"ORACLE 모니터링"의 두 판 사이의 차이

DB CAFE

2024년 1월 1일 (월) 23:41 기준 최신판

추천메뉴 바로가기

DBA { Oracle DBA 명령어 > DBA 초급 과정 > DBA 고급 과정 }
튜닝 { 오라클 튜닝 목록 }
모델링 { 데이터 모델링 가이드 }

DB 모니터링

1 세션 정보
2 SQL 찾기
3 SQL 과 커서 조회
4 락 모니터링
5 프로세스 정보
- 5.1 Oracle Process의 정보
- 5.2 오브젝트에 접속되어 있는 프로그램 조회
6 부하 발생 모니터링
7 I/O 부하 찾기
- 7.1 인스턴스 기동 이후 I/O 관련 대기현상 누적 값
- 7.2 전체 IO량(100%) 대비 사용자별 IO량
8 테이블스페이스
- 8.1 테이블스페이스 사용량
- 8.2 System 테이블스페이스에 비시스템 세그먼트 조회
9 CPU/메모리/사양
10 INDEX 정보 추출하기
11 ASH를 이용한 모니터링
12 AWR을 이용한 모니터링

1 세션 정보[편집]

1.1 User 별 Session 정보 조회[편집]

V$SESSION
V$PROCESS

SELECT S.USERNAME
     , S.SID
     , S.SERIAL#
     , P.SPID
     , S.OSUSER
     , S.MACHINE
     , S.PROGRAM
     , TO_CHAR(S.LOGON_TIME, 'MM/DD HH24:MI') "LOGON_TIME"
     , ROUND(S.LAST_CALL_ET / 60) "IDLE"
  FROM V$SESSION S
     , V$PROCESS P
 WHERE S.PADDR = P.ADDR 
   AND S.USERNAME LIKE UPPER('&DBUSER%') 
 ORDER BY 9;

1.2 세션 죽이기(SID,SERAIL#)[편집]

ALTER SYSTEM KILL SESSION '8,4093' immediate;
-- RAC 
ALTER SYSTEM KILL SESSION '8,4093,@1' immediate;

1.3 현재 TOP 세션 gv$session[편집]

select  inst_id||'_'||sid||' '||serial# inst_sid_ser,
    username||case when regexp_substr(program,' \(...') <> ' (TNS' then regexp_substr(program,' \(.+') end username,
    sql_id sql_id, 
    round((sysdate-sql_exec_start)*24*3600,1) sql_dT,
        last_call_et call_dT,
    case state when 'WAITING' then round(wait_time_micro/1000000,2) else round(time_since_last_wait_micro/1000000,2) end W_dT,
        decode(state,'WAITING',event,'CPU') event, 
    service_name||' '||substr(module,1,20)||' '||ACTION serv_mod_action,  
          nullif(row_wait_obj#,-1) obj#,decode(taddr,null,null,'NN') tr
from gv$session
where ((state='WAITING' and wait_class<>'Idle') or (state<>'WAITING' and status='ACTIVE'))
      --and audsid != to_number(sys_context('USERENV','SESSIONID')) -- this is clean but does not work on ADG so replaced by following line
      and (machine,port) <> (select machine,port from v$session where sid=sys_context('USERENV','SID')) --workaround for ADG
order by inst_id,sql_id

1.4 오래 걸리는 세션 정보[편집]

1.5 현재 작업중인 세션 정보 V$SESSION_LONGOPS[편집]

V$SESSION_LONGOPS
V$SESSION

SELECT B.USERNAME, A.SID, B.OPNAME, B.TARGET
     , ROUND(B.SOFAR*100/B.TOTALWORK,0) || '%' AS "%DONE", B.TIME_REMAINING
     , TO_CHAR(B.START_TIME,'YYYY/MM/DD HH24:MI:SS') START_TIME
  FROM V$SESSION_LONGOPS B
     , V$SESSION A
 WHERE A.SID = B.SID      
 ORDER BY 6 DESC;

1.6 세션당 리소스 사용률 gv$sessmetric[편집]

gv$sessmetric

select * from (
  select inst_id, session_id sid, to_char(begin_time,'hh24:mi:ss') begTime, round(intsize_csec/100,0) D_sec,  cpu,          
         physical_reads PhyReads, logical_reads LogicalReads, pga_memory, hard_parses, soft_parses 
  from gv$sessmetric
  order by cpu+physical_reads desc
)
where rownum<20

1.7 대기중인 세션[편집]

v$session_wait

select * 
  from v$session_wait;

1.8 1시간 이상 idle 상태인 세션[편집]

V$SESSION

SELECT SID
     , SERIAL#
     , USERNAME
     , TRUNC(LAST_CALL_ET / 3600, 2) || ' HR' LAST_CALL_ET
  FROM V$SESSION 
 WHERE LAST_CALL_ET > 3600 
   AND USERNAME IS NOT NULL;

1.9 Active Session 중 Idle Time이 긴 작업[편집]

V$SESSION
V$PROCESS

SELECT VS.SID || ',' || VS.SERIAL# " SID"
     , VP.SPID
     , VS.MACHINE
     , VS.PROGRAM
     , VS.MODULE
     , VS.STATUS
     , TO_CHAR(VS.LOGON_TIME, 'MM/DD HH24:MI') LOGIN_TIME
     , ROUND(VS.LAST_CALL_ET / 60) "IDLE"
  FROM V$SESSION VS
     , V$PROCESS VP
 WHERE VS.STATUS = 'ACTIVE' 
   AND VS.SID NOT IN (1, 2, 3, 4, 5, 6, 7) 
   AND VS.PADDR = VP.ADDR ORDER BY 8;

1.10 사용자 세션 중에서 2시간 이상 idle 상태가 지속되는 세션 kill[편집]

V$SESSION
V$PROCESS

SET PAGESIZE 0 SPOOL KILLIDLE3.SQL
SELECT DISTINCT '!KILL -9 ' || B.SPID, 'ALTER SYSTEM KILL SESSION '''|| A.SID || ',' || A.SERIAL# || ''';' 
  FROM V$SESSION A
     , V$PROCESS B
 WHERE A.PADDR IN (SELECT S.PADDR
                     FROM V$SESSION S
                    WHERE STATUS = 'INACTIVE'
                    GROUP BY  S.PADDR
                   HAVING MIN(ROUND(LAST_CALL_ET / 60)) > 120)
   AND A.PADDR = B.ADDR AND A.STATUS = 'INACTIVE';
SPOOL OFF

1.11 트랜젝션 중인 세션[편집]

v$transaction
gv$fast_start_transactions

USED_UBLK
- Number of undo blocks used , 사용된 언두 블럭 수
USED_UREC
- Number of undo records used , 사용된 언두 레코드 수

select sid, serial#, username, taddr, used_ublk, used_urec
  from v$transaction t
     , v$session s
where t.addr = s.taddr;

select inst_id,addr,start_time,used_ublk,xid 
  from gv$transaction;

select inst_id,state,undoblocksdone,undoblockstotal,xid 
 from gv$fast_start_transactions;

SELECT sess.sid, sess.status, sess.username, machine, sql_id, prev_sql_id, trans.USED_UBLK, trans.start_date
from gv$session sess
   , gv$transaction trans
WHERE sess.taddr=trans.addr and sess.inst_id=trans.inst_id;

1.12 연결되어 있는 OS 사용자 및 프로그램 조회[편집]

V$SESSION

SELECT SID
     , SERIAL#
     , OSUSER
     , SUBSTRB(USERNAME, 1, 10) AS USER_NAME
     , SUBSTRB(PROGRAM, 1, 30) AS PROGRAM_NAME
     , STATUS
     , TO_CHAR(LOGON_TIME, 'YYYY/MM/DD HH:MI') AS LOGON_TIME
  FROM V$SESSION 
 WHERE TYPE!= 'BACKGROUND' AND STATUS = 'ACTIVE';

1.13 Session별 사용 명령어[편집]

V$SESSION SESS
V$SESSTAT STAT
V$STATNAME NAME
V$PROCESS PROC

SELECT SESS.SID
     , SESS.SERIAL#
     , SUBSTR(SESS.USERNAME, 1, 10) "USER NAME"
     , SUBSTR(OSUSER, 1, 11) "OS USER"
     , SUBSTR(SESS.MACHINE, 1, 15) "MACHINE NAME"
     , STATUS
     , UPPER(
            DECODE(NVL(COMMAND, 0)
              , 0, '---'
              , 1, 'CREATE TABLE'
              , 2, 'INSERT -'
              , 3, 'SELECT -'
              , 4, 'CREATE CLUST'
              , 5, 'ALTER CLUST'
              , 6, 'UPDATE -'
              , 7, 'DELETE -'
              , 8, 'DROP -'
              , 9, 'CREATE INDEX'
              , 10, 'DROP INDEX'
              , 11, 'ALTER INDEX'
              , 12, 'DROP TABLE'
              , 13, 'CREATE SEQ'
              , 14, 'ALTER SEQ'
              , 15, 'ALTER TABLE'
              , 16, 'DROP SEQ'
              , 17, 'GRANT'
              , 18, 'REVOKE'
              , 19, 'CREATE SYN'
              , 20, 'DROP SYN'
              , 21, 'CREATE VIEW'
              , 22, 'DROP VIEW'
              , 23, 'VALIDATE IX'
              , 24, 'CREATE PROC'
              , 25, 'ALTER PROC'
              , 26, 'LOCK TABLE'
              , 27, 'NO OPERATION'
              , 28, 'RENAME'
              , 29, 'COMMENT'
              , 30, 'AUDIT'
              , 31, 'NOAUDIT'
              , 32, 'CREATE DBLINK'
              , 33, 'DROP DB LINK'
              , 34, 'CREATE DATABASE'
              , 35, 'ALTER DATABASE'
              , 36, 'CREATE RBS'
              , 37, 'ALTER RBS'
              , 38, 'DROP RBS'
              , 39, 'CREATE TABLESPACE'
              , 40, 'ALTER TABLESPACE'
              , 41, 'DROP TABLESPACE'
              , 42, 'ALTER SESSION'
              , 43, 'ALTER USER'
              , 44, 'COMMIT'
              , 45, 'ROLLBACK'
              , 47, 'PL/SQL EXEC'
              , 48, 'SET TRANSACTION'
              , 49, 'SWITCH LOG'
              , 50, 'EXPLAIN'
              , 51, 'CREATE USER'
              , 52, 'CREATE ROLE'
              , 53, 'DROP USER'
              , 54, 'DROP ROLE'
              , 55, 'SET ROLE'
              , 56, 'CREATE SCHEMA'
              , 58, 'ALTER TRACING'
              , 59, 'CREATE TRIGGER'
              , 61, 'DROP TRIGGER'
              , 62, 'ANALYZE TABLE'
              , 63, 'ANALYZE INDEX'
              , 69, 'DROP PROCEDURE'
              , 71, 'CREATE SNAP LOG'
              , 72, 'ALTER SNAP LOG'
              , 73, 'DROP SNAP LOG'
              , 74, 'CREATE SNAPSHOT'
              , 75, 'ALTER SNAPSHOT'
              , 76, 'DROP SNAPSHOT'
              , 85, 'TRUNCATE TABLE'
              , 88, 'ALTER VIEW'
              , 91, 'CREATE FUNCTION'
              , 92, 'ALTER FUNCTION'
              , 93, 'DROP FUNCTION'
              , 94, 'CREATE PACKAGE'
              , 95, 'ALTER PACKAGE'
              , 96, 'DROP PACKAGE'
              , 46, 'SAVEPOINT'
                  )
          )
            COMMAND
     , SESS.PROCESS "C.PROC"
     , PROC.SPID "S.PROC"
     , TO_CHAR(SESS.LOGON_TIME, 'YYYY-MM-DD HH24:MI')
  FROM V$SESSION SESS
     , V$SESSTAT STAT
     , V$STATNAME NAME
     , V$PROCESS PROC
 WHERE SESS.SID = STAT.SID 
   AND STAT.STATISTIC# = NAME.STATISTIC# 
   AND SESS.USERNAME IS NOT NULL 
   AND NAME.NAME = 'RECURSIVE CALLS' 
   AND SESS.PADDR = PROC.ADDR 
 ORDER BY 3, 1, 2;

2 SQL 찾기[편집]

2.1 SQL 통계 정보확인하기[편집]

2.2 Module LEVEL 통계[편집]

SELECT *
    FROM
        (SELECT module ,
                count(*) sql_cnt ,
                sum(executions) executions ,
                round(avg(buffer_gets/executions)) "lio(avg)" ,
                round(avg(disk_reads/executions)) "pio(avg)" ,
                round(avg(rows_processed/executions)) "rows(avg)" ,
                round(avg(elapsed_time/executions/1000000),2) "elapsed(avg)" ,
                count(CASE WHEN elapsed_time/executions/1000000>=3 THEN 1 END) bad_sql ,
                round(max(elapsed_time/executions/1000000),2) "elapsed(max)"
         FROM v$sql
         WHERE executions > 0
         GROUP BY module)
ORDER BY "lio(avg)" * executions DESC ;

