03 -oracle 10g performance tuning

Oracle 10g DBA

Performance Tuning

JENiSOFT SOLUTIONS

#40A, G2 Ponnammal Street, Kumaran Colony,

Vadapalni Chennai-600026 Mobile: 98409 41415

Nearest Land Mark: SRM UNIVERCITY

JENiSOFT SOLUTIONS Oracle 10g PT1

Oracle Performance Tuning

Optimizer Mode:

The goal of the query optimizer is the best throughput. This means that it chooses the

least amount of resources necessary to process all rows accessed by the statement.

Oracle can also optimize a statement with the goal of best response time. This means

that it uses the least amount of resources necessary to process the first row accessed by

a SQL statement.

RBO – Rule Based Optimizer mode

CBO – Cost Based Optimizer mode

Default keyword is “Choose”

o RULE

o ALL_ROWS

o FIRST_ROWS_n

o FIRST_ROWS

Choose Optimizer Mode:

Oracle will execute the rule-based optimizer if there are no statistics present for the

table or execute the cost-based optimizer if statistics are present.

RULE:

This Optimizer used to decide how to execute a given SQL statement.

One rule stated than an index would be used in favor of a full table scan

The Rule Based Optimizer (RBO) would always use this index to perform the query

EMP

ENO NUMBER PRIMARY KEY


ENAME VARCHAR2(30)

GENDER CHAR(1)

SQL > select * from emp where eno=101;

SQL > select * from emp where gender='F';

OPTIMIZER_INDEX_COST_ADJ :

Set between 1 and 10 to force index use more frequently and would pretty much

guarantee index use.

The range of values is 1 to 10000. Default of this parameter value is 100

SQL > show parameter optimizer_index_cost_adj;

SQL > alter system set optimizer_index_cost_adj=10;

OPTIMIZER_INDEX_CACHING :

The range of values 0 to 100 for this parameter.

OPTIMIZER_INDEX_CACHING indicates percentage of index blocks in the buffer

cache,

A value of 100 infers that 100% of the index blocks are likely to be found in the

buffer cache

SQL > show parameter optimizer_index_caching;

SQL > alter system set optimizer_index_caching=100;

OPTIMIZER_DYNAMIC_SAMPLING

If no statistics are available then the optimizer uses dynamic sampling.

This parameter should be set to a value of 2 or higher. The default value is 2

The range of value is 0 to 10 for this parameter.

A certain number of blocks are read by the dynamic sampling query based on this

parameter value.

Level 1 : Samples tables that appear in join or sub query conditions that have no


indexes and have more blocks than 32, the default for dynamic sampling.

Level 2 : Samples all unanalyzed tables that have more than 32 blocks.

Level 3 : Samples tables using a single column that applies selectivity to the table

being sampled.(2 X32)

Level : Samples tables using two or more columns that apply selectivity to the

table being sampled.(2 X 32)

Level 5 : Doubles the dynamic sample size and samples 64 blocks on tables.

Level 6 : Quadruples the dynamic sample size and samples 128 blocks on tables.

Level 7 : Samples 256 blocks on tables.

Level 8 : Samples 1,024 blocks on tables.

Level 9 : Samples 4,096 blocks on tables.

Level 10 : Samples all of the block in the tables.

SQL > show parameter optimizer_dynamic_sampling

SQL > alter system set optimizer_dynamic_sampling=2;

SQL > alter system set timed_statistices=TRUE;

SQL > ALTER SYSTEM SET statistics_level=typical;

SQL > ALTER SYSTEM SET statistics_level=all;

FIRST_ROWS

This is a cost-based optimizer mode that will return rows as soon as possible, even if the

overall query runs longer or consumes more resources.

The first_rows mode favors index scans over full-table scans,

ALL_ROWS:


This is a cost-based optimizer mode that ensures that the overall query time is

minimized, even if it takes longer to receive the first row.

The all_rows mode favors full-table scans,

The all_rows mode is best suited for batch-oriented queries where intermediate rows

are not required for viewing.

FIRST_ROWS_n:

optimizes with a goal of best responsetime to return the first n number of rows;

n can equal 1, 10, 100,or 1000

The first_rows_n mode using either full-table scans or fast index full scan

SQL > show parameter optimizer_mode;

SQL > alter system set optimizer_mode=choose;

Query Tuning or SQL Tuning:

Step 1:

$ top

SQL > select addr from v$process where spid=123;

X123H023

SQL > select sid,sql_id,sql_Address from v$session Where paddr=’ X123H023’;

SID SQL_ID SQL_ADDRESS

15 XA234 XhA234

SQL > select sql_text from v$sqlarea where SQL_ID=’ XA234’ and address=’ XhA234’;

Step 2:

SQL > select sid,sql_id,start_time from v$session_longops order by start_time;

SID SQL_ID StART_TIME

15 XA234 21-Nov-2010 10:30:33


SQL > select sql_text from v$sqlarea where SQL_ID=’ XA234’ ;

Step 3:

SQL > select b.username, a.disk_reads ,a.executions , a.disk_reads /a.executions, a.sql_text

from V$sqlarea a, dba_users b

where a.parsing_user_id = b.user_id and a.disk_reads > 100000

order by a.disk_reads desc;

USERNAME READS EXEC RDS_EXEC_RATIO STATEMENT

ADHOC1 7281934 1 7281934 select custno, ordno

from cust, orders

Step 4:

AWR report or statspack report or ADDM report.

