
creates it, uses it, and then drops it.
The
specified user must be able to issue INSERT, SELECT, and DELETE statements
against the table. If the table does not already exist, the user must
also be able to issue CREATE TABLE and DROP TABLE statements. For the
privileges to issue these statements, see the Oracle8 Server SQL Reference.
This
option allows multiple individuals to run TKPROF concurrently with the
same user in the EXPLAIN value. These individuals can specify different
TABLE values and avoid destructively interfering with each other's processing
on the temporary plan table.
If you use the EXPLAIN parameter without the TABLE parameter, TKPROF uses
the table PROF$PLAN_TABLE in the schema of the user specified by the EXPLAIN
parameter. If you use the TABLE parameter without the EXPLAIN parameter,
TKPROF ignores the TABLE parameter.
INSERT Creates a SQL script that stores the trace file statistics in the
database. TKPROF creates this script with the name filename3. This script
creates a table and inserts a row of statistics for each traced SQL statement
into the table.
SYS Enables and disables the listing of SQL statements issued by the user
SYS, or recursive SQL statements into the output file. The default value
of YES causes TKPROF to list these statements. The value of NO causes
TKPROF to omit them. Note that this parameter does not affect the optional
SQL script. The SQL script always inserts statistics for all traced SQL
statements, including recursive SQL statements.
SORT Sorts the traced SQL statements in descending order of the specified
sort option before listing them into the output file. If more than one
option is specified, the output is sorted in descending order by the sum
of the values specified in the sort options. If you omit this parameter,
TKPROF lists statements into the output file in order of first use.
The
sort options are as follows:
PRSCNT number of times parsed
PRSCPU CPU time spent parsing
PRSELA elapsed time spent parsing
PRSDSK number of physical reads from disk during parse
PRSMIS number of consistent mode block reads during parse
PRSCU number of current mode block reads during parse
PRSMIS number of library cache misses during parse
EXECNT number of executes
EXECPU CPU time spent executing
EXEELA elapsed time spent executing
EXEDSK number of physical reads from disk during execute
EXEQRY number of consistent mode block reads during execute
EXECU number of current mode block reads during execute
EXEROW number of rows processed during execute
EXEMIS number of library cache misses during execute
FCHCNT number of fetches
FCHCPU CPU time spent fetching
FCHELA elapsed time spent fetching
FCHDSK number of physical reads from disk during fetch
FCHQRY number of consistent mode block reads during fetch
FCHCU number of current mode block reads during fetch
FCHROW number of rows fetched
PRINT Lists only the first integer sorted SQL statements into the output
file. If you omit this parameter, TKPROF lists all traced SQL statements.
Note that this parameter does not affect the optional SQL script. The
SQL script always inserts statistics for all traced SQL statements.
RECORD Creates a SQL script with the specified filename with all of the
non-recursive SQL in the trace file. This can be used to replay the user
events from the trace file.
TKPROF Statement Examples
This
section provides two brief examples of TKPROF usage. For an extensive
example of TKPROF output, see "Sample TKPROF Output" on page
22-7.
Example
1
If you are processing a large trace file using a combination of SORT parameters
and the PRINT parameter, you can produce a TKPROF output file containing
only the highest resource-intensive statements. For example, the following
statement will print the ten statements in the trace file that have generated
the most physical I/O:
TKPROF
ora53269.trc ora 53269.prf
SORT = (PRSDSK, EXEDSK, FCHDSK)
PRINT = 10
Example
2
This example runs TKPROF, accepts a trace file named dlsun12_jane_fg_svrmgr_007.trc,
and writes a formatted output file named outputa.prf:
TKPROF
DLSUN12_JANE_FG_SVRMGR_007.TRC OUTPUTA.PRF
EXPLAIN=SCOTT/TIGER TABLE=SCOTT.TEMP_PLAN_TABLE_A INSERT=STOREA.SQL
SYS=NO SORT=(EXECPU,FCHCPU)
This
example is likely to be longer than a single line on your screen and you
may have to use continuation characters, depending on your operating system.
Note
the other parameters in this example:
The
EXPLAIN value causes TKPROF to connect as the user SCOTT and use the EXPLAIN
PLAN command to generate the execution plan for each traced SQL statement.
You can use this to get access paths and row source counts.
The TABLE value causes TKPROF to use the table TEMP_PLAN_TABLE_A in the
schema SCOTT as a temporary plan table.
The INSERT value causes TKPROF to generate a SQL script named STOREA.SQL
that stores statistics for all traced SQL statements in the database.
The SYS parameter with the value of NO causes TKPROF to omit recursive
SQL statements from the output file. In this way you can ignore internal
Oracle statements such as temporary table operations.
The SORT value causes TKPROF to sort the SQL statements in order of the
sum of the CPU time spent executing and the CPU time spent fetching rows
before writing them to the output file. For greatest efficiency, always
use SORT parameters.
Step
4:
Interpreting
TKPROF Output
This section provides pointers for interpreting TKPROF output.