2.3 SCHEMA LEVEL 통계[편집]

SELECT * --9i
FROM
    ( SELECT
         (SELECT username
          FROM dba_users
          WHERE user_id = parsing_schema_id) AS SCHEMA_NAME ,
             count(*) sql_cnt ,
             sum(executions) executions ,
             round(avg(buffer_gets/executions)) "lio(avg)" ,
             round(avg(disk_reads/executions)) "pio(avg)" ,
             round(avg(rows_processed/executions)) "rows(avg)" ,
             round(avg(elapsed_time/executions/1000000),2) "elapsed(avg)" ,
             count(CASE WHEN elapsed_time/executions/1000000>=10 THEN 1 END) bad_sql ,
             round(max(elapsed_time/executions/1000000),2) "elapsed(max)"
     FROM v$sqlarea
     WHERE executions > 0
     GROUP BY parsing_schema_id)
ORDER BY "lio(avg)" * executions DESC ;

SELECT * --10g
FROM
    (SELECT parsing_schema_name ,
            count(*) sql_cnt ,
            sum(executions) executions ,
            round(avg(buffer_gets/executions)) "lio(avg)" ,
            round(avg(disk_reads/executions)) "pio(avg)" ,
            round(avg(rows_processed/executions)) "rows(avg)" ,
            round(avg(elapsed_time/executions/1000000),2) "elapsed(avg)" ,
            count(CASE WHEN elapsed_time/executions/1000000>=10 THEN 1 END) bad_sql ,
            round(max(elapsed_time/executions/1000000),2) "elapsed(max)"
     FROM v$sqlarea
     WHERE executions > 0
     GROUP BY parsing_schema_name)
ORDER BY "lio(avg)" * executions DESC ;

2.4 DB Time 분석하기[편집]

SELECT STAT_NAME ,
       VALUE,
       ROUND(VALUE/
                 ( SELECT VALUE
                  FROM V$SYS_TIME_MODEL
                  WHERE STAT_NAME='DB TIME' ) *100 , 2) TIME_RATIO
FROM V$SYS_TIME_MODEL
WHERE STAT_NAME NOT IN ( 'BACKGROUND ELAPSED TIME' ,
                         'BACKGROUND CPU TIME' ,
                         'RMAN CPU TIME (BACKUP/RESTORE)' ,
                         'HARD PARSE ELAPSED TIME' ,
                         'HARD PARSE (SHARING CRITERIA) ELAPSED TIME' ,
                         'HARD PARSE (BIND MISMATCH) ELAPSED TIME' ,
                         'FAILED PARSE ELAPSED TIME' ,
                         'FAILED PARSE (OUT OF SHARED MEMORY) ELAPSED TIME' ,
                         'DB CPU')
ORDER BY TIME_RATIO DESC ;

2.5 SQL 추출하기 (10g)[편집]

SELECT *
    FROM
        ( SELECT rownum cnt,
                 t1.*
         FROM
             ( SELECT parsing_schema_name SCHEMA,    --> 1
                                          module,    --> 2
                                          sql_id,    --> 3
                                          hash_value,--> 4
                                          substr(sql_text,1,100) substr_sqltext, --> 5
                                          executions,  --> 6
                                          buffer_gets, --> 7 (전체 I/O 처리량)
                                          disk_reads,  --> 8
                                          rows_processed, --> 9
                                          round(buffer_gets/executions,1) lio, --> 10 (1회 수행 당 I/O)
                                          round(elapsed_time/executions/1000000,1) elapsed_sec, --> 11 (1회 수행 당 Elapsed)
                                          round(cpu_time/executions/1000000,1) cpu_sec, --> 12 (1회 수행 당 CPU)
                                          round(elapsed_time/tot_elapsed*100,1) ratio_elapsed, --> 13 (전체 대비 Elapsed )
                                          round(cpu_time/tot_cpu*100,1) ratio_cpu --> 14 (전체 대비 CPU 사용률)
                            FROM v$sqlarea s,
                  (SELECT sum(elapsed_time) tot_elapsed,
                          sum(cpu_time) tot_cpu
                   FROM v$sqlarea) t
              WHERE s.executions > 0
                  AND parsing_schema_name NOT IN ('SYS','SYSTEM')
                  AND ((module NOT LIKE 'TOAD%'
                        AND module NOT LIKE 'SQL De%'
                        AND module NOT LIKE 'Orange%'
                        AND module NOT LIKE 'PL/SQL%'
                        AND module NOT LIKE 'plsqldev.exe')
                       OR (module IS NULL))
              ORDER BY 14 DESC ) t1
         WHERE rownum <= 50) WHERE cnt >= 1 
;

2.6 엑사 DB[편집]

2.7 Exa . Smart Scan으로 수행되지 않고,Interconnect Bytes가 높은 SQL[편집]

WITH Cell_Mon_SQL AS
        ( SELECT sql_id,
                 io_cell_offload_eligible_bytes AS coeb,
                 io_interconnect_bytes AS ib,
                 io_cell_offload_returned_bytes AS corb,
                 io_cell_uncompressed_bytes AS cub,
                 optimized_phy_read_requests AS oprr,
                 physical_read_requests AS prr,
                 physical_read_bytes AS prb,
                 sql_text
         FROM v$sqlarea)
    SELECT /*** Smart Scan이 되지 않으면서, Interconnect Bytes가 높은 SQL ***/ *
    FROM
        ( SELECT sql_id ,
                 decode(coeb,0,'No','Yes') AS "Offload",
                 round(prb/1024/1024/1024,3) AS "Physical_Bytes(GB)",
                 round(coeb/1024/1024/1024,3) AS "Offload_Bytes(GB)",
                 round(ib/1024/1024/1024,3) AS "Interconnect_Bytes(GB)",
                 round(oprr*8192/1024/1024/1024,3) AS "Cell_Flash_Cache(GB)",
                 round(corb/1024/1024/1024,3) AS "Returned_Bytes(GB)",
                 round(decode(coeb,0,0,100*(coeb-ib)/coeb),2) AS "Saved_IO(%)",
                 sql_text
         FROM Cell_Mon_SQL
         WHERE coeb = 0
         ORDER BY ib DESC) WHERE rownum <= 10 ;

2.8 Smart Scan으로 수행되지 않고,Offload Returned Bytes가 높은 SQL[편집]

Reverting TO Block Shipping으로 수행된 SQL로 Smart Scan으로 동작할 수 있는 SQL이지만,Direct I/O로 동작하지 않는 경우 (FULL TABLE Scan도 cell single block physical read로 수행됨.)

Chained Row가 있는 블록들을 읽을 때 발생. : Smart Scan을 통해서 Chained Row가 있는 블록을 읽을 때 각 블록이 다른 STORAGE Cell에 있는 경우 STORAGE Cell 간 통신할 수 없기 때문에 Buffer Cache로 전체 블록을 전송한 후 처리하도록 Block Shipping Mode로 변경

읽기 일관성 모드로 변경 중인 데이터를 가진 Block에 대해서 Smart Scan을 중지하고, Single Block Read를 수행하는 Block Shipping Mode로 변경

WITH Cell_Mon_SQL AS
    ( SELECT sql_id,
             io_cell_offload_eligible_bytes AS coeb,
             io_interconnect_bytes AS ib,
             io_cell_offload_returned_bytes AS corb,
             io_cell_uncompressed_bytes AS cub,
             optimized_phy_read_requests AS oprr,
             physical_read_requests AS prr,
             physical_read_bytes AS prb,
             sql_text
     FROM v$sqlarea)
SELECT *
FROM
    ( SELECT sql_id ,
             decode(coeb,0,'No','Yes') AS "Offload",
             round(prb/1024/1024/1024,3) AS "Physical_Bytes(GB)",
             round(coeb/1024/1024/1024,3) AS "Offload_Bytes(GB)",
             round(ib/1024/1024/1024,3) AS "Interconnect_Bytes(GB)",
             round(oprr*8192/1024/1024/1024,3) AS "Cell_Flash_Cache(GB)",
             round(corb/1024/1024/1024,3) AS "Returned_Bytes(GB)",
             round(decode(coeb,0,0,100*(coeb-ib)/coeb),2) AS "Saved_IO(%)",
             sql_text
     FROM Cell_Mon_SQL
     WHERE coeb = 0
     ORDER BY corb DESC)
WHERE rownum <= 10 ;

2.9 Full Table Scan 처리가 있는 SQL문 정보 추출하기(10g)[편집]

SELECT *
FROM
    (SELECT rownum cnt,
            t1.*
     FROM
         (SELECT parsing_schema_name,
                 --> 1

                 module,
                 --> 2

                 sql_id,
                 --> 3

                 hash_value,
                 --> 4

                 substr(sql_text,1,100) substr_sqltext,
                 --> 5

                 executions,
                 --> 6

                 buffer_gets,
                 --> 7 (전체 I/O 처리량)

                 disk_reads,
                 --> 8

                 rows_processed,
                 --> 9

                 round(buffer_gets/executions,1) lio,
                 --> 10 (1회 수행 당 I/O)

                 round(elapsed_time/executions/1000000,1) elapsed_sec,
                 --> 11 (1회 수행 당 Elapsed)

                 round(cpu_time/executions/1000000,1) cpu_sec,
                 --> 12 (1회 수행 당 CPU)

                 round(elapsed_time/tot_elapsed*100,1) ratio_elapsed,
                 --> 13 (전체 대비 Elapsed )

                 round(cpu_time/tot_cpu*100,1) ratio_cpu --> 14 (전체 대비 CPU 사용률)

          FROM
              (SELECT /*+ leading(x) no_merge(x) use_hash(x s) */ s.*
               FROM v$sqlarea s,
                   ( SELECT DISTINCT hash_value
                    FROM v$sql_plan
                    WHERE OPERATION = 'TABLE ACCESS'
                        AND OPTIONS = 'FULL' --and object_owner like :b1||'%'
 ) x
               WHERE x.hash_value = s.hash_value) s,
              (SELECT sum(elapsed_time) tot_elapsed,
                      sum(cpu_time) tot_cpu
               FROM v$sqlarea) t
          WHERE executions > 0
              AND parsing_schema_name NOT IN ('SYS',
                                              'SYSTEM')
              AND ((module NOT LIKE 'TOAD%'
                    AND module NOT LIKE 'SQL De%'
                    AND module NOT LIKE 'Orange%'
                    AND module NOT LIKE 'PL/SQL%'
                    AND module NOT LIKE 'plsqldev.exe')
                   OR (module IS NULL))
          ORDER BY 10 DESC ) t1
     WHERE rownum <= 50)
WHERE cnt >= 1

2.10 하드 파싱(leteral,리터럴) SQL 찾기[편집]

SELECT *
    FROM
        ( SELECT
             ( SELECT parsing_schema_name
              FROM v$sqlarea
              WHERE sql_id=a.max_sql_id ) SCHEMA,
             ( SELECT MODULE
              FROM v$sqlarea
              WHERE sql_id=a.max_sql_id ) MODULE, a.literal_sql_cnt, a.execution_cnt, a.plan_cnt, a.max_sql_id,
             ( SELECT sql_fulltext
              FROM v$sqlarea
              WHERE sql_id = a.max_sql_id ) sql_text
         FROM
             ( SELECT s.force_matching_signature,
                      COUNT(s.exact_matching_signature) literal_sql_cnt,
                      SUM(s.executions) execution_cnt,
                      MAX(s.sql_id) max_sql_id,
                      COUNT(DISTINCT s.plan_hash_value) plan_cnt
              FROM v$sql s
              GROUP BY s.force_matching_signature HAVING COUNT(s.exact_matching_signature) >= 2 ) a
         ORDER BY 3 DESC ) WHERE ROWNUM <= 50
;

또는

SELECT /*+ leading(h) no_merge(h) use_nl(h s) */
       s.parsing_schema_name        SCHEMA                  , --> 1
       s.module                                             , --> 2
       s.sql_id                                             , --> 3
       s.hash_value                                         , --> 4
       SUBSTR (s.sql_text, 1, 100)       substr_sqltext     , --> 5
       s.executions                                         , --> 6
       s.buffer_gets                                        , --> 7 (전체 I/O 처리량)
       s.disk_reads                                         , --> 8
       s.rows_processed                                     , --> 9
       ROUND (s.buffer_gets / s.executions, 1)                lio     , --> 10 (1회 수행 당 I/O)
       ROUND (s.elapsed_time / s.executions / 1000000, 1)     elapsed_sec     , --> 11 (1회 수행 당 Elapsed)
       ROUND (s.cpu_time / s.executions / 1000000, 1)         cpu_sec --> 12 (1회 수행 당 CPU)
  FROM v$sqlarea  s
     , (SELECT *
          FROM (  SELECT *
                    FROM (  SELECT s.force_matching_signature
                                 , COUNT (s.exact_matching_signature)    literal_sql_cnt
                                 , SUM (s.executions)                    execution_cnt
                                 , MAX (s.sql_id)                        max_sql_id
                                 , COUNT (DISTINCT s.plan_hash_value)    plan_cnt
                              FROM v$sql s
                             WHERE s.executions > 0
                                   AND parsing_schema_name NOT IN
                                           ('SYS', 'SYSTEM')
                                   AND (   (    module NOT LIKE 'TOAD%'
                                            AND module NOT LIKE 'SQL De%'
                                            AND module NOT LIKE 'Orange%'
                                            AND module NOT LIKE 'PL/SQL%'
                                            AND module NOT LIKE 'plsqldev.exe')
                                        OR (module IS NULL))
                          GROUP BY s.force_matching_signature
                            HAVING COUNT (s.exact_matching_signature) >= 2) a
                ORDER BY 3 DESC)
         WHERE ROWNUM <= 50) h
 WHERE h.max_sql_id = s.sql_id;

SQL 수행 이력 추출하기

DBA_HIST_SQLTEXT : SQL TEXT가 CLOB으로 저장
DBA_HIST_SQL_PLAN : SQL문의 수행 PLAN이 저장되어 있으므로 플랜 변경등의 이력을 조회해 볼때 용이하다.
DBA_HIST_SQLBIND : SQL문을 수행한 BIND VALUE를 저장하고 있는데 시간정보가 같이 있으므로 조회패턴을 분석하기 용이하다. DBA_HIST_SQLSTAT : V$SQL 정보와 같은 SQL 수행이력이 있는 VIEW로 각 SNAP_ID 마다의 Snap Shot 정보를 담고 있으므로 수행이력 패턴을 조회할 수 있다.