Buffer Gets Exec Gets per Exec CPU Time (s) ElapTime (s) Hash Value

16,177,286 1 16,177,286.0 209.00 297.91 342588312

select sf.satz_id, f1.feld, f2.feld feld20, substr(sf.fehler

,instr(sf.fehler,'geschrieben:')+13) feld30

SQL tuning Method:

Autotrace method

Tkprof

Explain plan

Auotmatic SQL tuning


AutoTrace Method:

SET AUTOTRACE {ON | OFF | TRACEONLY | EXPLAIN | STATISTICS }

Step 1:

SQL > set autotrace on;

SQL > select * from emp where eno = 101;

ENO ENAME

101 Stanly

Execution Plan

0 SELECT STATEMENT Optimizer=CHOOSE (Cost=7 Card=1 Bytes=12)

1 0 TABLE ACCESS* (FULL) OF 'ENO' (Cost=7 Card=1 Bytes=12)

Statistics

0 db block gets

0 consistent gets

0 physical reads

0 sorts (memory)

0 sorts (disk)

Step 2:


SQL > analyze table EMP compute statistics;

Or

SQL > exec DBMS_STATS.GATHER_TABLE_STATS ('STANLY', 'EMP');



ENO ENAME

101 Stanly

Execution Plan


1 0 TABLE ACCESS* (FULL) OF 'ENO' (Cost=7 Card=1 Bytes=12)

Statistics

90 db block gets

110 consistent gets

100 physical reads

0 sorts (memory)

0 sorts (disk)

% Ratio = (1-(Physical reads/(consistent gets + db block gets)))*100

SQL > create index ind_eno on emp (eno);


SQL > analyze table EMP compute statistics;

Or

SQL > exec DBMS_STATS.GATHER_TABLE_STATS ('STANLY', 'EMP');


ENO ENAME

101 Stanly

Execution Plan



1 0 INDEX ACCESS OF 'IND_ENO' (Cost=2 Card=1 Bytes=12)

Statistics

90 db block gets

110 consistent gets

10 physical reads

0 sorts (memory)

0 sorts (disk)

Index Tuning:

SQL > analyze index IND_ENO validate structure;

Or

SQL > exec DBMS_STATS.GATHER_INDEX_STATS ('STANLY', 'IND_ENO');

INDEX_STATS:

NAME : Name of the index

DEL_LF_ROWS : Number of deleted leaf rows in the index

DEL_LF_ROWS_LEN : Total length of all deleted rows in the index

LF_ROWS : Number of leaf rows (values in the index)

LF_ROWS_LEN : Sum of the lengths of all the leaf rows

SQL > select (del_lf_row_len / lf_row_len)*100 from Index_stats where name=’IND_ENO’;

SQL > alter index IND_ENO monitoring usage;

SQL > select index_name,used,monitoring from v$object_usage where index_name=’IND_ENO’;

INDEX_NAME USED MONITORING

IND_ENO NO YES

SQL > alter index IND_ENO nomonitoring usage;


SQL > alter index IND_ENO coalesce;

SQL > alter index IND_ENO rebuild;

SQL > alter index IND_ENO rebuild online;

Row Chaining and Row Migration :

Method 1:

SQL > ANALYZE TABLE emp COMPUTE STATISTICS;

SQL > SELECT chain_cnt FROM dba_tables WHERE table_name = 'EMP';

Method 2:

SQL > @?/rdbms/admin/UTLCHAIN.SQL

SQL > ANALYZE TABLE emp LIST CHAINED ROWS;

SQL > select count(*) from chained_rows where owner_name='STANLY' and

Table_name='EMP';

Solutions:

1. Alter table emp move tablespace <old tablespace> to <new tablespace>;

2. export table / drop or truncate table/ import table

High Water Mark:

SQL > SELECT segment_name, blocks, segment_owner

FROM dba_segments WHERE segment_name='BIG_EMP1';

SEGMENT_NAME BLOCKS SEGMENT_OWNER

BIG_EMP1 1024 STANLY

SQL > ANALYZE TABLE stanly.big_emp1 ESTIMATE STATISTICS;

SQL> SELECT table_name, blocks, empty_blocks FROM dba_tables

WHERE table_name='BIG_EMP1' and owner='STANLY';

TABLE_NAME BLOCKS EMPTY_BLOCKS


BIG_EMP1 700 323

BLOCKS + EMPTY_BLOCKS (700+323=1023)

SQL> SELECT COUNT (DISTINCT DBMS_ROWID.ROWID_BLOCK_NUMBER (rowid)) "Used"

FROM stanly.big_emp1;

Used

700

SQL> DELETE from stanly.big_emp1;

SQL> commit;

SQL> ANALYZE TABLE stanly.big_emp1 ESTIMATE STATISTICS;




BIG_EMP1 700 323

SQL> SELECT COUNT (DISTINCT DBMS_ROWID.ROWID_BLOCK_NUMBER (rowid)) "Used"