Tabular
Statistics
Library Cache Misses
User Issuing the SQL Statement
Execution Plan
Deciding What Statements to Tune
While TKPROF provides a very useful analysis, the most accurate measure
of efficiency is the actual performance of the application in question.
Note that at the end of the TKPROF output is a summary of the work done
in the database engine by the process during the period that the trace
was running.
Tabular
Statistics
TKPROF
lists the statistics for a SQL statement returned by the SQL trace facility
in rows and columns. Each row corresponds to one of three steps of SQL
statement processing. The step for which each row contains statistics
is identified by the value of the call column:
Parse
This step translates the SQL statement into an execution plan. This step
includes checks for proper security authorization and checks for the existence
of tables, columns, and other referenced objects.
Execute This step is the actual execution of the statement by Oracle.
For INSERT, UPDATE, and DELETE statements, this step modifies the data.
For SELECT statements, the step identifies the selected rows.
Fetch This step retrieves rows returned by a query. Fetches are only performed
for SELECT
statements.
The other columns of the SQL trace facility output are combined statistics
for all parses, all executes, and all fetches of a statement. These values
are zero (0) if TIMED_STATISTICS is not turned on. The sum of query and
current is the total number of buffers accessed.
count
Number of times a statement was parsed, executed, or fetched.
cpu
Total CPU time in seconds for all parse, execute, or fetch calls for the
statement.
elapsed
Total elapsed time in seconds for all parse, execute, or fetch calls for
the statement.
disk
Total number of data blocks physically read from the datafiles on disk
for all parse, execute, or fetch calls.
query
Total number of buffers retrieved in consistent mode for all parse, execute,
or fetch calls. Buffers are usually retrieved in consistent mode for queries.
current
Total number of buffers retrieved in current mode. Buffers are retrieved
in current mode for statements such as INSERT, UPDATE, and DELETE.
Rows
Statistics about the processed rows appear in the "rows" column.
rows
Total number of rows processed by the SQL statement. This total does not
include rows processed by subqueries of the SQL statement.
For SELECT statements, the number of rows returned appears for the fetch
step. For UPDATE, DELETE, and INSERT statements, the number of rows processed
appears for the execute step.
Resolution
of Statistics
Since timing statistics have a resolution of one hundredth of a second,
any operation on a cursor that takes a hundredth of a second or less may
not be timed accurately. Keep this in mind when interpreting statistics.
In particular, be careful when interpreting the results from simple queries
that execute very quickly.
Recursive
Calls
Sometimes in order to execute a SQL statement issued by a user, Oracle
must issue additional statements. Such statements are called recursive
calls or recursive SQL statements. For example, if you insert a row into
a table that does not have enough space to hold that row, Oracle makes
recursive calls to allocate the space dynamically. Recursive calls are
also generated when data dictionary information is not available in the
data dictionary cache and must be retrieved from disk.
If
recursive calls occur while the SQL trace facility is enabled, TKPROF
produces statistics for the recursive SQL statements and clearly marks
them as recursive SQL statements in the output file. Note that the statistics
for a recursive SQL statement are included in the listing for that statement,
not in the listing for the SQL statement that caused the recursive call.
So when you are calculating the total resources required to process a
SQL statement, you should consider the statistics for that statement as
well as those for recursive calls caused by that statement. Note that
setting the SYS command line parameter to NO suppresses the listing of
recursive calls in the output file.
Library
Cache Misses
TKPROF also lists the number of library cache misses resulting from parse
and execute steps for each SQL statement. These statistics appear on separate
lines following the tabular statistics. If the statement resulted in no
library cache misses, TKPROF does not list the statistic. In the example,
the statement resulted in one library cache miss for the parse step and
no misses for the execute step.
User
Issuing the SQL Statement
TKPROF also lists the user ID of the user issuing each SQL statement.
If the SQL trace input file contained statistics from multiple users and
the statement was issued by more than one user, TKPROF lists the ID of
the last user to parse the statement. The user ID of all database users
appears in the data dictionary in the column ALL_USERS.USER_ID.
Execution
Plan
If you specify the EXPLAIN parameter on the TKPROF command line, TKPROF
uses the EXPLAIN PLAN command to generate the execution plan of each SQL
statement traced. TKPROF also displays the number of rows processed by
each step of the execution plan.
Note:
Trace files generated immediately after instance startup contain data
that reflects the activity of the startup process. In particular, they
reflect a disproportionate amount of I/O activity as caches in the System
Global Area (SGA) are filled. For the purposes of tuning, ignore such
trace files.
Deciding
What Statements to Tune
The following listing shows TKPROF output for one SQL statement as it
appears in the output file:
SELECT
* FROM emp, dept WHERE emp.deptno = dept.deptno;
call count cpu elapsed disk query current rows
---- ------- ------- --------- -------- -------- ------- ------
Parse 11 0.08 0.18 0 0 0 0
Execute 11 0.23 0.66 0 3 6 2
Fetch 35 6.70 6.83 100 12326 2 824
------------------------------------------------------------------
total 57 7.01 7.67 100 12329 8 826
Misses
in library cache during parse: 0
10
user SQL statements in session.