2.11 특정 SQL 수행내역 확인하기[편집]

-- SQL TEXT

SELECT module,
       sql_fulltext
FROM v$sqlarea
WHERE hash_value = :hash_value -- 수행내역
;

v$sqlarea

SELECT EXECUTIONS "Executions(total)",
           round(DISK_READS/executions,2) "Disk_reads(one)",
           round(BUFFER_GETS/executions,0) "Buffer_gets(one)",
           round(ROWS_PROCESSED/EXECUTIONS,0) "Rows(one)",
           round((ELAPSED_TIME/EXECUTIONS)/1000000,2) "Elapsed_time(one)",
           round((CPU_TIME/EXECUTIONS)/1000000,2) "Cpu_time(one)"
  FROM v$sqlarea sa 
 WHERE hash_value = :hash_value -- 바인드 변수값
;

2.12 SQL_ID를 이용하여 SQL TEXT 추출하기[편집]

DBA_HIST_SQLTEXT

SELECT *
  FROM DBA_HIST_SQLTEXT
 WHERE sql_id = :sql_id 
;

2.13 SQL_ID를 이용하여 SQL문의 수행내역 추출하기[편집]

dba_hist_sqlstat

SELECT snap_id,
           EXECUTIONS_TOTAL "Executions(total)",
           round(DISK_READS_TOTAL/executions_total,2) "Disk_reads(one)",
           round(BUFFER_GETS_TOTAL/executions_total,0) "Buffer_gets(one)",
           round(ROWS_PROCESSED_TOTAL/EXECUTIONS_TOTAL,0) "Rows(one)",
           round((ELAPSED_TIME_TOTAL/EXECUTIONS_TOTAL)/1000000,2) "Elapsed_time(one)",
           round((CPU_TIME_TOTAL/EXECUTIONS_TOTAL)/1000000,2) "Cpu_time(one)"
    FROM dba_hist_sqlstat WHERE sql_id = :sql_id
ORDER BY snap_id DESC 
;

2.14 프로세스 ID 를 이용하여 SQL 찾기[편집]

v$process
v$session
v$sqltext

select c.sql_text
     , b.SID
     , b.SERIAL#
     , b.machine
     , b.OSUSER
     , b.logon_time --이 쿼리를 호출한 시간
  from v$process a, v$session b, v$sqltext c
 where a.addr = b.paddr
   and b.sql_hash_value = c.hash_value
   and a.spid = '1708032' --1912870/
 order by c.PIECE

2.15 바인드 변수 값[편집]

바인드 변수 값의 이력을 가지고 있는 뷰 조회로 SQL문의 조회 패턴 분석 가능
v$sql_bind_capture

SELECT name,
           sql_id,
           to_char(LAST_CAPTURED,'yyyymmdd hh24:mi:ss') LAST_CAPTURED,
           datatype,
           VALUE_STRING
    FROM v$sql_bind_capture 
   WHERE hash_value = :hash_value -- or sql_id=:sql_id;

2.16 바인드 변수 값2[편집]

gv$sql_monitor

select BINDS_XML from 
* gv$sql_monitor 
 where sql_id='&1';

2.17 바인드 변수 값3[편집]

dba_hist_sqlbind

SELECT name,
           to_char(LAST_CAPTURED,'yyyymmdd hh24:mi:ss') LAST_CAPTURED,
           datatype,
           VALUE_STRING
    FROM dba_hist_sqlbind 
   WHERE sql_id = :sql_id -- PLAN
;

2.18 플랜 정보 조회[편집]

V$SQL_PLAN

SELECT --cardinality , cost,
 LPAD('.', 1*(LEVEL-1),'.')|| OPERATION || decode(OPTIONS,NULL,NULL,' (') || OPTIONS || decode(OPTIONS,NULL,NULL,')') || decode(object_owner,NULL,NULL,' :')||object_owner || decode(object_name,NULL,NULL,'.') || OBJECT_NAME || decode(other_tag,NULL,NULL,'(')||other_tag||decode(other_tag,NULL,NULL,')') || DECODE(ACCESS_PREDICATES,NULL,NULL,' [AP] '||ACCESS_PREDICATES) ||DECODE(FILTER_PREDICATES,NULL,NULL,' [FP] '||FILTER_PREDICATES) "operation"
    FROM
        (SELECT /*+ NO_MERGE */ *
         FROM V$SQL_PLAN
         WHERE HASH_VALUE = :hash_value
             AND CHILD_NUMBER = 0) CONNECT BY
    PRIOR id = parent_id START WITH id=0
ORDER BY ID 
;

2.19 SQL_ID를 이용하여 SQL문의 PLAN 추출하기[편집]

DBA_HIST_SQL_PLAN

SELECT LPAD(' ',4*(depth-1))||OPERATION||DECODE(OTHER_TAG,NULL,'','*')|| DECODE(OPTIONS,NULL,'',' ('||OPTIONS||')')|| DECODE(OBJECT_NAME,NULL,'',' OF '|| OBJECT_NAME||'')|| DECODE(OBJECT_TYPE,NULL,'',' ('||OBJECT_TYPE||')')|| DECODE(ID,0,DECODE(OPTIMIZER,NULL,'',' Optimizer='||OPTIMIZER))|| DECODE(COST,NULL,'',' (Cost='||COST|| DECODE(CARDINALITY,NULL,'',' Card='||CARDINALITY)|| DECODE(BYTES,NULL,'',' Bytes='||BYTES)||')')|| DECODE(ACCESS_PREDICATES,NULL,' ',' [AP] = '||ACCESS_PREDICATES)|| DECODE(ACCESS_PREDICATES,NULL,' ',' [FP] = '||FILTER_PREDICATES) AS PLAN_OUTPUT
    FROM DBA_HIST_SQL_PLAN WHERE sql_id = :sql_id
ORDER BY ID,
         POSITION 
;

3 SQL 과 커서 조회[편집]

3.1 현재 커서 수 확인[편집]

V$OPEN_CURSOR
v$session_wait
v$transaction

3.2 sid별 열린커셔[편집]

SELECT sid, count(sid) cursor
  FROM V$OPEN_CURSOR
 WHERE user_name = 'dbcafe'
 GROUP BY sid
 ORDER BY cursor DESC;

3.3 sql 별 열린 커서[편집]

SELECT sql_text, count(sid) cnt
  FROM v$OPEN_CURSOR
 GROUP BY sql_text
 ORDER BY cnt DESC

4 락 모니터링[편집]

4.1 V$LOCK 을 사용한 잠금 경합 모니터링[편집]

V$SESSION
DBA_OBJECTS
V$LOCK

SELECT S.USERNAME, S.SID, S.SERIAL#, S.LOGON_TIME,
    DECODE(L.TYPE, 'TM', 'TABLE LOCK',
                   'TX', 'ROW LOCK',
             NULL) "LOCK LEVEL",
    O.OWNER, O.OBJECT_NAME, O.OBJECT_TYPE
FROM V$SESSION S, V$LOCK L, DBA_OBJECTS O
WHERE S.SID = L.SID
AND O.OBJECT_ID = L.ID1
AND S.USERNAME IS NOT NULL

4.2 락이 걸린 세션 간단히 보기[편집]

SELECT A.SID, A.SERIAL#, B.TYPE, C.OBJECT_NAME, A.PROGRAM, A.LOCKWAIT,
       A.LOGON_TIME, A.PROCESS, A.OSUSER, A.TERMINAL
  FROM V$SESSION   A
     , V$LOCK      B
     , DBA_OBJECTS C
 WHERE A.SID = B.SID
   AND B.ID1 = C.OBJECT_ID
   AND B.TYPE = 'TM';

4.3 락이 걸린 세션 상세 보기[편집]

V$SESSION
DBA_OBJECTS
V$LOCK

SELECT A.SID, A.SERIAL#,A.USERNAME,A.PROCESS,B.OBJECT_NAME
     , DECODE(C.LMODE,2,'RS',3,'RX',4,'S',5,'SRX',8,'X','NO') "TABLE LOCK"
     , DECODE (A.COMMAND,2,'INSERT',3,'SELECT',6,'UPDATE',7,'DELETE',12,'DROP TABLE',26,'LOCK TABLE','UNKNOWN') "SQL"
     , DECODE(A.LOCKWAIT, NULL,'NO WAIT','WAIT') "STATUS"
  FROM V$SESSION A
     , DBA_OBJECTS B
     , V$LOCK C
 WHERE A.SID=C.SID 
   AND B.OBJECT_ID=C.ID1
   AND C.TYPE='TM'

락이 걸린 세션을 찾아 내어 세션을 죽이려고 해도 죽지 않는 경우
아래 쿼리문으로 OS단의 PROCESS ID를 찾아내어 OS에서 죽인다
kill -9 프로세스아이디

SELECT SUBSTR(S.USERNAME,1,11) "ORACLE USER", P.PID "PROCESS ID"
     , S.SID "SESSION ID", S.SERIAL#, OSUSER "OS USER"
     , P.SPID "PROC SPID",S.PROCESS "SESS SPID", S.LOCKWAIT "LOCK WAIT"
  FROM V$PROCESS P
     , V$SESSION S
     , V$ACCESS  A
 WHERE A.SID=S.SID 
   AND P.ADDR=S.PADDR 
   AND S.USERNAME != 'SYS'

위 쿼리문의 결과가 있다면 락이 걸린 세션이 있다는것이므로 아래의 쿼리문으로 세션을 죽인다

ALTER SYSTEM KILL SESSION '11,39061'

alter session으로 죽지않는 프로세스 죽이기

1.oracle이 설치된 서버에 텔넷으로 root로 접속한다
2.su -오라클계정
3.sqlplus '/as sysdba''
4.connect system/sys
5.ALTER SYSTEM KILL SESSION '137,1723'

4.4 1일 이상 존재 하는 조회 / kill / AWS_KILL_CMD[편집]

-- DROP VIEW RTIS_DBA.V_DBA_KILL_SESS_OVER_1DAY;

CREATE OR REPLACE FORCE VIEW V_DBA_KILL_SESS_OVER_1DAY
(
    STATUS
  , USERNAME
  , OSUSER
  , MACHINE
  , EXEC_TIME
  , SID
  , SERIAL#
  , CMD_KILL
)
BEQUEATH DEFINER
AS
      SELECT STATUS
           , USERNAME
           , OSUSER
           , MACHINE
           , NVL (A.SQL_EXEC_START, A.PREV_EXEC_START)    EXEC_TIME
           , SID
           , SERIAL#
           ,    'EXEC RDSADMIN.RDSADMIN_UTIL.KILL('
             || TO_CHAR (SID)
             || ','
             || TO_CHAR (SERIAL#)
             || ');'                                      CMD_KILL
        FROM V$SESSION A
       WHERE     STATUS = 'INACTIVE'
             AND TYPE = 'USER'
             AND MACHINE NOT LIKE 'ip%'
             AND (SYSDATE - NVL (A.SQL_EXEC_START, A.PREV_EXEC_START)) * 24 > 1
    ORDER BY EXEC_TIME DESC;

5 프로세스 정보[편집]

5.1 Oracle Process의 정보[편집]

V$SESSION
V$PROCESS
SYS.V_$SESS_IO

SELECT S.STATUS "STATUS"
     , S.SERIAL# "SERIAL#"
     , S.TYPE "TYPE"
     , S.USERNAME "DB USER"
     , S.OSUSER "CLIENT USER"
     , S.SERVER "SERVER"
     , S.MACHINE "MACHINE"
     , S.MODULE "MODULE"
     , S.TERMINAL "TERMINAL"
     , S.PROGRAM "PROGRAM"
     , P.PROGRAM "O.S. PROGRAM"
     , S.LOGON_TIME "CONNECT TIME"
     , LOCKWAIT "LOCK WAIT"
     , SI.PHYSICAL_READS "PHYSICAL READS"
     , SI.BLOCK_GETS "BLOCK GETS"
     , SI.CONSISTENT_GETS "CONSISTENT GETS"
     , SI.BLOCK_CHANGES "BLOCK CHANGES"
     , SI.CONSISTENT_CHANGES "CONSISTENT CHANGES"
     , S.PROCESS "PROCESS"
     , P.SPID
     , P.PID
     , S.SERIAL#
     , SI.SID
     , S.SQL_ADDRESS "ADDRESS"
     , S.SQL_HASH_VALUE "SQL HASH"
     , S.ACTION
  FROM V$SESSION S
     , V$PROCESS P
     , SYS.V_$SESS_IO SI
 WHERE S.PADDR = P.ADDR(+) 
   AND SI.SID(+) = S.SID 
   AND S.USERNAME IS NOT NULL 
   AND NVL(S.OSUSER, 'X') <> 'SYSTEM' 
   AND S.TYPE <> 'BACKGROUND' 
 ORDER BY 3;