FROM stanly.big_emp1;

Used

0

Solutions :

Method 1:

Export stanly.big_emp1 table

SQL > TRUNCATE TABLE stanly.big_emp1;

Import stanly.big_emp1 table

Method 2:


SQL > alter table stanly.big_emp1 enable row movement;

SQL > alter table stanly.big_emp1 shrink space;

SQL > ANALYZE TABLE big_emp1 ESTIMATE STATISTICS;




BIG_EMP1 0 511

Tkprof Method:

SQL > alter session set sql_trace=true;

SQL > SELECT * FROM emp, dept WHERE emp.deptno = dept.deptno;

SQL > alter session set sql_trace=false

$ tkprof /u01/admin/prod/udump/ora_prod_124.trc /u01/emp_trace.txt

$ more /u01/emp_trace.txt

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 0

Fetch 35 6.70 6.83 100 12326 2 824

Total 57 7.01 7.67 100 12329 8 824

%Ratio = ( 1 - ( Disk / ( Query + Current ) ) ) * 100

Rows Execution Plan


14 MERGE JOIN

4 SORT JOIN

4 TABLE ACCESS (FULL) OF 'DEPT'

14 SORT JOIN

14 TABLE ACCESS (FULL) OF 'EMP'

Explain Plan Method:

SQL > @?/rdbms/admin/utlxplan.sql

SQL > EXPLAIN PLAN SET STATEMENT_ID = 'st1' FOR SELECT last_name FROM emp;

SQL > select operation, options, object_name, cost, cardinality, bytes, cpu_cost

from plan_table where statement_id='st1';

OPERATION OPTIONS OBJECTNAME COST CARD BYTES CPUCOST

Select statement Full Access EMP 7 100 1024 75

Automatic SQL Tuning:

SQL > exec DBMS_SQLTUNE.create_tuning_task

(

sql_text => 'SELECT e.eno,e.ename, d.dno,d.dname

FROM emp e, dept d where e.deptno = d.deptno ',

user_name => 'scott',

scope => DBMS_SQLTUNE.scope_comprehensive,

time_limit => 60,

task_name => 'emp_dept_tuning_task',

description => 'Tuning task for an EMP to DEPT join query.'

);

SQL > exec EXEC DBMS_SQLTUNE.execute_tuning_task (task_name => 'emp_dept_tuning_task');


SQL > SELECT task_name, status FROM dba_advisor_log WHERE owner = 'SCOTT';

TASK_NAME STATUS

emp_dept_tuning_task COMPLETED

SQL > SELECT DBMS_SQLTUNE.report_tuning_task('emp_dept_tuning_task') AS

recommendations FROM dual;

RECOMMENDATIONS

GENERAL INFORMATION SECTION

Tuning Task Name : emp_dept_tuning_task

Scope : COMPREHENSIVE

Time Limit(seconds) : 60

Completion Status : COMPLETED

Started at : 05/11/2010 09:29:13

Completed at : 05/11/2010 09:29:15

-------------------------------------------------------------------------------

SQL ID : 0wrmfv2yvswx1

SQL Text : SELECT e.eno,e.ename, d.dno,d.dname FROM emp e, dept d

where e.deptno = d.deptno

-------------------------------------------------------------------------------

FINDINGS SECTION (1 findings)

1- Statistics Finding

Table "SCOTT"."EMP" and its indices were not analyzed.

Recommendation


Consider collecting optimizer statistics for this table and its indices.

execute dbms_stats.gather_table_stats(ownname => 'SCOTT', tabname =>

'EMP', estimate_percent => DBMS_STATS.AUTO_SAMPLE_SIZE,

method_opt => 'FOR ALL COLUMNS SIZE AUTO', cascade => TRUE)

SQL > execute dbms_stats.gather_table_stats(ownname => 'SCOTT',

tabname =>'EMP', estimate_percent => DBMS_STATS.AUTO_SAMPLE_SIZE,

method_opt => 'FOR ALL COLUMNS SIZE AUTO', cascade => TRUE)

SQL > EXEC DBMS_SQLTUNE.reset_tuning_task (task_name => 'emp_dept_tuning_task');

SQL > exec EXEC DBMS_SQLTUNE.execute_tuning_task (task_name => 'emp_dept_tuning_task');

SQL > SELECT DBMS_SQLTUNE.report_tuning_task('emp_dept_tuning_task') AS

recommendations FROM dual;

RECOMMENDATIONS

GENERAL INFORMATION SECTION

Tuning Task Name : emp_dept_tuning_task

Scope : COMPREHENSIVE

Time Limit(seconds) : 60

Completion Status : COMPLETED

Started at : 05/11/2010 09:29:13

Completed at : 05/11/2010 09:29:15

There are no recommendations...

SQL > exec DBMS_SQLTUNE.drop_tuning_task (task_name => 'emp_dept_tuning_task');

Statspack Report:

SQL > @?/rdbms/admin/spcreate.sql

Choose the PERFSTAT user's password


Enter value for perfstat_password : perfstat

Choose the Default tablespace for the PERFSTAT user

TABLESPACE_NAME CONTENTS STATSPACK

PERFSTAT PERMANENT

SYSAUX PERMANENT *

USERS PERMANENT

Enter value for default_tablespace: PERFSTAT

Choose the Temporary tablespace for the PERFSTAT user

TABLESPACE_NAME CONTENTS DB

TEMP TEMPORARY *

Enter value for temporary_tablespace: TEMP

.....