0 internal SQL statements in session.
10 SQL statements in session.
If it is acceptable to expend 7.01 CPU seconds to insert, update or delete
2 rows and to retrieve 824 rows, then you need not look any further at
this trace output. In fact, a major use of TKPROF reports in a tuning
exercise is to eliminate processes from the detailed tuning phase.
You
can also see from this summary that 1 unnecessary parse call was made
(because there were more parse calls than SQL statements) and that array
fetch operations were performed. (You know this because more rows were
fetched than there were fetches performed.)
Finally,
you can see that very little physical I/O was performed; this is suspicious
and probably means that the same database blocks were being continually
re-visited.
Having
established that the process has used excessive resource, the next step
is to discover which SQL statements are the culprits. Normally only a
small percentage of the SQL statements in any process need to be tuned,
and these are the ones which use the greatest resource.
The
examples which follow were all produced with TIMED_STATISTICS=TRUE. However,
with the exception of locking problems and inefficient PL/SQL loops, neither
the CPU nor the elapsed time are necessary to find the problem statements.
The key is the number of block visits both query (that is, subject to
read consistency) and current (not subject to read consistency). Segment
headers and blocks which are going to be updated are always acquired in
current mode, but all query and sub-query processing requests the data
in query mode. These are precisely the same measures as the instance statistics
consistent gets and db block gets.
The
SQL parsed as SYS is recursive SQL used to maintain the dictionary cache,
and is not normally of great benefit; if the number of internal SQL statements
looks high it may be worth checking to see what has been going on. (There
may be excessive space management overhead.)
Step
5:
Storing
SQL Trace Facility Statistics
This section covers:
Generating
the TKPROF Output SQL Script
Editing the TKPROF Output SQL Script
Querying the Output Table
You may want to keep a history of the statistics generated by the SQL
trace facility for your application and compare them over time. TKPROF
can generate a SQL script that creates a table and inserts rows of statistics
into it. This script contains
a
CREATE TABLE statement that creates an output table named TKPROF_TABLE
INSERT statements that add rows of statistics, one for each traced SQL
statement, to the TKPROF_TABLE
After running TKPROF, you can run this script to store the statistics
in the database.
Generating
the TKPROF Output SQL Script
When you run TKPROF, use the INSERT parameter to specify the name of the
generated SQL script. If you omit this parameter, TKPROF does not generate
a script.
Editing
the TKPROF Output SQL Script
After TKPROF has created the SQL script, you may want to edit the script
before running it.
If
you have already created an output table for previously collected statistics
and you want to add new statistics to the existing table, remove the CREATE
TABLE statement from the script. The script will then insert the new rows
into the existing table.
If you have created multiple output tables, perhaps to store statistics
from different databases in different tables, edit the CREATE TABLE and
INSERT statements to change the name of the output table.
Querying
the Output Table
The following CREATE TABLE statement creates the TKPROF_TABLE:
CREATE
TABLE tkprof_table
(date_of_insert DATE,
cursor_num NUMBER,
depth NUMBER,
user_id NUMBER,
parse_cnt NUMBER,
parse_cpu NUMBER,
parse_elap NUMBER,
parse_disk NUMBER,
parse_query NUMBER,
parse_current NUMBER,
parse_miss NUMBER,
exe_count NUMBER,
exe_cpu NUMBER,
exe_elap NUMBER,
exe_disk NUMBER,
exe_query NUMBER,
exe_current NUMBER,
exe_miss NUMBER,
exe_rows NUMBER,
fetch_count NUMBER,
fetch_cpu NUMBER,
fetch_elap NUMBER,
fetch_disk NUMBER,
fetch_query NUMBER,
fetch_current NUMBER,
fetch_rows NUMBER,
clock_ticks NUMBER,
sql_statement LONG)
These
columns help you identify a row of statistics:
SQL_STATEMENT
The column value is the SQL statement for which the SQL trace facility
collected the row of statistics. Note that because this column has datatype
LONG, you cannot use it in expressions or WHERE clause conditions.
DATE_OF_INSERT
The column value is the date and time when the row was inserted into the
table. Note that this value is not exactly the same as the time the statistics
were collected by the SQL trace facility. Most output table columns correspond
directly to the statistics that appear in the formatted output file. For
example, the PARSE_CNT column value corresponds to the count statistic
for the parse step in the output file.
DEPTH
This column value indicates the level of recursion at which the SQL statement
was issued. For example, a value of 1 indicates that a user issued the
statement. A value of 2 indicates Oracle generated the statement as a
recursive call to process a statement with a value of 1 (a statement issued
by a user). A value of n indicates Oracle generated the statement as a
recursive call to process a statement with a value of n-1.
USER_ID
This column value identifies the user issuing the statement. This value
also appears in the formatted output file.
CURSOR_NUM
This column value is used by Oracle to keep track of the cursor to which
each SQL statement was assigned. Note that the output table does not store
the statement's execution plan.