5.2 오브젝트에 접속되어 있는 프로그램 조회[편집]

V$SESSION
V$ACCESS

SELECT SUBSTR(B.OBJECT, 1, 15) AS OBJECT, SUBSTR(A.PROGRAM, 1, 15) AS PROGRAM, COUNT(*) AS CNT 
  FROM V$SESSION A
     , V$ACCESS B
 WHERE A.SID = B.SID 
   AND B.OWNER NOT IN ('SYS') 
   AND A.TYPE!= 'BACKGROUND' 
   AND B.OBJECT LIKE UPPER('&OBJECT_NAME') || '%' 
 GROUP BY B.OBJECT, SUBSTR(A.PROGRAM, 1, 15);

6 부하 발생 모니터링[편집]

6.1 cpu를 많이 사용하는 쿼리문과 프로세스아이디,시리얼번호,머신 알아내기[편집]

V$PROCESS
V$SESSION
V$SQLTEXT

SELECT C.SQL_TEXT
     , B.SID
     , B.SERIAL#
     , B.MACHINE
     , B.OSUSER
     , B.LOGON_TIME --이 쿼리를 호출한 시간
  FROM V$PROCESS A
     , V$SESSION B
     , V$SQLTEXT C
 WHERE A.ADDR = B.PADDR
   AND B.SQL_HASH_VALUE = C.HASH_VALUE
--AND A.SPID = '675958'
 ORDER BY C.PIECE

6.2 롤백 세그먼트 경합 조회[편집]

V$ROLLSTAT
V$ROLLNAME

SELECT NAME T0
     , GETS T1
     , WAITS T2
     , TO_CHAR(TRUNC(WAITS / GETS * 100, 2), 099.99) || '%' T3
     , TO_CHAR(ROUND(RSSIZE / 1024)) T4
     , SHRINKS T5
     , EXTENDS T6
  FROM V$ROLLSTAT
     , V$ROLLNAME
 WHERE V$ROLLSTAT.USN = V$ROLLNAME.USN;

6.3 버퍼 캐시 히트율(Buffer Cache Hit Ratio)[편집]

V$SYSSTAT

SELECT ROUND(((1-(SUM(DECODE(name, 'physical reads', value,0))/(SUM(DECODE(name, 'db block gets', value,0))+ (SUM(DECODE(name, 'consistent gets', value, 0))))))*100),2) || '%' "Buffer Cache Hit Ratio"
  FROM V$SYSSTAT;

6.4 라이브러리 캐시 히트율(Library Cache Hit Ratio)[편집]

V$LIBRARYCACHE

SELECT (1-SUM (reloads)/SUM(pins))*100 "Library Cache Hit Ratio"
From V$LIBRARYCACHE;

6.5 데이터 딕셔너리 캐시 히트율(Data Dictionary Cache Hit Ratio)[편집]

V$ROWCACHE

SELECT (1-SUM(getmisses)/SUM(gets))*100 "Data Dictionary Hit Ratio"
  FROM V$ROWCACHE;

7 I/O 부하 찾기[편집]

V$IOFUNCMETRIC

-- iometric values in 11g

select min(begin_time) b_time, min(end_time) e_time, function_name,
       round(sum(small_read_iops+large_read_iops)) read_TOT_iops, round(sum(small_write_iops+large_write_iops)) write_TOT_iops, round(sum(large_read_mbps+small_read_mbps)) read_TOT_mbps, round(sum(large_write_mbps+small_write_mbps)) write_TOT_mbps
       from GV$IOFUNCMETRIC
       group by rollup(function_name)
       having round(sum(small_read_iops+large_read_iops)) + round(sum(large_read_mbps+small_read_mbps)) + round(sum(small_write_iops+large_write_iops)) + round(sum(large_write_mbps+small_write_mbps))  >0
       order by function_name;

-- query to iofuncmetric view in 11g

select inst_id,begin_time,function_name,
       round(small_read_iops) RD_IOPS_sm, round(large_read_iops) RD_IOPS_lg, 
       round(small_read_mbps) RD_MBPS_sm, round(large_read_mbps) RD_MBPS_lg, 
       round(small_write_iops) WT_IOPS_sm, round(large_write_iops) WT_IOPS_lg, 
       round(small_write_mbps) WT_MBPS_sm, round(large_write_mbps) WT_MBPS_lg 
from GV$IOFUNCMETRIC
--where function_name in ('Buffer Cache Reads','LGWR','DBWR','Direct Reads','Direct Writes','RMAN') 
--where round(small_read_iops+large_read_iops+small_write_iops+large_write_iops) >0
order by function_name,inst_id;

select min(begin_time) b_time, min(end_time) e_time, round(sum(small_read_iops+large_read_iops)) read_TOT_iops, round(sum(large_read_mbps+small_read_mbps)) read_TOT_mbps, round(sum(small_write_iops+large_write_iops)) write_TOT_iops, round(sum(large_write_mbps+small_write_mbps)) write_TOT_mbps from GV$IOFUNCMETRIC;

7.1 인스턴스 기동 이후 I/O 관련 대기현상 누적 값[편집]

- v$system_event의 time_waited는 centisecond (1/100 초)

SELECT EVENT ,
       TOTAL_WAITS ,
       TIME_WAITED / 100 AS "TIME_WAITED(SEC)" ,
       AVERAGE_WAIT
FROM V$SYSTEM_EVENT
WHERE EVENT IN ('DB FILE SEQUENTIAL READ' ,
                'LOG FILE SYNC' ,
                'DB FILE SCATTERED READ') 
;

7.2 전체 IO량(100%) 대비 사용자별 IO량[편집]

WITH io AS (
  SELECT /*+ materialize */ decode(statistic_name,'logical reads', 'lio','pio') io_type
       , sum(value) io_value
  FROM v$segment_statistics
 WHERE statistic_name IN ( 'logical reads' ,
                           'physical reads' ,
                           'physical reads direct' )
 GROUP BY decode(statistic_name,'logical reads', 'lio','pio')),
per_io AS
( SELECT OWNER,
         object_name,
         object_type,
         seg_io,
         seg_value,
         round(seg_value / decode(seg_io,'lio',
                                      (SELECT io_value
                                       FROM io
                                       WHERE io_type='lio'),
                                      (SELECT io_value
                                       FROM io
                                       WHERE io_type='pio') )*100,5) per_seg_io
 FROM
     ( SELECT ss.OWNER,
                 ss.object_name,
                 ss.object_type,
                 decode(ss.statistic_name,'logical reads', 'lio','pio') seg_io,
                                                                        sum(ss.value) seg_value
      FROM v$segment_statistics ss
      WHERE ss.statistic_name IN ( 'logical reads' ,
                                   'physical reads' ,
                                   'physical reads direct' )
      GROUP BY ss.OWNER,
                  ss.object_name,
                  ss.object_type,
                  decode(ss.statistic_name,'logical reads', 'lio','pio') ))
SELECT *
FROM
( SELECT OWNER,
         sum(decode(seg_io,'lio',per_seg_io)) AS lio,
         sum(decode(seg_io,'pio',per_seg_io)) AS pio
 FROM per_io
 GROUP BY OWNER
 ORDER BY sum(decode(seg_io,'lio',per_seg_io)) DESC)
WHERE rownum <= 50 - Segment별 전체 IO량(100%) 대비 IO량 (TABLE/INDEX 분리) WITH io AS
( SELECT /*+ materialize */ decode(statistic_name,'logical reads', 'lio','pio') io_type,
                                                                                sum(value) io_value
 FROM v$segment_statistics
 WHERE statistic_name IN ( 'logical reads' ,
                           'physical reads' ,
                           'physical reads direct' )
 GROUP BY decode(statistic_name,'logical reads', 'lio','pio')),
                                                                        per_io AS
( SELECT OWNER,
         object_name,
         object_type,
         seg_io,
         seg_value,
         round(seg_value / decode(seg_io,'lio',
                                      (SELECT io_value
                                       FROM io
                                       WHERE io_type='lio'),
                                      (SELECT io_value
                                       FROM io
                                       WHERE io_type='pio') )*100,5) per_seg_io
 FROM
     ( SELECT ss.OWNER,
                 ss.object_name,
                 ss.object_type,
                 decode(ss.statistic_name,'logical reads', 'lio','pio') seg_io,
                                                                        sum(ss.value) seg_value
      FROM v$segment_statistics ss
      WHERE ss.statistic_name IN ( 'logical reads' ,
                                   'physical reads' ,
                                   'physical reads direct' )
      GROUP BY ss.OWNER,
                  ss.object_name,
                  ss.object_type,
                  decode(ss.statistic_name,'logical reads', 'lio','pio') ))

SELECT *
  FROM per_io
ORDER BY per_seg_io DESC 
;

8 테이블스페이스[편집]

8.1 테이블스페이스 사용량[편집]

dba_data_files
dba_free_space

SELECT A.TABLESPACE_NAME,
             A.TOTAL "TOTAL(MB)",
             A.TOTAL - B.FREE "USED(MB)",
             NVL(B.FREE,0) "FREE(MB)",
             ROUND((A.TOTAL - NVL(B.FREE,0))*100/TOTAL,0)  "USED(%)"
  FROM ( SELECT TABLESPACE_NAME
                , ROUND((SUM(BYTES)/1024/1024),0) AS TOTAL
             FROM DBA_DATA_FILES
            GROUP BY TABLESPACE_NAME
       ) A
     , ( SELECT TABLESPACE_NAME
              , ROUND((SUM(BYTES)/1024/1024),0) AS FREE
           FROM DBA_FREE_SPACE
          GROUP BY  TABLESPACE_NAME
       ) B
 WHERE A.TABLESPACE_NAME = B.TABLESPACE_NAME(+)
 ORDER BY A.TABLESPACE_NAME;

8.2 System 테이블스페이스에 비시스템 세그먼트 조회[편집]

DBA_SEGMENTS

SELECT OWNER
     , SEGMENT_NAME
     , SEGMENT_TYPE
     , TABLESPACE_NAME
  FROM DBA_SEGMENTS 
 WHERE OWNER NOT IN ('SYS', 'SYSTEM') 
   AND TABLESPACE_NAME = 'SYSTEM';

9 CPU/메모리/사양[편집]

9.1 오라클서버의 메모리[편집]

v$sgastat

SELECT * FROM V$SGASTAT;

SELECT POOL, SUM(BYTES) "SIZE"
  FROM V$SGASTAT
 WHERE POOL = 'SHARED POOL'
 GROUP BY POOL

9.2 총 CPU Time 대비 SQL Parsing Time[편집]

V$SYSSTAT

SELECT MAX(DECODE(Y.RNO,1,DECODE(X.NAME,'CPU USED BY THIS SESSION',X.VALUE))) AS "CPU USED BY THIS SESSION" ,
       MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'PARSE TIME CPU',X.VALUE))) AS "PARSE TIME CPU" ,
       ROUND((MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'PARSE TIME CPU',X.VALUE)))/MAX(DECODE(Y.RNO,1,DECODE(X.NAME,'CPU USED BY THIS SESSION',X.VALUE))) * 100),1)||'%' AS "PARSETIME/CPU(%)"
FROM
    ( SELECT NAME,
             VALUE
     FROM V$SYSSTAT
     WHERE NAME IN ('CPU USED BY THIS SESSION',
                    'PARSE TIME CPU') ) X,
    (SELECT LEVEL AS RNO
     FROM DUAL CONNECT BY LEVEL<=2) Y ;

9.3 총 Parsing Time 대비 Hard Parsing Time[편집]

V$SYS_TIME_MODEL

SELECT MAX(DECODE(Y.RNO,1,DECODE(X.NAME,'PARSE TIME ELAPSED',X.VALUE))) AS "PARSE TIME ELAPSED" ,
       MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'HARD PARSE ELAPSED TIME',X.VALUE))) AS "HARD PARSE ELAPSED TIME" ,
       ROUND((MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'HARD PARSE ELAPSED TIME',X.VALUE)))/MAX(DECODE(Y.RNO,1,DECODE(X.NAME,'PARSE TIME ELAPSED',X.VALUE))) * 100),1)||'%' AS "HARD/PARSETORAL(%)"
FROM
    ( SELECT STAT_NAME AS NAME,
             VALUE
     FROM V$SYS_TIME_MODEL
     WHERE STAT_NAME IN ('PARSE TIME ELAPSED',
                         'HARD PARSE ELAPSED TIME') ) X,
    (SELECT LEVEL AS RNO
     FROM DUAL CONNECT BY LEVEL<=2) Y ;

9.4 Log file sync 관련 대기현상 Sync Writes Time[편집]

redo synch time의 단위는 centisecond (1/100 초)
v$sysstat의 stat 중 time 정보를 담고 있는 stat의 value 단위는 centisecond이다.