Creating Package STATSPACK...

No errors.

NOTE:

SPCPKG complete. Please check spcpkg.lis for any errors.

SQL > show user;

User “PERFSTAT”

SQL> SELECT * FROM stats$level_description ORDER BY snap_level;

Level 0 : This level captures general statistics, including rollback segment, row

cache, SGA, system events, background events, session events, system

statistics, wait statistics, lock statistics, and Latch information.

Level 5 : This level includes capturing high resource usage SQL Statements, along


with all data captured by lower levels.

Level 6 : This level includes capturing SQL plan and SQL plan usage information for

high resource usage SQL Statements, along with all data captured by lower

levels.

Level 7 : This level captures segment level statistics, including logical and

physical reads, row lock, itl and buffer busy waits, along with all data

captured by lower levels.

Level 10 : This level includes capturing Child Latch statistics, along with all data

captured by lower levels.

SQL> exec statspack.snap(i_snap_level => 6, i_modify_parameter => 'true');

SQL> exec statspack.snap;

SQL> exec statspack.snap;

SQL> select snap_id,to_char(snap_time,'DD.MM.YYYY:HH24:MI:SS') "Date/Time"

from stats$snapshot ;

SNAP_ID Date/Time

1 13.11.2010:08:48:47

2 13.11.2010:09:00:01

3 13.11.2010:10:30:48

4 13.11.2010:11:00:01

SQL> @?/rdbms/admin/spreport.sql

SNAP_ID Date/Time

1 13.11.2010:08:48:47

2 13.11.2010:09:00:01

3 13.11.2010:10:30:48

4 13.11.2010:11:00:01


Enter Begin Snap Id : 3

Enter End Snap Id : 4

Enter statspack report Name : /u01/Prod_snap_3_4_report.txt

$ more /u01/Prod_snap_3_4_report.txt

STATSPACK report for

DB Name DB Id Instance Inst Num Release RAC Host

--------------------------------------------------------------------------------------------------------------------------------

PROD 2006521736 PROD 1 10.2.0.4.0 NO Linux1_prod

Snap Id Snap Time Sessions Curs/Sess Comment

--------------------------------------------------------------------------------------------

Begin Snap: 3 13-Nov-10 10:30:48 15 14.3

End Snap: 4 13-Nov-10 11:00:01 15 10.2

Elapsed: 30.00 (mins)

Cache Sizes (end)

~~~~~~~~~~~~~~~~~

Buffer Cache : 24M Std Block Size : 4K

Shared Pool Size : 764M Log Buffer : 1,000K

Instance Efficiency Percentages (Target 100%)

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Buffer Nowait % : 100.00 Redo NoWait % : 99.99

Buffer Hit % : 95.39 In-memory Sort % : 100.00

Library Hit % : 99.42 Soft Parse % : 94.45


Execute to Parse % : 36.39 Latch Hit % : 100.00

Parse CPU to Parse Elapsd % : 59.15 % Non-Parse CPU : 99.31

Top 5 Timed Events

~~~~~~~~~~~~~~~~~~ % Total

Event Waits Time (s) Call Time

-------------------------------------------- ------------ ----------- ---------

CPU time 122 91.65

db file sequential read 1,571 2 1.61

db file scattered read 1,174 2 1.59

log file sequential read 342 2 1.39

control file parallel write 450 2 1.39

SQL ordered by Gets

---------------------------

Buffer Gets Executions Gets per Exec %Total CPU Time (s) Hash Value

--------------- ------------ -------------- ---------- ----------------- --------- ----------

16,926 1 16,926.0 2.3 2.36 1279400914

Module: SQL*Plus

create table test as select * from all_objects

SQL > @?/rdbms/admin/sprepsql.sql

Enter the Hash Value : 1279400914

SQL > @?/rdbms/admin/spauto.sql

SQL > @?/rdbms/admin/sppurge.sql


Enter lower snap id : 1

Enter upper snap id : 50

SQL > @?/rdbms/admin/spdrop.sql

AWR Report:

SQL > exec DBMS_WORKLOAD_REPOSITORY.CREATE_SNAPSHOT ( );

SQL > exec DBMS_WORKLOAD_REPOSITORY.MODIFY_SNAPSHOT_SETTINGS

( retention => 43200, interval => 30)

SQL > @?/rdbms/admin/awrrpt.sql

Enter value for report_type [ HTML / TEXT ] : text

Enter value for num_days : 2

SNAP_ID Date/Time

150 13.11.2010:08:48:47

151 13.11.2010:09:00:01

152 13.11.2010:10:30:48

153 13.11.2010:11:00:01

Enter value for begin_snap : 150

Enter value for end_snap : 153

Enter value for report_name : /u01/awrrpt_1_150_153

$ more /u01/awrrpt_1_150_153

SQL > select * from DBA_HIST_SNAPSHOT;

SQL > exec DBMS_WORKLOAD_REPOSITORY.DROP_SNAPSHOT_RANGE

(low_snap_id => 22, high_snap_id => 32)


ADDM Report:

SQL > @$?/rdbms/admin/addmrpt.sql

Instance DB Name Snap Id Snap Started

PROD PROD 136 20 Oct 2010 18:30

137 20 Oct 2010 19:00

138 20 Oct 2010 19:30

139 20 Oct 2010 20:00

140 20 Oct 2010 20:30

141 20 Oct 2010 21:00

142 20 Oct 2010 21:30

Enter value for begin_snap : 137

Enter value for end_snap : 141

Enter value for report_name : /u01/addmrpt_1_137_145.txt

$ more /u01/addmrpt_1_137_145.txt

FINDING 1 : 59% impact (944 seconds)

The buffer cache was undersized causing significant additional read I/O.