The following query returns the statistics from the output table. These
statistics correspond to the formatted output shown in the section "Step
4: Interpreting TKPROF Output"
SELECT
* FROM tkprof_table;
DATE_OF_INSERT CURSOR_NUM DEPTH USER_ID PARSE_CNT PARSE_CPU PARSE_ELAP
-------------- ---------- ----- ------- --------- --------- ----------
27-OCT-1993 1 0 8 1 16 29
PARSE_DISK PARSE_QUERY PARSE_CURRENT PARSE_MISS EXE_COUNT EXE_CPU
---------- ----------- ------------- ---------- --------- -------
3 13 0 1 1 0
EXE_ELAP
EXE_DISK EXE_QUERY EXE_CURRENT EXE_MISS EXE_ROWS FETCH_COUNT
-------- -------- --------- ----------- -------- -------- -----------
0 0 0 0 0 0 1
FETCH_CPU
FETCH_ELAP FETCH_DISK FETCH_QUERY FETCH_CURRENT FETCH_ROWS
--------- ---------- ---------- ----------- ------------- ----------
3 26 2 2 4 14
SQL_STATEMENT
---------------------------------------------------------------------
SELECT * FROM EMP, DEPT WHERE EMP.DEPTNO = DEPT.DEPTNO
Avoiding
Pitfalls in TKPROF Interpretation
This section describes some fine points of TKPROF interpretation:
Finding
the Statements Which Constitute the Bulk of the Load
The Argument Trap
The Read Consistency Trap
The Schema Trap
The Time Trap
The Trigger Trap
The "Correct" Version
See Also: "EXPLAIN PLAN Restrictions" on page 21-16 for information
about TKPROF and bind variables.
Finding
the Statements Which Constitute the Bulk of the Load
Look at the totals and try to identify the statements, which constitute
the bulk of the load.
Do
not attempt to perform many different jobs within a single query. It is
more effective to separate out the different queries which should be used
when certain optional parameters are present, and when the parameters
provided contain wild cards.
If
certain parameters are not specified by the report user, the query uses
bind variables which have been set to "%" to cause the LIKE
clauses in the query to operate as if they were not there. It would be
more efficient to run a query in which these clauses are not present.
Note:
TKPROF cannot tell the TYPE of the bind variables simply by looking at
the text of the SQL statement. It assumes that TYPE is CHARACTER; if this
is not the case, you should put appropriate type conversions in the SQL
statement.
The
Argument Trap
Especially where the LIKE operator is used, the query may be markedly
less efficient for certain values, or types of value, in a bind variable.
This is because the optimizer must make an assumption about the probable
selectivity without knowing the value. If you are not aware of the values
being bound at run time then it is possible to fall into the "argument
trap".
The Read Consistency Trap
The next example illustrates the read consistency trap. Without knowing
that an uncommitted transaction had made a series of updates to the NAME
column it is very difficult to see why so many block visits would be incurred.
Cases
like this are not normally repeatable: if the process were run again,
it is unlikely that another transaction would interact with it in the
same way.
select
NAME_ID
from CQ_NAMES where NAME = 'FLOOR'
call count cpu elapsed disk query current rows
---- ----- --- ------- ---- ----- ------- ----
Parse 1 0.11 0.21 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.15 0.24 4 150 0 1
Misses in library cache during parse: 1
Parsing user id: 13 (DJONES)
Rows Execution Plan
---- --------- ----
0 SELECT STATEMENT
1 TABLE ACCESS (BY ROWID) OF 'CQ_NAMES'
2 INDEX (RANGE SCAN) OF 'CQ_NAMES_NAME' (NON_UNIQUE)
The
Schema Trap
This example shows an extreme and thus easily detected example of the
schema trap. At first it is difficult to see why such an apparently straightforward
indexed query needs to look at so many database blocks, or why it should
access any blocks at all in current mode.
select
NAME_ID
from CQ_NAMES where NAME = 'FLOOR'
call
count cpu elapsed disk query current rows
-------- ------- -------- --------- ------- ------ ------- ----
Parse 1 0.04 0.12 0 0 0 0
Execute 1 0.01 0.01 0 0 0 0
Fetch 1 0.32 0.32 38 44 3 1
Misses
in library cache during parse: 0
Parsing user id: 13 (JAUSTEN
Rows
Execution Plan
------- ---------------------------------------------------
0 SELECTSTATEMENT
3519 TABLE ACCESS (BY ROWID) OF 'CQ_NAMES'
0 INDEX (RANGE SCAN) OF 'CQ_NAMES_NAME' (NON-UNIQUE)
Two
statistics suggest that the query may have been executed via a full table
scan. These statistics are the current mode block visits, plus the number
of rows originating from the Table Access row source in the execution
plan. The explanation is that the required index was built after the trace
file had been produced, but before TKPROF had been run.