SELECT MAX(DECODE(Y.RNO,1,DECODE(X.NAME,'USER COMMITS',X.VALUE))) AS "USER COMMITS" ,
       MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'REDO SYNCH WRITES',X.VALUE))) AS "REDO SYNCH WRITES" ,
       MAX(DECODE(Y.RNO,3,DECODE(X.NAME,
        'REDO SYNCH TIME',X.VALUE/100))) AS "REDO SYNCH TIME" ,
       ROUND((MAX(DECODE(Y.RNO,3,DECODE(X.NAME,'REDO SYNCH TIME',X.VALUE/100)))/MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'REDO SYNCH WRITES',X.VALUE)))),3)||'초' AS "SYNCTIME/SYNCWRITES(초)"
FROM
    ( SELECT NAME,
             VALUE
     FROM V$SYSSTAT
     WHERE NAME IN ('USER COMMITS' ,
                    'REDO SYNCH TIME' ,
                    'REDO SYNCH WRITES') ) X,
    (SELECT LEVEL AS RNO
     FROM DUAL CONNECT BY LEVEL<=3) Y 
;

10 INDEX 정보 추출하기[편집]

10.1 Function-based Index 컬럼 찾기[편집]

SELECT table_owner,
       TABLE_NAME,
       index_name,
       column_expression
FROM dba_ind_expressions
WHERE table_owner LIKE upper(ltrim(rtrim(:owner)))
    AND TABLE_NAME LIKE upper(ltrim(rtrim(:table))) ;
;

10.2 중복 인덱스 찾기(불필요 인덱스)[편집]

SELECT di.table_owner "OWNER" ,
       di.TABLE_NAME ,
          dic1.index_name || chr(10) || ' (' || replace(dic1.index_cols,' ',',') || decode(sign(dic1.cnt-6),1,'...') || ')' "삭제대상 INDEX" , dic2.index_name || chr(10) || ' (' || replace(dic2.index_cols,' ',',') || decode(sign(dic1.cnt-7),1,'...') || ')' "삭제원인 INDEX"
FROM dba_indexes di ,
    ( SELECT table_owner,
             TABLE_NAME,
             index_owner,
             index_name,
             TRIM(MAX(DECODE(column_position, 1, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 2, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 3, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 4, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 5, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 6, COLUMN_NAME))) INDEX_COLS,
             count(*) cnt
     FROM dba_ind_columns dic
     WHERE table_owner NOT IN ('SYS',
                               'SYSTEM')
     GROUP BY table_owner,
              TABLE_NAME,
              index_owner,
              index_name ) dic1 ,
    ( SELECT table_owner,
             TABLE_NAME,
             index_owner,
             index_name,
             TRIM(MAX(DECODE(column_position, 1, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 2, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 3, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 4, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 5, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 6, COLUMN_NAME)) || ' ' || MAX(DECODE(column_position, 7, COLUMN_NAME))) INDEX_COLS,
             count(*) cnt
     FROM dba_ind_columns dic
     WHERE table_owner NOT IN ('SYS',
                               'SYSTEM')
     GROUP BY table_owner,
              TABLE_NAME,
              index_owner,
              index_name ) dic2
WHERE di.table_owner NOT IN ('SYS','SYSTEM')
    AND di.owner=dic1.index_owner
    AND di.index_name=dic1.index_name
    AND di.table_owner=dic1.table_owner
    AND di.TABLE_NAME=dic1.TABLE_NAME
    AND dic1.table_owner=dic2.table_owner
    AND dic1.TABLE_NAME=dic2.TABLE_NAME
    AND dic1.index_name<>dic2.index_name
    AND dic2.index_cols LIKE dic1.index_cols || '%'
    AND dic2.cnt>dic1.cnt 
;

10.3 CONSTRAINTS 확인하기[편집]

-- 해당 테이블에 걸린 Constraints 확인하기

SELECT t.owner ,
           t.constraint_name ,
           t.constraint_type ,
           t.TABLE_NAME ,
             c.COLUMN_NAME ,
               c.position ,
               t.r_owner ,
               t.r_constraint_name ,
               t.status ,
               t.last_change ,
               t.index_owner ,
               t.index_name
    FROM dba_constraints t ,
         dba_cons_columns c WHERE t.TABLE_NAME = :table_name
    AND t.TABLE_NAME = c.TABLE_NAME
    AND t.constraint_name = c.constraint_name
    AND t.constraint_type <> 'C'
ORDER BY TABLE_NAME ASC, constraint_type DESC,
                         CONSTRAINT_NAME ASC, POSITION ASC 
;

10.4 해당 테이블을 참조하는 테이블(자식) Constraints 확인하기[편집]

SELECT t.owner ,
       t.CONSTRAINT_NAME ,
         t.constraint_type ,
         t.TABLE_NAME ,
           c.COLUMN_NAME ,
             c.POSITION ,
               t.r_owner ,
               t.r_constraint_name ,
               t.status ,
               t.last_change ,
               t.index_owner ,
               t.index_name
FROM dba_constraints t ,
     dba_cons_columns c
WHERE t.R_CONSTRAINT_NAME = '%'|| :table_name ||'%'
    AND t.TABLE_NAME = c.TABLE_NAME
    AND t.CONSTRAINT_NAME = c.CONSTRAINT_NAME
    AND t.constraint_type <> 'C'
ORDER BY TABLE_NAME ASC, constraint_type DESC,
                         CONSTRAINT_NAME ASC, POSITION ASC 
;

10.5 v$segment_statistics 활용하기[편집]

10.6 Parameter 확인하기[편집]

SET linesize 200
SET pagesize 100 col name
FOR a30 col value
FOR a10 col display_value
FOR a10 col isdefault
FOR a10 col ismodified
FOR a10

SELECT name,
       value,
       isdefault,
       ismodified
FROM v$parameter
WHERE name IN ( 'lock_sga' ,
                'db_cache_advice' ,
                'optimizer_dynamic_sampling' ,
                'session_cached_cursors' ,
                'sga_max_size' ,
                'sga_target' ,
                'db_cache_size' ,
                'shared_pool_size' ,
                'shared_pool_reserved_size' ,
                'log_buffer' ,
                'skip_unusable_indexes' ,
                'pga_aggregate_target' ,
                'workarea_size_policy' ,
                'cpu_count' ,
                'statistics_level') ;

SELECT ksppinm AS name,
       ksppstvl AS value
FROM sys.x$ksppi x
   , sys.x$ksppcv y
WHERE (x.indx = y.indx)
AND (translate(ksppinm,'_','#') LIKE '%_optim_peek_user_binds%'
     OR translate(ksppinm,'_','#') LIKE '%_kks_use_mutex_pin%'
     OR translate(ksppinm,'_','#') LIKE '%_gby_hash_aggregation_enabled%'
     OR translate(ksppinm,'_','#') LIKE '%_gc_affinity_time%'
     OR translate(ksppinm,'_','#') LIKE '%_optimizer_skip_scan_enabled%'
     OR translate(ksppinm,'_','#') LIKE '%_pga_max_size%'
     OR translate(ksppinm,'_','#') LIKE '%_smm_max_size%'
     OR translate(ksppinm,'_','#') LIKE '%_b_tree_bitmap_plans%'
     OR translate(ksppinm,'_','#') LIKE '%_undo_autotune%' ) ;

10.7 Bind Peeked 확인하기[편집]

SELECT p.plan_table_output
  FROM
( SELECT sql_id ,
         child_number
 FROM v$sql
 WHERE hash_value=:hash_value ) s 
     , TABLE(dbms_xplan.display_cursor(s.sql_id , s.child_number , 'typical +peeked_binds')) p 
;

-- Group By 처리하는 방식 변경

파라미터 : _gby_hash_aggregation_enabled VERSION : 10g에서 신규 추가된 파라미터 해석방법 : FALSE 
-
GROUP BY +
ORDER BY TRUE -
GROUP BY 주의사항 : 9i에서 10g로 업그레이드 수행 시 해당 파라미터가 True이면 기존
GROUP BY 절이 있는 SQL문의 정렬이 되지 않기 때문에 도출되는 결과가 바뀔 수 있다. 이럴 경우에는 해당 프로그램들에
ORDER by를 추가하여야 한다. _b_tree_bitmap_plans=TRUE

10.8 JOB 확인하기[편집]

SELECT *
  FROM dba_jobs
 WHERE last_date >= to_date('20110906 14:00:00' , 'yyyymmdd hh24:mi:ss')
ORDER BY last_date

SELECT *
  FROM dba_objects
 WHERE object_name = upper('proc_olap_summ_measure_run_02')

SELECT *
FROM v$sqlarea WHERE program_id = 100286 
;

SELECT *
FROM TABLE(DBMS_XPLAN.DISPLAY_CURSOR(NULL,NULL,’ALLSTATS COST LAST’));

10.9 DB Link SQL CURSOR 확인하기[편집]

SELECT inst_id ,
       kglnaown ,
       kglnaobj ,
       kglnadlk , -- DB Link
       decode(kglhdnsp , 0 , 'CURSOR' , 1 , 'TABLE/PROCEDURE' , 2 , 'BODY' , 3 , 'TRIGGER' , 4 , 'INDEX' , 5 , 'CLUSTER' , 6 , 'OBJECT' , 13 , 'JAVA SOURCE' , 14 , 'JAVA RESOURCE' , 15 , 'REPLICATED TABLE OBJECT' , 16 , 'REPLICATION INTERNAL PACKAGE' , 17 , 'CONTEXT POLICY' , 18 , 'PUB_SUB' , 19 , 'SUMMARY' , 20 , 'DIMENSION' , 21 , 'APP CONTEXT' , 22 , 'STORED OUTLINE' , 23 , 'RULESET' , 24 , 'RSRC PLAN' , 25 , 'RSRC CONSUMER GROUP' , 26 , 'PENDING RSRC PLAN' , 27 , 'PENDING RSRC CONSUMER GROUP' , 28 , 'SUBSCRIPTION' , 29 , 'LOCATION' , 30 , 'REMOTE OBJECT' , 31 , 'SNAPSHOT METADATA' , 32 , 'JAVA SHARED DATA' , 33 , 'SECURITY PROFILE' , 'INVALID NAMESPACE') ,
       decode(bitand(kglobflg , 3) , 0 , 'NOT LOADED' , 2 , 'NON-EXISTENT' , 3 , 'INVALID STATUS' , decode(kglobtyp , 0 , 'CURSOR' , 1 , 'INDEX' , 2 , 'TABLE' , 3 , 'CLUSTER' , 4 , 'VIEW' , 5 , 'SYNONYM' , 6 , 'SEQUENCE' , 7 , 'PROCEDURE' , 8 , 'FUNCTION' , 9 , 'PACKAGE' , 10 , 'NON-EXISTENT' , 11 , 'PACKAGE BODY' , 12 , 'TRIGGER' , 13 , 'TYPE' , 14 , 'TYPE BODY' , 15 , 'OBJECT' , 16 , 'USER' , 17 , 'DBLINK' , 18 , 'PIPE' , 19 , 'TABLE PARTITION' , 20 , 'INDEX PARTITION' , 21 , 'LOB' , 22 , 'LIBRARY' , 23 , 'DIRECTORY' , 24 , 'QUEUE' , 25 , 'INDEX-ORGANIZED TABLE' , 26 , 'REPLICATION OBJECT GROUP' , 27 , 'REPLICATION PROPAGATOR' , 28 , 'JAVA SOURCE' , 29 , 'JAVA CLASS' , 30 , 'JAVA RESOURCE' , 31 , 'JAVA JAR' , 32 , 'INDEX TYPE' , 33 , 'OPERATOR' , 34 , 'TABLE SUBPARTITION' , 35 , 'INDEX SUBPARTITION' , 36 , 'REPLICATED TABLE OBJECT' , 37 , 'REPLICATION INTERNAL PACKAGE' , 38 , 'CONTEXT POLICY' , 39 , 'PUB_SUB' , 40 , 'LOB PARTITION' , 41 , 'LOB SUBPARTITION' , 42 , 'SUMMARY' , 43 , 'DIMENSION' , 44 , 'APP CONTEXT' , 45 , 'STORED OUTLINE' , 46 , 'RULESET' , 47 , 'RSRC PLAN' , 48 , 'RSRC CONSUMER GROUP' , 49 , 'PENDING RSRC PLAN' , 50 , 'PENDING RSRC CONSUMER GROUP' , 51 , 'SUBSCRIPTION' , 52 , 'LOCATION' , 53 , 'REMOTE OBJECT' , 54 , 'SNAPSHOT METADATA' , 55 , 'IFS' , 56 , 'JAVA SHARED DATA' , 57 , 'SECURITY PROFILE' , 'INVALID TYPE')) ,
       kglobhs0+kglobhs1+kglobhs2+kglobhs3+kglobhs4+kglobhs5+kglobhs6 ,
       kglhdldc ,
       kglhdexc ,
       kglhdlkc ,
       kglobpc0 ,
       decode(kglhdkmk , 0 , 'NO' , 'YES') ,
       kglhdclt ,
       kglhdivc ,
       kglhdpar AS paddress,
       ------->>> 추가한 부분

       kglhdadr AS address,
       ------->>> 추가한 부분

       kglnahsh AS hash_value,
       ------->>> 추가한 부분

       kglobt03 AS sql_id ------->>> 추가한 부분
FROM x$kglob
WHERE kglhdnsp = 0 -- NAMESPACE='CURSOR'
AND kglnadlk IS NOT NULL ;

10.10 INDEX Deleted ROWS CHECK [1][편집]

- Validate + index_stats

SET serveroutput ON
SET pagesize 0
SET linesize 200
SET feedback OFF
SET timing OFF 