RECOMMENDATION 1: DB Configuration, 59% benefit (944 seconds)

ACTION: Increase SGA target size by increasing the value of parameter

"sga_target" by 28 M.

FINDING 2 : 31% impact (7798 seconds)

SQL statements were not shared due to the usage of literals. This resulted in

additional hard parses which were consuming significant database time.


RECOMMENDATION 1: Application Analysis, 31% benefit (7798 seconds)

ACTION: Investigate application logic for possible use of bind variables

instead of literals. Alternatively, you may set the parameter

"cursor_sharing" to "force".

Memory Tuning :

Database Buffer Cache

Redo log Buffer

Shared Pool

Database Buffer Cache:

V$sysstat

Statspack

AWR report

ADDM report

SQL > select (1 – ( p.value / ( d.value + c.value ) ) )*100 from v$sysstat p, v$sysstat d, v$sysstat c

Where p.name =’physical reads’ and d.name = ‘db block gets’ and

c.name = ‘consistent gets’ ;

Solutions :

Method 1:

SQL > alter system set db_cache_advice=on;

SQL > SELECT size_for_estimate, estd_physical_read_factor, estd_physical_reads

FROM V$DB_CACHE_ADVICE WHERE name = 'DEFAULT'

Size for Estimate Estd_physical read factor Estd_physical reads

243 1.33 13,720,149

273 1.13 11,583,180

304 1.00 10,282,475


334 .93 9,515,878

364 .87 8,909,026

395 .83 8,909,026

424 .79 8,909,026

SQL > alter system set db_cache_size=364M;

Method 2:

SQL > alter system set db_keep_cache_size=180M;

SQL > alter system set db_recycle_cache_size=364M;

SQL > SELECT o.OBJECT_NAME, COUNT(*) FROM DBA_OBJECTS o, V$BH bh

WHERE o.DATA_OBJECT_ID = bh.OBJD AND o.OWNER != 'SYS' GROUP BY o.OBJECT_NAME

Object_name Count

OM_EXT_HEADER 701

OM_SHELL 1,765

OM_HEADER 5,826

OM_INSTANCE 12,644

SQL > alter table OM_HEADER storage (buffer_pool keep);

SQL > alter table OM_INSTANCE storage (buffer_pool recycle);

SQL > select segment_name, buffer_pool from dba_segments;

SQL > SELECT NAME, 1 - (PHYSICAL_READS / (DB_BLOCK_GETS + CONSISTENT_GETS))

FROM V$BUFFER_POOL_STATISTICS order by name;

Redo log Buffer:

V$sysstat

Statspack

AWR report

ADDM report

SQL > SELECT SUM(value) "Redo Buffer Waits" FROM v$sysstat


WHERE name = 'redo log space wait time';

SQL > Select (r.value/e.value)*100 from v$sysstat r, v$sysstat e

Where r.name = 'redo buffer allocation retries' and e.name = 'redo entries';

SQL > alter system set log_buffer = scope=spfile;

SQL > shutdown immediate;

SQL > startup;

Checkpoint tuning:

LOG_CHECKPOINT_TIMEOUT = <no of sconds>

LOG_CHECKPOINT_INTERVAL = <no of OS blocks>

LOG_CHECKPOINTS_TO_ALERT = <TRUE / FALSE >

Thread 1 advanced to log sequence 248

Current log# 2 seq# 248 mem# 0: /prod1/oradata/logs/redologs02.log

Thread 1 cannot allocate new log, sequence 249

Checkpoint not complete

SQL > select name,value from v$sysstat where name like 'background checkpoint%'

NAME VALUE

Background checkpoint – start 30

Background checkpoint completed 27

SQL> select event, time_waited from v$system_event

where event = 'log file switch (checkpoint incomplete)';

SQL > alter system set FAST_START_MTTR_TARGET=120;

SQL > SELECT TARGET_MTTR, ESTIMATED_MTTR FROM V$INSTANCE_RECOVERY;

TARGET_MTTR ESTIMATED_MTTR

214 12


SQL > alter system set FAST_START_MTTR_TARGET=214;

SQL > SELECT OPTIMAL_LOGFILE_SIZE FROM V$INSTANCE_RECOVERY;

OPTIMAL_LOGFILE_SIZE

180

SQL> select event, time_waited from v$system_event

where event = 'log file switch (archiving needed)';

SQL > alter system set log_archive_max_processes=5;

Shared Pool Tuning:

V$librarycache

V$rowcache

Statspack Report

AWR Report

ADDM Report

SQL > select ( 1 - ( sum ( pinhits ) / sum( pins ) ) ) *100 from v$librarycache;

SQL > select ( 1 - ( sum ( reloads ) / sum ( pins ) ) ) * 100 from v$librarycache;

SQL > select ( 1 - ( sum ( getmiss ) / sum ( gets ) ) ) * 100 from v$rowcahce;

Solutions:

Method 1:

SQL > SELECT shared_pool_size_for_estimate ,shared_pool_size_factor,

estd_lc_time_saved, estd_lc_time_saved_factor FROM V$SHARED_POOL_ADVICE;

Size Size factor Time Saved Time Saved Factor

96 .8571 289654 .89

112 1 289660 .97

128 1.1429 289665 1


144 1.2857 289704 1.0002

160 1.4286 289707 1.0002

176 1.5714 289708 1.0002

SQL > alter system set shared_pool_size = 128M;

Method 2:

SQL > alter system set shared_pool_reserved_size=<10% of shared_pool size>

SQL > @?/rdbms/admin/dbmspool.sql

SQL > execute dbms_shared_pool.keep('PNAME');

SQL > execute dbms_shared_pool.keep('DBMS_DDL');

SQL > select name,sharable_mem,executions from v$db_object_cache where kept='YES';

SQL > execute dbms_shared_pool.unkeep('PNAME');

SQL > select (FREE_SPACE/1024/1024), REQUEST_MISSES, REQUEST_FAILURES

from V$SHARED_POOL_RESERVED;

FREE_SPACE REQUEST_MISSES REQUEST_FAILURES

47 0 0

Method 3:

SQL> ALTER SYSTEM FLUSH SHARED_POOL;

SQL > select * from emp where deptno=12;

SQL > select * from emp where deptno=13

SQL > select * from emp where deptno=14;

SQL> SELECT sql_text, executions FROM v$sqlarea

WHERE sql_text like 'select * from emp where deptno%'

SQL_TEXT EXECUTIONS


select * from emp where deptno=12 1



SQL> ALTER SYSTEM FLUSH SHARED_POOL;

SQL> variable a number

SQL> exec : a := 12

SQL> select * from emp where deptno = :a;

SQL> exec : a:= 13


SQL> exec : a := 14


SQL> SELECT sql_text, executions FROM v$sqlarea

WHERE sql_text like 'select * from emp where deptno%'

SQL_TEXT EXECUTIONS

select * from emp where deptno = :a 3

Method 4:

1. OPEN_CURSOR

SQL > alter system set OPEN_CURSOR=500;

SQL> select max(a.value) as Curr_Used_open_cur, p.value as max_open_cur

from v$sesstat a, v$statname b, v$parameter p

where a.statistic# = b.statistic# and b.name = 'opened cursors current'

and p.name= 'open_cursors' group by p.value;

Currently_Used_OPEN_CUR MAX_OPEN_CUR


490 500

2. CURSOR_SPACE_FOR_TIME = { TRUE | FALSE }

SQL > alter system set cursor_space_for_time=TRUE;

3. CURSOR_SHARING={ exact | similar | force }

SQL > alter system set cursor_sharing=similar;

Automatic Shared Memory Management (ASMM) :

SQL > show parameter SGA_MAX_SIZE;

SQL > show parameter SGA_TARGET;

SQL > alter system set sga_target=540M;

SQL > alter system set statistics_level=typical;

Auto Tuned parameter or Components

o DB_CACHE_SIZE

o SHARED_POOL_SIZE

o LARGE_POOL_SIZE

o JAVE_POOL_SIZE

o STREAMS_POOL_SIZE

Manual tuned parameter or Components

o LOG_BUFFER

o DB_KEEP_CACHE_SIZE

o DB_RECYCLE_CACHE_SIZE

SQL > select component, oper_type, initial_size/1024/1024 ,

TARGET_SIZE/1024/1024, FINAL_SIZE/1024/1024, status

from v$sga_resize_ops;


COMPONENT OPER_TYPE INITIAL TARGET FINAL STATUS

DEFAULT buffer cache SHRINK 160 148 148 COMPLETE

streams pool GROW 0 12 12 COMPLETE

SQL > select component, current_size/1024/1024, user_specified_size/1024/1024 ,

last_oper_type from v$sga_dynamic_components;

COMPONENT CURRENT_SIZE USER_SPECIFIED_SIZE TYPE

shared pool 80 80 STATIC

large pool 8 8 STATIC

java pool 48 48 STATIC

streams pool 12 12 GROW

DEFAULT buffer cache 48 24 SHRINK

SQL > select * from V$SGA_DYNAMIC_FREE_MEMORY;

Shows free memory available for future SGA memory resize.

SQL > select * from V$SGA_CURRENT_RESIZE_OPS

Currently SGA resize operations in progress.