The
Time Trap
Sometimes, as in the following example, you may wonder why a particular
query has taken so long.
update
CQ_NAMES set ATTRIBUTES = lower(ATTRIBUTES)
where ATTRIBUTES = :att
call
count cpu elapsed disk query current rows
-------- ------- -------- --------- -------- -------- ------- ------
----
Parse 1 0.08 0.24 0 0 0 0
Execute 1 0.63 19.63 33 526 13 7
Fetch 0 0.00 0.00 0 0 0 0
Misses
in library cache during parse: 1
Parsing user id: 13 (JAUSTEN)
Rows
Execution Plan
------- ---------------------------------------------------
0 UPDATE STATEMENT
3519 TABLE ACCESS (FULL) OF 'CQ_NAMES'
Again, the answer is interference from another transaction. In this case
another
transaction held a shared lock on the table CQ_NAMES for several seconds
before and after the update was issued. It takes a fair amount of experience
to
diagnose that interference effects are occurring. Comparative data is
essential
when the interference is only contributing a short delay (or a small increase
in
block visits in the previous example). On the other hand, if the interference
is
only contributing a modest overhead, and the statement is essentially
efficient,
its statistics may never have to be subjected to analysis.
The
Trigger Trap
The resources reported for a statement include those for all of the SQL
issued while the statement was being processed. They therefore include
any resources used within a trigger, along with the resources used by
any other recursive SQL (such as that used in space allocation). With
the SQL trace facility enabled, TKPROF will report these resources twice.
Avoid trying to tune the DML statement, if the resource is actually being
consumed at a lower level of recursion.
You
may need to inspect the raw trace file to see exactly where the resource
is being expended. The entries for recursive SQL follow the PARSING IN
CURSOR entry for the user's statement. Within the trace file, the order
is less easily defined.
The
"Correct" Version
For comparison with the output which results from one of these traps having
sprung, here is the TKPROF output for the indexed query with the index
in place and without any contention effects.
select
NAME_ID
from CQ_NAMES where NAME = 'FLOOR'
call
count cpu elapsed disk query current rows
----- ------ ------ -------- ----- ------ ------- -----
Parse 1 0.01 0.01 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.00 0.00 0 4 0 1
Misses
in library cache during parse: 0
Parsing user id: 13 (JAUSTEN)
Rows
Execution Plan
------- ---------------------------------------------------
0 SELECT STATEMENT
1 TABLE ACCESS (BY ROWID) OF 'CQ_NAMES'
2 INDEX (RANGE SCAN) OF 'CQ_NAMES_NAME' (NON-UNIQUE)
One
of the marked features of this correct version is that the parse call
took 10 milliseconds of both elapsed and CPU time, but the query apparently
took no time at all to execute and no time at all to perform the fetch.
In fact, no parse took place because the query was already available in
parsed form within the shared SQL area. These anomalies are due to the
clock tick of 10 msec being too long to reliably record simple and efficient
queries.
TKPROF
Sample Output
This section provides an extensive example of TKPROF output. Note that
portions have been edited out for the sake of brevity.
Header
Copyright
(c) Oracle Corporation 1979, 1997. All rights reserved.
Trace file: v80_ora_2758.trc
Sort options: default
********************************************************************************
count = number of times OCI procedure was executed
cpu = cpu time in seconds executing
elapsed = elapsed time in seconds executing
disk = number of physical reads of buffers from disk
query = number of buffers gotten for consistent read
current = number of buffers gotten in current mode (usually for update)
rows = number of rows processed by the fetch or execute call
********************************************************************************
The following statement encountered a error during parse:
select deptno, avg(sal) from emp e group by deptno
having exists (select deptno from dept
where dept.deptno = e.deptno
and dept.budget > avg(e.sal)) order by 1
Error encountered: ORA-00904
********************************************************************************
Body
alter
session set sql_trace = true
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 0 0.00 0.00 0 0 0 0
Execute 1 0.00 0.15 0 0 0 0
Fetch 0 0.00 0.00 0 0 0 0
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 1 0.00 0.15 0 0 0 0
Misses in library cache during parse: 0
Misses in library cache during execute: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
.
.
.
********************************************************************************
select emp.ename, dept.dname from emp, dept
where emp.deptno = dept.deptno
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.12 0.14 2 0 2 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.00 0.00 2 2 4 14
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.12 0.14 4 2 6 14
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 SELECT STATEMENT GOAL: CHOOSE
14 MERGE JOIN
4 SORT (JOIN)
4 TABLE ACCESS (FULL) OF 'DEPT'
14 SORT (JOIN)
14 TABLE ACCESS (FULL) OF 'EMP'
********************************************************************************
select a.ename name, b.ename manager from emp a, emp b
where a.mgr = b.empno(+)
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.01 0.01 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.01 0.01 1 54 2 14
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.02 0.02 1 54 2 14
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 SELECT STATEMENT GOAL: CHOOSE
13 NESTED LOOPS (OUTER)
14 TABLE ACCESS (FULL) OF 'EMP'
13 TABLE ACCESS (BY ROWID) OF 'EMP'
26 INDEX (RANGE SCAN) OF 'EMP_IND' (NON-UNIQUE)
.