--------------------------------------------------------------------------------
-- Work Time
--------------------------------------------------------------------------------

COLUMN the_date new_value run_date
SELECT to_char(sysdate,'yyyymmddhh24mi') the_date
FROM dual;

--------------------------------------------------------------------------------
-- File Creation - Index Size Check Script
--------------------------------------------------------------------------------
spool index_validate_check.SQL DECLARE
CURSOR cur_index_name IS
SELECT OWNER,
       index_name
FROM dba_indexes
WHERE OWNER = 'SYSTEM' ;

BEGIN dbms_output.enable(buffer_size=>2000000);
FOR index_rec IN cur_index_name LOOP dbms_output.put_line('Validate index '||index_rec.OWNER||'.'||index_rec.index_name||';');
dbms_output.put_line(' ');
dbms_output.put_line('select  '''||index_rec.OWNER||''',');
dbms_output.put_line('        name index_name,');
dbms_output.put_line('        lf_rows Tot_rows,');
dbms_output.put_line('        del_lf_rows Deleted_Rows,');
dbms_output.put_line('        (del_lf_rows * 100)/decode(lf_rows,0,0.1,lf_rows) "deleted(%)",');
dbms_output.put_line('        round(btree_space/1024/1024/1024,3) "tot_space(GB)",');
dbms_output.put_line('        round(used_space/1024/1024/1024,3) "used_space(GB)",');
dbms_output.put_line('        pct_used "used_pct(%)",');
dbms_output.put_line('        100 - pct_used "empty_pct(%)",');
dbms_output.put_line('        decode( sign((del_lf_rows * 100)/decode(lf_rows,0,0.1,lf_rows) - 20),1,');
dbms_output.put_line('                ''--->> Rebuild required'',');
dbms_output.put_line('                ''--->> Good'') Check_Result');
dbms_output.put_line(' from index_stats ;');
END LOOP;
END;
/

--------------------------------------------------------------------------------
-- Run - Index Size Check Script
--------------------------------------------------------------------------------
col OWNER format a10;
col index_name format a33;
col tot_rows format 99999999 col deleted_rows format 99999999 spool c:\total_index_size_check_&run_date..txt prompt OWNER INDEX name total ROWS deleted ROWS deleted(%) tot_space used_space used_pct(%) empty_pct(%) prompt 
---------------------------------------------------------------------------------------------------------------------------------
@index_validate_check spool OFF
SET feedback ON
SET serveroutput OFF 
################################################################################ 
## INDEX Deleted ROWS CHECK [2] - INDEX STATISTICS 
################################################################################ 
col blevel heading "INDEX|BLEVEL" 
col leaf_blocks heading "LEAF|BLOCKS" 
col num_rows heading "NUM|ROWS" 
col usable_key_count_per_block heading "USABLE_KEY|COUNT|PER_BLOCK" 
col current_key_count_per_block heading "CURRENT_KEY|COUNT|PER_BLOCK" 
col used_percent heading "USED|PERCENT|(%)" 
col empty_percent heading "EMPTY|PERCENT|(%)"
SELECT *
FROM
( SELECT /*+ leading(i) use_hash(i s) */ i.index_name,
                                         max(i.blevel) AS blevel,
                                         max(i.leaf_blocks) AS leaf_blocks,
                                         max(i.num_rows) AS num_rows,
                                         sum(s.blocks) AS "BLOCKS",
                                         sum(s.bytes)/1024/1024 AS "SIZE(MB)",
                                         max(i.usable_key_count_per_block) AS usable_key_count_per_block,
                                         max(i.current_key_count_per_block) AS current_key_count_per_block,
                                         max(i.used_percent) AS used_percent,
                                         max(i.empty_percent) AS empty_percent
 FROM
     ( SELECT i.index_name,
              i.blevel,
              i.leaf_blocks,
              i.num_rows,
              round(i.usable_block_size / (i.avg_key_length),0) AS usable_key_count_per_block,
              round(i.num_rows / i.leaf_blocks,0) AS current_key_count_per_block,
              round((i.num_rows / i.leaf_blocks) / (i.usable_block_size / i.avg_key_length) * 100,0) AS used_percent,
              round(100 -((i.num_rows / i.leaf_blocks) / (i.usable_block_size / i.avg_key_length) * 100)) AS empty_percent
      FROM
          ( SELECT i.index_name,
                   sum(c1.avg_col_len)+9 AS avg_key_length,
                   max(t.block_size*0.98) usable_block_size,
                   max(i.blevel) blevel,
                   max(i.leaf_blocks) leaf_blocks,
                   max(i.num_rows) num_rows
           FROM dba_indexes i,
                dba_tab_columns c1,
                dba_ind_columns c2,
                dba_tablespaces t
           WHERE i.OWNER=upper('GCKDPROD') --and    i.table_name = upper('TSD_PTSEQ_I')
 --and    i.index_name LIKE '%'

               AND i.OWNER=c1.OWNER
               AND i.TABLE_NAME=c1.TABLE_NAME
               AND c1.OWNER=c2.table_owner
               AND c1.TABLE_NAME=c2.TABLE_NAME
               AND c1.COLUMN_NAME=c2.COLUMN_NAME
               AND i.OWNER=c2.index_owner
               AND i.index_name=c2.index_name
               AND i.tablespace_name=t.tablespace_name
           GROUP BY i.index_name HAVING max(i.leaf_blocks) >= 10000 ) i ) i,
                                                                          dba_segments s
 WHERE s.OWNER = 'GCKDPROD'
     AND i.index_name = s.segment_name
 GROUP BY i.index_name
 ORDER BY max(i.empty_percent) DESC)
WHERE rownum <= 50 ;

################################################################################ 
## Locking Contention Query 
################################################################################
SELECT OS_USER_NAME AS OSUSER,
       s.serial# AS "SERIAL NO.",
       PROCESS AS "PID",
       ORACLE_USERNAME AS "USERNAME",
       l.SID AS "SID",
       DECODE(l.TYPE -- Long locks, 'TM', 'dml/data enq (TM)', 'TX', 'transac enq (TX) ', 'UL', 'pls usr lock (UL)'
-- Short locks, 'BL', 'buf hash tbl (BL)', 'CF', 'control file (CF)', 'CI', 'cross inst f (CI)', 'CU', 'cursor bind (CU) ', 'DF', 'data file (CF) ', 'DL', 'direct load (DL) ', 'DM', 'mount/strtup (DM)', 'DR', 'reco lock (DR) ', 'DX', 'distrib tran (DX)', 'FI', 'sga opn file (FI)', 'FS', 'file set (FS) ', 'IN', 'instance num (IN)', 'IR', 'instce recvr (IR)', 'IS', 'get state (IS) ', 'IV', 'libcache inv (IV)', 'JQ', 'job queue (JQ) ', 'KK', 'log sw kick (KK) ', 'LS', 'log switch (LS) ', 'MM', 'mount def (MM) ', 'MR', 'media recvry (MR)', 'PF', 'pwfile enq (PF) ', 'PR', 'process strt (PR)', 'RW', 'row wait (RW) ', 'RT', 'redo thread (RT) ', 'SC', 'scn enq (SC) ', 'SM', 'smon lock (SM) ', 'SN', 'seqno instce (SN)', 'SQ', 'seqno enq (SQ) ', 'ST', 'space transc (ST)', 'SV', 'seqno value (SV) ', 'TA', 'generic enq (TA) ', 'TD', 'dll enq (TD) ', 'TE', 'extend seg (TE) ', 'TS', 'temp segment (TS)', 'TT', 'temp table (TT) ', 'UN', 'user name (UN) ', 'WL', 'write redo (WL) ', 'TYPE = ' || l.TYPE) AS type,
 decode(l.lmode, 0, 'none', 1, 'null', 2, 'RS', 3, 'RX', 4, 'S', 5, 'SRX', 6, 'Exclusive', TO_CHAR(l.lmode)) AS lmode,
 decode(l.request, 0, 'none', 1, 'null', 2, 'RS', 3, 'RX', 4, 'S', 5, 'SRX', 6, 'X', TO_CHAR(l.request)) AS lrequest,
 decode(BLOCK, 0, 'Not Blocking', 1, 'Blocking', 2, 'Global', block) AS "DETAIL",
 OWNER,
 object_name
FROM sys.v_$locked_object lo,
                          dba_objects DO,
                                      sys.v_$lock l,
(SELECT a.sid,
        a.serial#
 FROM v$session a,
                v$bgprocess b
 WHERE a.paddr = b.paddr(+)) s
WHERE lo.object_id = DO.object_id
AND l.sid = lo.session_id
AND s.sid = l.sid 
;

11 ASH를 이용한 모니터링[편집]

최근 가장 많이 수행 된 SQL과 수행 점유율(수행 횟수)
v$active_session_history

SELECT sql_id ,
       COUNT(*) ,
       COUNT(*) *100/sum(COUNT(*)) over() pctload
  FROM v$active_session_history 
 WHERE sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
AND sample_time < to_date(:to_time,'yyyymmdd hh24miss')
GROUP BY sql_id
ORDER BY COUNT(*) DESC ;

11.1 특정 Session이 가장 많이 수행 된 SQL과 수행 점유율(수행 횟수)[편집]

v$active_session_history

SELECT sql_id ,
       COUNT(*) ,
       COUNT(*) *100/sum(COUNT(*)) over() pctload
FROM v$active_session_history
WHERE sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
AND sample_time < to_date(:to_time,'yyyymmdd hh24miss')
AND session_id = :b1
GROUP BY sql_id
ORDER BY COUNT(*) DESC ;

11.2 주요한 이벤트 메트릭 표시 gv$eventmetric v$event_name[편집]

sqlplus 스크립트

/* 
   eventmetric.sql - sqlplus script - displays significant event metrics
*/

col "Time /Delta" for a14
col name for a40
col INST_ID for 999
set linesize 140
set pagesize 1000

set wrap off 
select "Time /Delta",inst_id,name, 
        T_per_wait_fg*10 "Avg_FG_wait_ms", round(T_waited_fg/100,1) "Waited_FG_sec", W_count_fg "W_count_FG",
        round(T_waited/100,1) "Waited_tot_sec", W_count "W_count_tot"       
from (
  select to_char(min(begin_time),'hh24:mi:ss')||' /'||round(avg(intsize_csec/100),0)||'s' "Time /Delta",
       em.inst_id,en.name,
       sum(em.time_waited_fg) T_waited_fg, sum(em.time_waited) T_waited,sum(wait_count) W_count, sum(wait_count_fg) W_count_fg,
       sum(decode(em.wait_count, 0,0,round(em.time_waited/em.wait_count,2))) T_per_wait,
       sum(decode(em.wait_count_fg, 0,0,round(em.time_waited_fg/em.wait_count_fg,2))) T_per_wait_fg
  from gv$eventmetric em
     , v$event_name en
  where em.event#=en.event#
      and en.wait_class <>'Idle'
  group by em.inst_id,en.name,em.event_id
  order by T_waited_fg desc
  ) 
where rownum<=20;
set wrap on

11.3 중요한 시스템 메트릭 표시 gv$sysmetric[편집]

select "Time+Delta", "Metric", 
       case when "Total" >10000000 then '* '||round("Total"/1024/1024,0)||' M' 
            when "Total" between 10000 and 10000000 then '+ '||round("Total"/1024,0)||' K'
            when "Total" between 10 and 1024 then '  '||to_char(round("Total",0))
            else '  '||to_char("Total") 
       end "Total"
from (
 select to_char(min(begin_time),'hh24:mi:ss')||' /'||round(avg(intsize_csec/100),0)||'s' "Time+Delta",
       metric_name||' - '||metric_unit "Metric", 
       nvl(sum(value_inst1),0)+nvl(sum(value_inst2),0)+nvl(sum(value_inst3),0)+nvl(sum(value_inst4),0)+
       nvl(sum(value_inst5),0)+nvl(sum(value_inst6),0)+nvl(sum(value_inst7),0)+nvl(sum(value_inst8),0) "Total",
       sum(value_inst1) inst1, sum(value_inst2) inst2, sum(value_inst3) inst3, sum(value_inst4) inst4,
       sum(value_inst5) inst5, sum(value_inst6) inst6, sum(value_inst7) inst7, sum(value_inst8) inst8
 from
  ( select begin_time,intsize_csec,metric_name,metric_unit,metric_id,group_id,
       case inst_id when 1 then round(value,1) end value_inst1,
       case inst_id when 2 then round(value,1) end value_inst2,
       case inst_id when 3 then round(value,1) end value_inst3,
       case inst_id when 4 then round(value,1) end value_inst4,
       case inst_id when 5 then round(value,1) end value_inst5,
       case inst_id when 6 then round(value,1) end value_inst6,
       case inst_id when 7 then round(value,1) end value_inst7,
       case inst_id when 8 then round(value,1) end value_inst8
  from gv$sysmetric
  where metric_name in ('Host CPU Utilization (%)','Current OS Load', 'Physical Write Total IO Requests Per Sec', 
        'Physical Write Total Bytes Per Sec', 'Physical Write IO Requests Per Sec', 'Physical Write Bytes Per Sec',
         'I/O Requests per Second', 'I/O Megabytes per Second',
        'Physical Read Total Bytes Per Sec', 'Physical Read Total IO Requests Per Sec', 'Physical Read IO Requests Per Sec',
        'CPU Usage Per Sec','Network Traffic Volume Per Sec','Logons Per Sec','Redo Generated Per Sec','Redo Writes Per Sec',
        'User Transaction Per Sec','Average Active Sessions','Average Synchronous Single-Block Read Latency',
        'Logical Reads Per Sec','DB Block Changes Per Sec')
  )
 group by metric_id,group_id,metric_name,metric_unit
 order by metric_name
);