SQL > select sga_size, (1 - estd_db_time_factor ) *100 “ESTD FACTOR”

from v$sga_target_advice order by sga_size;

SGA_SIZE ESTD FACTOR

171 -2,41

228 0

285 1,2

342 2,41

399 2,41


456 2,41

SQL > alter system set sga_target=342M;

Automatic PGA Memory Management (APMM) :

Session Data

Private SQL Area

Persistent Area Run Time Area

SQL work Area

Bitmap

Merge

Area

Create Bitmap

Area

Hash

Join

Area

Sort

Area

BITMAP_MERGE_AREA_SIZE = 1M

CREATE_BITMAP_AREA_SIZE = 8M

HASH_AREA_SIZE = (2 * SORT_AREA_SIZE)

SORT_AREA_SIZE = 64K

SQL > show parameter PGA_AGGREGATE_TARGET;

WORKAREA_SIZE_POLICY= { AUTO | MANUAL }

SQL > show parameter WORKAREA_SIZE_POLICY;

For OLTP : PGA_AGGREGATE_TARGET = (total_mem * 80%) * 20%

For DSS : PGA_AGGREGATE_TARGET = (total_mem * 80%) * 50%


SQL > SELECT * FROM V$PGASTAT;

NAME VALUE UNIT

aggregate PGA target parameter 41156608 bytes

cache hit percentage 98.15 percent

SQL > SELECT PGA_TARGET_FOR_ESTIMATE/1024/1024,

ESTD_PGA_CACHE_HIT_PERCENTAGE, ESTD_OVERALLOC_COUNT

FROM V$PGA_TARGET_ADVICE;

TARGET_MB CACHE_HIT_PERC ESTD_OVERALLOC_COUNT

63 60 367

125 84 30

250 95 0

375 100 0

500 100 0

600 100 0

Sort Area Tuning:

SQL > select ( 1 - ( d.value / m.value ) ) * 100 from v$sysstat d, v$sysstat m

where d.name='sorts (disk)' and m.name='sorts (memory)';

SORT_AREA_SIZE= (SORT_AREA_SIZE + (MAX_USED_BLOCKS * DB_BLOCK_SIZE))

SQL > show parameter SORT_AREA_SIZE;

SQL > show parameter DB_BLOCK_SIZE;

SQL > select MAX_USED_BLOCKS from v$sort_segment;

SQL > alter system set SORT_AREA_SIZE= ;

Disk IO Tuning:


SQL > select d.name, f.phyrds, f.readtim / f.phyrds,

f.phywrts, f.writetim / f.phywrts

from v$filestat f, v$database d where f.file# = d.file# ;

Oracle Locks:

Session 1

SQL> update emp set sal=sal+1 where empno=7566;

Session 2

SQL> update emp set sal=sal+1 where empno=7566;

Hanged….

Session 3

SQL> drop table emp ;

ORA-00054: resource busy and acquire with NOWAIT specified

SQL > select XIDUSN,object_id,session_id,oracle_username from v$locked_object;

XIDUSN OBJECT ID SESSION_ID ORACLE USERNAME

10 123 9 JOHN

15 147 12 ANTONY

0 123 15 BRITTO

0 147 21 STANLY

SQL > select sid,serial# from v$session where sid= 9;

SID SERIAL#

9 151

SQL > alter system kill session ‘ 9,151 ‘;


SQL > select sid,serial# from v$session where sid= 12;

SID SERIAL#

12 181

SQL > alter system kill session ‘ 12,181 ‘;

Data Migration from 8.1.7.0 to 10.2.0.1

Pre-check and post check :

SQL > select sum(bytes)/1024/1024/1024 from dba_data_files;

SQL > select sum(bytes)/1024/1024/1024 from dba_segments;

SQL > select tablespace_name,sum(bytes)/1024/1024 from dba_data_files

group by tablespace_name;

SQL > select tablespace_name,sum(bytes)/1024/1024 from dba_segments


SQL > select count(*) from dba_users where username not in ('SYS','SYSTEM');

SQL > select owner,count(*) from dba_tables where owner not in ('SYS','SYSTEM')

group by owner;

SQL > select owner,count(*) from dba_indexes where owner not in ('SYS','SYSTEM')

group by owner;

SQL > select owner,count(*) from dba_constraints where owner not in ('SYS','SYSTEM')

group by owner;

SQL > select owner,count(*) from dba_objects where owner not in ('SYS','SYSTEM')

group by owner;

SQL > select owner,count(*) from dba_objects where owner not in ('SYS','SYSTEM')


and status <> 'VALID' group by owner;

Target Sever 8.1.7.4.0

$ . oraenv

ORACLE_SID: prod

$ sqlplus “/ as sysdba”

SQL> shutdown immediate;

$ lsnrctl stop <Listener_name>


SQL > startup restrict

SQL > exit;

$ exp \'/ as sysdba \' FULL=Y FILE=full_prod.dmp BUFFER=157286400 STATISTICS=NONE

LOG=full_prod.log

$ scp –r full_prod.dmp [email protected]:/opt/oradata

Destination Server 10.2.0.1.0

$ . oraenv

ORACLE_SID: prod


SQL > select instance_name,status from v$instance;

SQL > select tablespace_name,sum(bytes)/1024/1024 from dba_data_files


mailto:[email protected]:/opt/oradata


$ imp \'/ as sysdba \' FULL=Y FILE=full_prod.dmp BUFFER=157286400 STATISTICS=NONE

LOG=full_prod.log

Upgrade 9.2.0.7.0 to 10.2.0.1.0

$ . oraenv

ORACLE_SID:prod


SQL > @/opt/oracle/product/10.2.0/rdbms/admin/utlu102i.sql

(1) Logfiles: [make adjustments in the current environment]

*************************************************************************

The existing log files are adequate. No changes are required.