.
.
********************************************************************************
select ename,job,sal
from emp
where sal =
(select max(sal)
from emp)
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.00 0.00 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.00 0.00 0 16 4 1
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.00 0.00 0 16 4 1
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 SELECT STATEMENT GOAL: CHOOSE
14 FILTER
14 TABLE ACCESS (FULL) OF 'EMP'
14 SORT (AGGREGATE)
14 TABLE ACCESS (FULL) OF 'EMP'
********************************************************************************
select deptno
from emp
where job = 'clerk'
group by deptno
having count(*) >= 2
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.00 0.00 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.00 0.00 0 1 2 0
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.00 0.00 0 1 2 0
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 SELECT STATEMENT GOAL: CHOOSE
0 FILTER
0 SORT (GROUP BY)
14 TABLE ACCESS (FULL) OF 'EMP'
********************************************************************************
select dept.deptno,dname,job,ename
from dept,emp
where dept.deptno = emp.deptno(+)
order by dept.deptno
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.00 0.00 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.00 0.00 0 2 4 15
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.00 0.00 0 2 4 15
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 SELECT STATEMENT GOAL: CHOOSE
14 MERGE JOIN (OUTER)
4 SORT (JOIN)
4 TABLE ACCESS (FULL) OF 'DEPT'
14 SORT (JOIN)
14 TABLE ACCESS (FULL) OF 'EMP'
********************************************************************************
select grade,job,ename,sal
from emp,salgrade
where sal between losal and hisal
order by grade,job
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.06 0.08 2 18 1 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.01 0.01 1 11 12 14
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.07 0.09 3 29 13 14
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 SELECT STATEMENT GOAL: CHOOSE
14 SORT (ORDER BY)
14 NESTED LOOPS
5 TABLE ACCESS (FULL) OF 'SALGRADE'
70 TABLE ACCESS (FULL) OF 'EMP'
********************************************************************************
select lpad(' ',level*2)||ename org_chart,level,empno,mgr,job,deptno
from emp
connect by prior empno = mgr
start with ename = 'clark'
or ename = 'blake'
order by deptno
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.01 0.01 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.01 0.01 0 1 2 0
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.02 0.02 0 1 2 0
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 SELECT STATEMENT GOAL: CHOOSE
0 SORT (ORDER BY)
0 CONNECT BY
14 TABLE ACCESS (FULL) OF 'EMP'
0 TABLE ACCESS (BY ROWID) OF 'EMP'
0 TABLE ACCESS (FULL) OF 'EMP'
********************************************************************************
create table tkoptkp (a number, b number)
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.00 0.00 0 0 0 0
Execute 1 0.01 0.01 1 0 1 0
Fetch 0 0.00 0.00 0 0 0 0
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 2 0.01 0.01 1 0 1 0
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 CREATE TABLE STATEMENT GOAL: CHOOSE
.
.
.
********************************************************************************
insert into tkoptkp
values
(1,1)
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.07 0.09 0 0 0 0
Execute 1 0.01 0.20 2 2 3 1
Fetch 0 0.00 0.00 0 0 0 0
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 2 0.08 0.29 2 2 3 1
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 INSERT STATEMENT GOAL: CHOOSE
.
.
.
********************************************************************************
insert into tkoptkp select * from tkoptkp
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.00 0.00 0 0 0 0
Execute 1 0.02 0.04 0 2 3 12
Fetch 0 0.00 0.00 0 0 0 0
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 2 0.02 0.04 0 2 3 12
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 INSERT STATEMENT GOAL: CHOOSE
12 TABLE ACCESS (FULL) OF 'TKOPTKP'
********************************************************************************
select *
from
tkoptkp where a > 2
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.01 0.01 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 1 0.00 0.00 0 1 2 12
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 3 0.01 0.01 0 1 2 12
Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 8 (SCOTT)
Rows Execution Plan
------- ---------------------------------------------------
0 SELECT STATEMENT GOAL: CHOOSE
24 TABLE ACCESS (FULL) OF 'TKOPTKP'
********************************************************************************
Summary
OVERALL
TOTALS FOR ALL NON-RECURSIVE STATEMENTS
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 18 0.40 0.53 30 182 3 0
Execute 19 0.05 0.41 3 7 10 16
Fetch 12 0.05 0.06 4 105 66 78
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 49 0.50 1.00 37 294 79 94
Misses in library cache during parse: 18
Misses in library cache during execute: 1
OVERALL TOTALS FOR ALL RECURSIVE STATEMENTS
call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 69 0.49 0.60 9 12 8 0
Execute 103 0.13 0.54 0 0 0 0
Fetch 213 0.12 0.27 40 435 0 162
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 385 0.74 1.41 49 447 8 162
Misses in library cache during parse: 13
19 user SQL statements in session.
69 internal SQL statements in session.
88 SQL statements in session.
17 statements EXPLAINed in this session.