11.4 RAC 인스턴스별로 상세한 중요한 시스템 메트릭 gv$sysmetric[편집]

select to_char(min(begin_time),'hh24:mi:ss')||' /'||round(avg(intsize_csec/100),0)||'s' "Time+Delta",
       metric_name||' - '||metric_unit "Metric", 
       sum(value_inst1) inst1, sum(value_inst2) inst2, sum(value_inst3) inst3, sum(value_inst4) inst4,
       sum(value_inst5) inst5, sum(value_inst6) inst6
 from
  ( select begin_time,intsize_csec,metric_name,metric_unit,metric_id,group_id,
       case inst_id when 1 then round(value,1) end value_inst1,
       case inst_id when 2 then round(value,1) end value_inst2,
       case inst_id when 3 then round(value,1) end value_inst3,
       case inst_id when 4 then round(value,1) end value_inst4,
       case inst_id when 5 then round(value,1) end value_inst5,
       case inst_id when 6 then round(value,1) end value_inst6
  from gv$sysmetric
  where metric_name in ('Host CPU Utilization (%)','Current OS Load', 'Physical Write Total IO Requests Per Sec',
        'Physical Write Total Bytes Per Sec', 'Physical Write IO Requests Per Sec', 'Physical Write Bytes Per Sec',
         'I/O Requests per Second', 'I/O Megabytes per Second',
        'Physical Read Total Bytes Per Sec', 'Physical Read Total IO Requests Per Sec', 'Physical Read IO Requests Per Sec',
        'CPU Usage Per Sec','Network Traffic Volume Per Sec','Logons Per Sec','Redo Generated Per Sec',
        'User Transaction Per Sec','Average Active Sessions','Average Synchronous Single-Block Read Latency',
        'Logical Reads Per Sec','DB Block Changes Per Sec')
  )
 group by metric_id,group_id,metric_name,metric_unit
 order by metric_name;

11.5 특정 구간 이벤트 별 대기 시간 v$active_session_history[편집]

SELECT NVL(a.event, 'ON CPU') AS event,
       COUNT(*) AS total_wait_time
FROM v$active_session_history a
WHERE sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
AND sample_time < to_date(:to_time,'yyyymmdd hh24miss')
GROUP BY a.event
ORDER BY total_wait_time DESC;

11.6 특정 구간 CPU 점유율 순 - TOP SQL[편집]

SELECT ash.session_id ,
       ash.session_serial# ,
       ash.user_id ,
       ash.program ,
       MAX(en.name) event_name ,
       SUM(decode(ash.session_state , 'ON CPU' , 1 , 0)) "CPU" ,
       SUM(decode(ash.session_state , 'WAITING' , 1 , 0)) - SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "WAITING" ,
       SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "IO" ,
       SUM(decode(session_state , 'ON CPU' , 1 , 1)) "TOTAL"
FROM v$active_session_history ash 
   , v$event_name en
WHERE en.event# = ash.event#
AND ash.sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
AND ash.sample_time < to_date(:to_time,'yyyymmdd hh24miss')
GROUP BY session_id ,
         user_id ,
         session_serial# ,
         program
ORDER BY SUM(decode(session_state , 'ON CPU' , 1 , 1)) 
;

11.7 특정 구간 CPU 점유율 순 - TOP SESSION v$active_session_history[편집]

SELECT ash.session_id ,
       ash.session_serial# ,
       ash.user_id ,
       ash.program ,
       SUM(decode(ash.session_state , 'ON CPU' , 1 , 0)) "CPU" ,
       SUM(decode(ash.session_state , 'WAITING' , 1 , 0)) - SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "WAITING" ,
       SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "IO" ,
       SUM(decode(session_state , 'ON CPU' , 1 , 1)) "TOTAL"
FROM v$active_session_history ash 
   , v$event_name en
WHERE en.event# = ash.event#
AND ash.sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
AND ash.sample_time < to_date(:to_time,'yyyymmdd hh24miss')
GROUP BY session_id ,
         user_id ,
         session_serial# ,
         program
ORDER BY SUM(decode(session_state , 'ON CPU' , 1 , 1)) 
;

11.8 특정 구간 수행 이력 v$active_session_history[편집]

SELECT ash.sample_time TIME ,
       ash.session_id ,
       ash.session_serial# ,
       ash.user_id ,
       ash.program ,
       ash.module ,
       ash.client_id ,
       ash.machine ,
       ash.session_state ,
       ash.event ,
       ash.sql_id ,
       ash.blocking_session ,
       ash.current_obj# ,
       ash.current_file# ,
       ash.pga_allocated ,
       ash.temp_space_allocated
FROM v$active_session_history ash
WHERE ash.sample_time >= TO_DATE(:from_time , 'yyyymmdd hh24miss')
AND ash.sample_time < TO_DATE(:to_time , 'yyyymmdd hh24miss')
ORDER BY ash.sample_time DESC 
;

12 AWR을 이용한 모니터링[편집]

12.1 AWR을 이용한 literal SQL 추출 방법[편집]

get_literal_sql.SQL

SET linesize 120 col SCHEMA_NAME
FOR a8 col module_name
FOR a15 col sql_text
FOR a35 col SCHEMA_NAME heading "SCHEMA|NAME" col module_name heading "MODULE|NAME" col literal_sql_cnt heading "LITERAL|CNT"
FOR 99999 col execution_cnt heading "EXECUTION|CNT"
FOR 99999 col plan_cnt heading "PLAN|CNT"
FOR 999 col literal_sql_ratio heading "LITERAL|RATIO(%)"
FOR 999 accept i_begin_time prompt 'Enter begin time[YYYYMMDDHH24]: ' accept i_end_time prompt 'Enter end   time[YYYYMMDDHH24]: ' VARIABLE v_begin_time char(10) VARIABLE v_end_time char(10) EXEC :v_begin_time:=&i_begin_time EXEC :v_end_time :=&i_end_time

SELECT a.SCHEMA_NAME , a.module_name , a.literal_sql_cnt , a.execution_cnt , round(a.literal_sql_cnt/a.execution_cnt*100,1) literal_sql_ratio, a.plan_cnt,
    ( SELECT sql_text
     FROM dba_hist_sqltext
     WHERE dbid = a.dbid
         AND sql_id = a.max_sql_id ) sql_text
FROM
    ( SELECT s.force_matching_signature,
             s.dbid ,
             s.parsing_schema_name SCHEMA_NAME ,
                                   s.module module_name ,
                                   count(*) literal_sql_cnt ,
                                   sum(s.executions_total) execution_cnt ,
                                   max(s.sql_id) max_sql_id,
                                   count(DISTINCT s.plan_hash_value) plan_cnt
     FROM sys.wrm$_snapshot w ,
                            dba_hist_sqlstat s
     WHERE w.begin_interval_time >= to_timestamp(:v_begin_time , 'yyyymmddhh24')
         AND w.end_interval_time <= to_timestamp(:v_end_time , 'yyyymmddhh24')
         AND w.snap_id = s.snap_id
         AND w.dbid = s.dbid
         AND w.instance_number = s.instance_number
         AND s.parsing_schema_name <> 'SYS'
     GROUP BY s.dbid ,
              s.force_matching_signature ,
              s.parsing_schema_name ,
              s.module
     ORDER BY count(*) DESC ) a
WHERE rownum<=10;

12.2 AWR SQL ordered BY Elapsed Time[편집]

def dbid="&1" 
def inst_num="&2" 
def snap_start="&3" 
def snap_end="&4" 
def top_num="&5" 
col ela_Time format 999990.00 heading "ElaTime|(s)" 
col EXEC format 9999990 heading "Exec" 
col elapexec format 999990.00 heading "ElaTimeperexec (s)" 
--col Total               format 9999990.00         heading "DBTime%"
col CPU format 999990.00 heading "CPU%" 
col IO format 999990.00 heading "IO%" 
col SQLId format a15 heading "SQL_ID" 
col SQLModule format a50 heading "SQL_Module" 
col SQLText format a50 heading "SQL_Text" 

WITH sqt AS
(SELECT elap,
        cput,
        EXEC,
        norm_val,
        iowt,
        sql_id,
        module,
        rnum
 FROM
     (SELECT sql_id,
             module,
             elap,
             norm_val,
             cput,
             EXEC,
             iowt,
             rownum rnum
      FROM
          (SELECT sql_id,
                  MAX(module) module,
                              SUM(elapsed_time_delta) elap, (100 * (SUM(elapsed_time_delta) / NULLIF(
                                                                                                         (SELECT (SUM(e.VALUE) - SUM(b.VALUE)) / 1000000 / 60
                                                                                                          FROM DBA_HIST_SYS_TIME_MODEL e, DBA_HIST_SYS_TIME_MODEL b
                                                                                                          WHERE e.SNAP_ID = &snap_end
                                                                                                              AND e.DBID = &dbid
                                                                                                              AND e.INSTANCE_NUMBER = &inst_num
                                                                                                              AND e.STAT_NAME = 'DB time'
                                                                                                              AND b.SNAP_ID = &snap_start
                                                                                                              AND b.DBID = &dbid
                                                                                                              AND b.INSTANCE_NUMBER = &inst_num
                                                                                                              AND b.STAT_NAME = 'DB time'), 0))) norm_val,
                                                                                                                                                 SUM(cpu_time_delta) cput,
                                                                                                                                                                     SUM(executions_delta) EXEC,
                                                                                                                                                                                           SUM(iowait_delta) iowt
           FROM dba_hist_sqlstat
           WHERE dbid = &dbid
               AND instance_number = &inst_num
               AND &snap_start < snap_id
               AND snap_id <= &snap_end
           GROUP BY sql_id
           ORDER BY nvl(SUM(elapsed_time_delta), -1) DESC, sql_id))
 WHERE rnum < &top_num --  and (rnum <= :tsql_min or norm_val > :top_pct_sql)
 )
;

SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.elap / 1000000), to_number(NULL)) ela_Time,
                            sqt.EXEC EXEC,
                                     decode(sqt.EXEC, 0, to_number(NULL), (sqt.elap / sqt.EXEC / 1000000)) elapexec,
                                     sqt.norm_val Total ,
                                     --(100 * (sum(elap) /nullif(dbtime.dbtime,0))) norm_val
 --null Total,

                                     decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))) CPU,
                                     decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.iowt / sqt.elap))) IO,
                                     sqt.sql_id SQLId,
                                     to_clob(decode(sqt.module, NULL, NULL, 'Module: ' || sqt.module)) SQLModule,
                                     nvl(substr(st.sql_text, 0, 50), to_clob('** SQL Text Not Available **')) SQLText
FROM sqt,
     dba_hist_sqltext st
WHERE st.sql_id(+) = sqt.sql_id
AND st.dbid(+) = &dbid
ORDER BY sqt.rnum 
/

12.3 AWR SQL ordered BY Elapsed Time[편집]

undef dbid 
undef inst_num 
undef snap_start 
undef snap_end 
undef sql_id 
undef top_num 

WITH sqt AS
(SELECT elap,
        cput,
        EXEC,
        iowt,
        norm_val,
        sql_id,
        module,
        rnum
 FROM
     (SELECT sql_id,
             module,
             elap,
             norm_val,
             cput,
             EXEC,
             iowt,
             rownum rnum
      FROM
          (SELECT sql_id ,
                  MAX(module) module ,
                              SUM(elapsed_time_delta) elap , (100 * (SUM(elapsed_time_delta) / NULLIF(:dbtime,0))) norm_val ,
                                                                                                                   SUM(cpu_time_delta) cput ,                                                                                                                                       SUM(executions_delta) EXEC ,                                                                                                                                                             SUM(iowait_delta) iowt
           FROM dba_hist_sqlstat
           WHERE dbid = :dbid
               AND instance_number = :inst_num
               AND :bid < snap_id
               AND snap_id <= :eid
           GROUP BY sql_id
           ORDER BY nvl(SUM(elapsed_time_delta), -1) DESC, sql_id))
 WHERE rnum < :tsql_max
     AND (rnum <= :tsql_min
          OR norm_val > :top_pct_sql))
;

SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.elap/1000000), to_number(NULL)),
                            sqt.EXEC,
                                decode(sqt.EXEC, 0, to_number(NULL),(sqt.elap / sqt.EXEC / 1000000)),
                                sqt.norm_val,
                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))),
                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.iowt / sqt.elap))),
                                sqt.sql_id,
                                to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
                                nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
FROM sqt,
     dba_hist_sqltext st
WHERE st.sql_id(+) = sqt.sql_id
AND st.dbid(+) = :dbid
ORDER BY sqt.rnum 
;