(2) Tablespaces: [make adjustments in the current environment]

********************************************************************** *******

--> SYSTEM tablespace is adequate for the upgrade ... minimum required size: 532 MB

--> TEMP tablespace is adequate for the upgrade. ... minimum required size: 58 MB

--> USERS tablespace is adequate for the upgrade ... minimum required size: 10199 MB

--> UNDO1 tablespace is adequate for the upgrade ... minimum required size: 185 MB

(3) Update Parameters: [Update Oracle Database 10.2 init.ora or spfile]

**********************************************************************

WARNING: --> "streams_pool_size" is not currently defined and needs a value of at least


50331648

WARNING: --> "pga_aggregate_target" needs to be increased to at least 25165824

WARNING: --> "session_max_open_files" needs to be increased to at least 20

(4) Renamed Parameters: [Update Oracle Database 10.2 init.ora or spfile]

**********************************************************************

-- No renamed parameters found. No changes are required.

(5) Obsolete/Deprecated Parameters: [Update Oracle Database 10.2 init.ora or spfile]

**********************************************************************

--> "hash_join_enabled"

--> "log_archive_start"

--> "parallel_automatic_tuning"

(6) SYSAUX Tablespace:

[Create tablespace in the Oracle Database 10.2 environment]

**********************************************************************

--> New "SYSAUX" tablespace .... minimum required size for database upgrade: 500 MB

SQL > create tablespace sysaux

datafile ‘/opt/oradata/sysaux01.dbf’ size 500M

extent management local

segment space management auto

online;

SQL > create pfile from spfile;

SQL > select comp_name, status, version from dba_registry;


$ cp /opt/oracle/product/9.2.0/dbs/initprod.ora /opt/oracle/product/10.2.0/dbs


$ cp /opt/oracle/product/9.2.0/dbs/orapwprod /opt/oracle/product/10.2.0/dbs

# vi /etc/oratab

#prod:/opt/oracle/product/9.2.0:N

prod:/opt/oracle/product/10.2.0:N

$ lsnrctl stop <listener_name>

$ cd /opt/oracle/product/10.2.0/rdbms/admin/

$. oraenv

ORACLE_SID:prod

$ echo $ORACLE_HOME

/opt/oracle/product/10.2.0

$ sqlplus /nolog

SQL > conn sys/sys as sysdba

SQL > startup upgrade

SQL > spool upgrade.log

SQL > @?/rdbms/admin/catupgrd.sql

SQL > SPOOL OFF


SQL > startup

SQL >@utlrp.sql

SQL > @utlu102s.sql TEXT


BACKOUT FROM 10.2.0.1 to 9.2.0.7

$. oraenv

ORACLE_SID:prod



SQL > STARTUP DOWNGRADE

SQL > SPOOL downgrade. log

SQL > @/opt/oracle/product/10.2.0/rdbms/admin/catdwgrd.sql

SQL > SPOOL OFF

SQL > SHUTDOWN IMMEDIATE

# vi /etc/oratab

prod:/opt/oracle/product/9.2.0:N

# prod:/opt/oracle/product/10.2.0:N

$. oraenv

ORACLE_SID:prod

$ echo $ORACLE_HOME

/opt/oracle/product/9.2.0

$ sqlplus /nolog

SQL > CONNECT SYS/SYS_password AS SYSDBA

SQL > STARTUP MIGRATE

SQL > SPOOL reload.log

SQL > @/opt/oracle/product/10.2.0/rdbms/admin/catrelod.sql

SQL > SPOOL OFF

SQL > SHUTDOWN IMMEDIATE

SQL > STARTUP

SQL > @/opt/oracle/product/9.2.0/rdbms/admin/UTLRP.SQL



Apply CPU /PSU patches:

$. oraenv

ORACLE_SID : prod

SQL > sqlplus “/ as sysdba”

SQL > shutdown immediate

$ lsnrctl stop <Listener_name>

$ unzip p919293_1024_linux86.zip

$ cd p919293

$ opatch apply

$ opatch lsinventory

$ . oraenv

ORACLE_SID : prod

$ cd $ORACLE_HOME/rdbms/admin

$ sqlplus /nolog

SQL > CONNECT / AS SYSDBA

SQL > STARTUP

SQL > @?/rdbms/admin/catbundle.sql psu apply

SQL > @?/cpu/view_recompile/recompile_precheck_jan2008cpu.sql

SQL > shutdown immediate

SQL > CONNECT / AS SYSDBA

SQL > STARTUP UPGRADE

SQL > @?/cpu/view_recompile/view_recompile_jan2008cpu.sql

SQL > SHUTDOWN;

SQL > STARTUP;


SQL > @?/rdbms/admin/utlrp.sql

SQL > alter package schemaname.packagename compile;

SQL > select * from dba_registry_history;

Putty.exe Tool:


Xmanager Tool:


03 -oracle 10g performance tuning

Documents