********************************************************************************
Trace file: v80_ora_2758.trc
Trace file compatibility: 7.03.02
Sort options: default
1 session in tracefile.
19 user SQL statements in trace file.
69 internal SQL statements in trace file.
88 SQL statements in trace file.
41 unique SQL statements in trace file.
17 SQL statements EXPLAINed using schema:
SCOTT.prof$plan_table
Default table was used.
Table was created.
Table was dropped.
1017 lines in trace file.
Tracking the Explain Plan and Stats without doing anything but setting
the parameters at SQL*Plus
SET
AUTOTRACE OFF No AUTOTRACE report is generated. This is the default
SET AUTOTRACE ON The AUTOTRACE report includes both the optimizer execution
path and the SQL statement execution statistics
SET AUTOTRACE ON EXPLAIN The AUTOTRACE report shows only the optimizer
execution path
SET AUTOTRACE ON STATISTICS The AUTOTRACE report shows only the SQL statement
execution statistics.
SET AUTOTRACE TRACEONLY Like SET AUTOTRACE ON, but suppresses the printing
of the user's query output, if any.
Utilities to gather Stats for the tables/indexes/schemas/database
DBMS_STATS
DBMS_STATS
provides a mechanism for you to view and modify optimizer statistics gathered
for database objects.The statistics can reside in two different locations:
The
dictionary.
A
table created in the user's schema for this purpose.
Only
statistics stored in the dictionary itself have an impact on the cost-based
optimizer.
Example
Scenario
There
has been a lot of modification against the emp table since the last time
statistics were gathered. To ensure that the cost-based optimizer is still
picking the best plan, statistics should be gathered once again; however,
the user is concerned that new statistics will cause the optimizer to
choose bad plans when the current ones are acceptable. The user can do
the following:
BEGIN
DBMS_STATS.CREATE_STAT_TABLE ('scott', 'savestats');
DBMS_STATS.GATHER_TABLE_STATS ('scott', 'emp', 5, stattab => 'savestats');
END;
This operation gathers new statistics on emp, but first saves the original
statistics in a user stat table: emp.savestats.
If
the user believes that the new statistics are causing the optimizer to
generate poor plans, then the original stats can be restored as follows:
BEGIN
DBMS_STATS.DELETE_TABLE_STATS ('scott', 'emp');
DBMS_STATS.IMPORT_TABLE_STATS ('scott', 'emp', stattab => 'savestats');
END;
DBMS_UTILITY
This
package provides various utility subprograms.
DBMS_UTILITY
submits a job for each partition. It is the users responsibility to control
the number of concurrent jobs by setting the INIT.ORA parameter JOB_QUEUE_PROCESSES
correctly. There is minimal error checking for correct syntax. Any error
is reported in SNP trace files.
DBMS_UTILITY
runs with the privileges of the calling user for the NAME_RESOLVE, COMPILE_SCHEMA,
and ANALYZE_SCHEMA procedures. This is necessary so that the SQL works
correctly.
This
does not run as SYS. The privileges are checked using DBMS_DDL.
DBMS_SYSTEM
This
package enables you to gather information about events set in the current
session and manipulate other user's sessions to set events and change
the values of certain init.ora parameters. It provides some of the capability
of DBMS_SESSION but with the ability to affect any session.
Oracle
version 9.2 introduced a procedure to reset the IO counters in dynamic
performance views (KCFRMS) and a suite of procedures that allow privileged
users to write messages to the session trace file and instance alert log.
Note:
I
should add that this package has been left undocumented for good reason
and since Oracle8, it has been wrapped and moved into prvtutil.plb “for
obscurity”.
Requirements
This
package is owned by SYS and is not generally available. I recommend that
you only use this when logged on as SYS or connected as SYSDBA
Privileges
Before
using this package, you must run the DBMSUTIL.SQL and PRVTUTIL.PLB scripts
to create the DBMS_SYSTEM package.
DIST_TXN_SYNC
Procedure
There
is no public information available for this procedure, but it somehow
sounds like distributed transaction synchronisation. According to Tom
Kyte it is used in XA interfaces and is nothing you would ever need to
call directly.
Syntax
DBMS_SYSTEM.DIST_TXN_SYNC
(inst_num IN NUMBER);
KCFRMS
Procedure
This
procedure resets the timers displayed by MAX_WAIT in V$SESSION_EVENT and
MAXIORTM, MAXIOWTM in V$FILESTAT (X$KCFIO) views.
Syntax
DBMS_SYSTEM.KCFRMS
;
Example
To
check for the current timer values:
SQL>
select MAX_WAIT from V$SESSION_EVENT;
SQL> select MAXIORTM, MAXIOWTM from V$FILESTAT;
KSDDDT Procedure
This
procedure prints the timestamp to the trace file. It can't be used for
the alert log.
Syntax
DBMS_SYSTEM.KSDDDT
;
Example
SQL>
exec dbms_system.ksdddt ;
Output
in tracefile looks like:
***
2003-03-10 21:23:17.000
KSDFLS
Procedure
This
procedure flushes any pending output to the target file (alert log and/or
trace file).
Syntax
DBMS_SYSTEM.KSDFLS
;
KSDIND Procedure
This
procedure does an 'indent' before the next write (ksdwrt) by printing
that many colons (:) before the next write.
Syntax
DBMS_SYSTEM.KSDIND
(lvl IN BINARY_INTEGER);
Example
Write
in the same line 5 ':' and the text 'Test Alert Msg' to the tracefile
and the alertlog:
SQL>
exec dbms_system.ksdind (5);
SQL> exec dbms_system.ksdwrt (3, 'Test Alert Msg');
Output
in tracefile/alertlog looks like:
:::::Test
Alert Msg
KSDWRT
Procedure
This
procedure prints the message to the target file (alert log and/or trace
file).
Syntax
DBMS_SYSTEM.KSDWRT
(
dest IN BINARY_INTEGER,
tst IN VARCHAR2);
Example
Write
the text 'Test Alert Msg' to the tracefile and the alertlog:
SQL>
exec dbms_system.ksdwrt (3, ' ');
SQL> exec dbms_system.ksdwrt (3, '--- Start ---');
SQL> exec dbms_system.ksdddt;
SQL> exec dbms_system.ksdwrt (3, ' Test Alert Msg');
SQL> exec dbms_system.ksdwrt (3, '--- End ---');
SQL> exec dbms_system.ksdwrt (3, ' ');
The
output in the aleret log file is as under
(This is the output from Oracle 10.1)
Wed
Aug 04 22:13:47 2004
ARC1: Evaluating archive log 2 thread 1 sequence 138
Committing creation of archivelog 'D:\ORACLE\FLASH_RECOVERY_AREA\ORA10G\ARCHIVELOG\2004_08_04\O1_MF_1_138_0K39HD5M_.ARC'
Created Oracle managed file D:\ORACLE\FLASH_RECOVERY_AREA\ORA10G\ARCHIVELOG\2004_08_04\O1_MF_1_138_0K39HD5M_.ARC
Wed Aug 04 22:15:54 2004
--- Start ---
Test Alert Msg
--- End ---
ORADEBUG
ORADEBUG is an undocumented debugging utility supplied with Oracle
The
ORADEBUG HELP command lists the commands available within ORADEBUG
These vary by release and platform. Commands appearing in this help do
not necessarily work for the release/platform on which the database is
running
For example in Oracle 10.1.0.2.0 (Windows 2000) the command
ORADEBUG HELP
Returns the following
HELP [command] Describe one or all commands
SETMYPID Debug current process
SETOSPID <ospid> Set OS pid of process to debug
SETORAPID <orapid> ['force'] Set Oracle pid of process to debug
DUMP <dump_name> <lvl> [addr] Invoke named dump
DUMPSGA [bytes] Dump fixed SGA
DUMPLIST Print a list of available dumps
EVENT <text> Set trace event in process
SESSION_EVENT <text> Set trace event in session
DUMPVAR <p|s|uga> <name> [level] Print/dump a fixed PGA/SGA/UGA
variable
SETVAR <p|s|uga> <name> <value> Modify a fixed PGA/SGA/UGA
variable
PEEK <addr> <len> [level] Print/Dump memory
POKE <addr> <len> <value> Modify memory
WAKEUP <orapid> Wake up Oracle process
SUSPEND Suspend execution
RESUME Resume execution
FLUSH Flush pending writes to trace file
CLOSE_TRACE Close trace file
TRACEFILE_NAME Get name of trace file
LKDEBUG Invoke global enqueue service debugger
NSDBX Invoke CGS name-service debugger
-G <Inst-List | def | all> Parallel oradebug command prefix
-R <Inst-List | def | all> Parallel oradebug prefix (return output
SETINST <instance# .. | all> Set instance list in double quotes
SGATOFILE <SGA dump dir> Dump SGA to file; dirname in double quotes
DMPCOWSGA <SGA dump dir> Dump & map SGA as COW; dirname in double
quotes
MAPCOWSGA <SGA dump dir> Map SGA as COW; dirname in double quotes
HANGANALYZE [level] [syslevel] Analyze system hang
FFBEGIN Flash Freeze the Instance
FFDEREGISTER FF deregister instance from cluster
FFTERMINST Call exit and terminate instance
FFRESUMEINST Resume the flash frozen instance
FFSTATUS Flash freeze status of instance
SKDSTTPCS <ifname> <ofname> Helps translate PCs to names
WATCH <address> <len> <self|exist|all|target> Watch
a region of memory
DELETE <local|global|target> watchpoint <id> Delete a watchpoint
SHOW <local|global|target> watchpoints Show watchpoints
CORE Dump core without crashing process
UNLIMIT Unlimit the size of the trace file
PROCSTAT Dump process statistics
CALL <func> [arg1] ... [argn] Invoke function with arguments
This
will be discussed in another lesson later as this lesson has already touched
30 pages
|