12.4 AWR SQL ordered BY CPU Time[편집]

WITH sqt AS
(SELECT elap,
        cput,
        EXEC,
        uiot,
        norm_val,
        sql_id,
        module,
        rnum
 FROM
     (SELECT sql_id,
             module,
             elap,
             norm_val,
             cput,
             EXEC,
             uiot,
             rownum rnum
      FROM
          (SELECT sql_id ,
                  MAX(module) module ,
                              SUM(elapsed_time_delta) elap , (100 * (SUM(cpu_time_delta) / NULLIF(:tcpu,0))) norm_val ,
                                                                                                             SUM(cpu_time_delta) cput ,                                                                                                                                 SUM(executions_delta) EXEC ,                                                                                                                                                       SUM(iowait_delta) uiot
           FROM dba_hist_sqlstat
           WHERE dbid = :dbid
               AND instance_number = :inst_num
               AND :bid < snap_id
               AND snap_id <= :eid
           GROUP BY sql_id
           ORDER BY nvl(SUM(cpu_time_delta), -1) DESC, sql_id))
 WHERE rnum < :tsql_max
     AND (rnum <= :tsql_min
          OR norm_val > :top_pct_sql))
;

SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.cput/1000000), to_number(NULL)),
                            sqt.EXEC,
                                decode(sqt.EXEC, 0, to_number(NULL), (sqt.cput / sqt.EXEC /1000000)),
                                sqt.norm_val,
                                nvl((sqt.elap/1000000), to_number(NULL)),
                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))),
                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.uiot / sqt.elap))),
                                sqt.sql_id,
                                to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
                                nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
FROM sqt,
     dba_hist_sqltext st
WHERE st.sql_id(+) = sqt.sql_id
AND st.dbid(+) = :dbid
ORDER BY sqt.rnum 
;

12.5 AWR SQL ordered BY USER I/O Wait Time[편집]

WITH sqt AS
(SELECT elap,
        cput,
        EXEC,
        uiot,
        norm_val,
        sql_id,
        module,
        rnum
 FROM
     (SELECT sql_id,
             module,
             elap,
             norm_val,
             cput,
             EXEC,
             uiot,
             rownum rnum
      FROM
          (SELECT sql_id ,
                  MAX(module) module ,
                              SUM(elapsed_time_delta) elap , (100 * (SUM(iowait_delta) / NULLIF(:uiowt,0))) norm_val ,                                                                                                            SUM(cpu_time_delta) cput ,                                                                                                                                SUM(executions_delta) EXEC ,                                                                                                                                                      SUM(iowait_delta) uiot
           FROM dba_hist_sqlstat
           WHERE dbid = :dbid
               AND instance_number = :inst_num
               AND :bid < snap_id
               AND snap_id <= :eid
           GROUP BY sql_id
           ORDER BY nvl(SUM(iowait_delta), -1) DESC, sql_id))
 WHERE rnum < :tsql_max
     AND (rnum <= :tsql_min
          OR norm_val > :top_pct_sql))
;

SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.uiot/1000000), to_number(NULL)),
                            sqt.EXEC,
                                decode(sqt.EXEC, 0, to_number(NULL), (sqt.uiot / sqt.EXEC /1000000)),
                                sqt.norm_val,
                                nvl((sqt.elap/1000000), to_number(NULL)),
                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))),
                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.uiot / sqt.elap))),
                                sqt.sql_id,
                                to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
                                nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
FROM sqt,
     dba_hist_sqltext st
WHERE st.sql_id(+) = sqt.sql_id
AND st.dbid(+) = :dbid
ORDER BY sqt.rnum 
;

12.6 AWR SQL ordered BY Gets[편집]

WITH sqt AS
(SELECT elap,
        cput,
        EXEC,
        uiot,
        bget,
        norm_val,
        sql_id,
        module,
        rnum
 FROM
     (SELECT sql_id,
             module,
             elap,
             norm_val,
             cput,
             EXEC,
             uiot,
             bget,
             rownum rnum
      FROM
          (SELECT sql_id ,
                  MAX(module) module ,
                              SUM(elapsed_time_delta) elap , (100 * (SUM(buffer_gets_delta) / NULLIF(:slr,0))) norm_val ,                                                                                                               SUM(cpu_time_delta) cput ,                                                                                                                                   SUM(executions_delta) EXEC ,                                                                                                                                                         SUM(iowait_delta) uiot ,                                                                                                                                                                           SUM(buffer_gets_delta) bget
           FROM dba_hist_sqlstat
           WHERE dbid = :dbid
               AND instance_number = :inst_num
               AND :bid < snap_id
               AND snap_id <= :eid
           GROUP BY sql_id
           ORDER BY nvl(SUM(buffer_gets_delta), -1) DESC, sql_id))
 WHERE rnum < :tsql_max
     AND (rnum <= :tsql_min
          OR norm_val > :top_pct_sql))
;

SELECT /*+ NO_MERGE(sqt) */ sqt.bget,
                            sqt.EXEC,
                                decode(sqt.EXEC, 0, to_number(NULL), (sqt.bget / sqt.EXEC)),
                                sqt.norm_val,
                                nvl((sqt.elap/1000000), to_number(NULL)),
                                decode(sqt.elap, 0, '     ', lpad(to_char(round((100 * (sqt.cput / sqt.elap)),1), 'TM9'),5)),
                                decode(sqt.elap, 0, '     ', lpad(to_char(round((100 * (sqt.uiot / sqt.elap)),1), 'TM9'),5)),
                                sqt.sql_id,
                                to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
                                nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
FROM sqt,
     dba_hist_sqltext st
WHERE st.sql_id(+) = sqt.sql_id
AND st.dbid(+) = :dbid
ORDER BY sqt.rnum 
;

12.7 AWR SQL ordered BY READS[편집]

WITH sqt AS
(SELECT elap,
        cput,
        EXEC,
        uiot,
        dskr,
        norm_val,
        sql_id,
        module,
        rnum
 FROM
     (SELECT sql_id,
             module,
             elap,
             norm_val,
             cput,
             EXEC,
             uiot,
             dskr,
             rownum rnum
      FROM
          (SELECT sql_id ,
                  MAX(module) module ,
                              SUM(elapsed_time_delta) elap , (100 * (SUM(disk_reads_delta) / NULLIF(:phyr,0))) norm_val ,                                                                                                               SUM(cpu_time_delta) cput ,                                                                                                                                   SUM(executions_delta) EXEC ,                                                                                                                                                         SUM(iowait_delta) uiot ,                                                                                                                                                                           SUM(disk_reads_delta) dskr
           FROM dba_hist_sqlstat
           WHERE dbid = :dbid
               AND instance_number = :inst_num
               AND :bid < snap_id
               AND snap_id <= :eid
           GROUP BY sql_id
           ORDER BY nvl(SUM(disk_reads_delta), -1) DESC, sql_id))
 WHERE rnum < :tsql_max
     AND (rnum <= :tsql_min
          OR norm_val > :top_pct_sql))
;

SELECT /*+ NO_MERGE(sqt) */ sqt.dskr,
                            sqt.EXEC,
                                decode(sqt.EXEC, 0, to_number(NULL), (sqt.dskr / sqt.EXEC)),
                                sqt.norm_val,
                                nvl((sqt.elap / 1000000), to_number(NULL)),
                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))),
                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.uiot / sqt.elap))),
                                sqt.sql_id,
                                to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
                                nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
FROM sqt,
     dba_hist_sqltext st
WHERE st.sql_id(+) = sqt.sql_id
AND st.dbid(+) = :dbid
ORDER BY sqt.rnum 
;

12.8 AWR SQL ordered BY READS[편집]

WITH sqt AS
(SELECT EXEC,
        pdr,
        prq,
        porq,
        norm_val,
        sql_id,
        module,
        rnum
 FROM
     (SELECT sql_id,
             module,
             norm_val,
             EXEC,
             pdr,
             prq,
             porq,
             rownum rnum
      FROM
          (SELECT sql_id ,
                  MAX(module) module , (100 * ((SUM(physical_read_requests_delta) - SUM(optimized_physical_reads_delta)) / NULLIF(:phydrq,0))) norm_val ,                                                                                                                                               SUM(executions_delta) EXEC ,                                                                                                                                                                     SUM(physical_read_requests_delta) - SUM(optimized_physical_reads_delta) pdr ,
SUM(physical_read_requests_delta) prq ,
                                  SUM(optimized_physical_reads_delta) porq
           FROM dba_hist_sqlstat
WHERE dbid = :dbid
AND instance_number = :inst_num
AND :bid < snap_id
AND snap_id <= :eid
           GROUP BY sql_id
           ORDER BY nvl(SUM(physical_read_requests_delta) - SUM(optimized_physical_reads_delta), -1) DESC, sql_id))
WHERE rnum < :tsql_max
AND (rnum <= :tsql_min
 OR norm_val > :top_pct_sql))
;

SELECT /*+ NO_MERGE(sqt) */ sqt.pdr,
                sqt.prq,
                sqt.EXEC,
        decode(sqt.EXEC, 0, to_number(NULL), (sqt.prq/ sqt.EXEC)),
        decode(sqt.prq, 0, to_number(NULL), (100 * (sqt.porq / sqt.prq))),
        sqt.norm_val,
        sqt.sql_id,
        to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
        nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
FROM sqt,
dba_hist_sqltext st
WHERE st.sql_id(+) = sqt.sql_id
AND st.dbid(+) = :dbid
ORDER BY sqt.rnum 
;

12.9 AWR SQL ordered BY Sharable Memory[편집]

WITH sqt AS
(SELECT EXEC,
sharable_mem,
sql_id,
module,
rnum
FROM
(SELECT sql_id,
    module,
    EXEC,
    sharable_mem,
    rownum rnum
FROM
 (SELECT sql_id,
         module,
         EXEC,
         sharable_mem
  FROM
      (SELECT sharable_mem,
              sql_id
       FROM dba_hist_sqlstat
       WHERE dbid = :dbid
           AND snap_id = :eid
           AND instance_number = :inst_num
           AND sharable_mem > :shmem_thres) y
  LEFT OUTER JOIN
      (SELECT sql_id ,
              MAX(module) module ,
                          SUM(executions_delta) EXEC
       FROM dba_hist_sqlstat
       WHERE dbid = :dbid
           AND instance_number = :inst_num
           AND :bid < snap_id
           AND snap_id <= :eid
       GROUP BY sql_id) x USING (sql_id)
  ORDER BY nvl(sharable_mem, -1) DESC, sql_id))
WHERE rnum <= :tsql_max)
;

SELECT /*+ NO_MERGE(sqt) */ sqt.sharable_mem,
                sqt.EXEC,
        decode(:espm, 0, 0, 100 * sqt.sharable_mem/:espm),
        sqt.sql_id,
        to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
        nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
FROM sqt,
dba_hist_sqltext st
WHERE st.sql_id(+) = sqt.sql_id
AND st.dbid(+) = :dbid
ORDER BY sqt.rnum 
;

12.10 AWR SQL ordered BY VERSION COUNT[편집]

WITH sqt AS
(SELECT EXEC,
version_count,
sql_id,
module,
rnum
FROM
(SELECT sql_id,
    module,
    EXEC,
    version_count,
    rownum rnum
FROM
 (SELECT sql_id,
         module,
         EXEC,
         version_count
  FROM
      (SELECT version_count,
              sql_id
       FROM dba_hist_sqlstat
       WHERE dbid = :dbid
           AND snap_id = :eid
           AND instance_number = :inst_num
           AND version_count > :vcnt_thres) y
  LEFT OUTER JOIN
      (SELECT sql_id ,
              MAX(module) module ,
                          SUM(executions_delta) EXEC
       FROM dba_hist_sqlstat
       WHERE dbid = :dbid
           AND instance_number = :inst_num
           AND :bid < snap_id
           AND snap_id <= :eid
       GROUP BY sql_id) x USING (sql_id)
  ORDER BY nvl(y.version_count, -1) DESC, sql_id))
WHERE rnum <= :tsql_max)
;

SELECT /*+ NO_MERGE(sqt) */ sqt.version_count,
                sqt.EXEC,
        sqt.sql_id,
        to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
        nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
FROM sqt,
dba_hist_sqltext st
WHERE st.sql_id(+) = sqt.sql_id
AND st.dbid(+) = :dbid
ORDER BY sqt.rnum

원본 주소 "http://dbcafe.co.kr/wiki/index.php?title=ORACLE_모니터링&oldid=10585"

@@ 3번째 줄: / 3번째 줄: @@
 ----
-== 세션/접속사용자 정보 ==
+= 세션 정보 =
+== User 별 Session 정보 조회 ==
-=== 세션 죽이기(SID,SERAIL#) ===
+* V$SESSION
+* V$PROCESS
 <source lang="sql">
-ALTER SYSTEM KILL SESSION '8,4093'
+SELECT S.USERNAME
+     , S.SID
+     , S.SERIAL#
+     , P.SPID
+     , S.OSUSER
+     , S.MACHINE
+     , S.PROGRAM
+     , TO_CHAR(S.LOGON_TIME, 'MM/DD HH24:MI') "LOGON_TIME"
+     , ROUND(S.LAST_CALL_ET / 60) "IDLE"
+  FROM V$SESSION S
+     , V$PROCESS P
+ WHERE S.PADDR = P.ADDR
+   AND S.USERNAME LIKE UPPER('&DBUSER%')
+ ORDER BY 9;
+</source>
-=== 오라클 세션과 관련된 테이블 ===
+== 세션 죽이기(SID,SERAIL#) ==
-* v$session
 <source lang="sql">
- select count(*)
+ALTER SYSTEM KILL SESSION '8,4093' immediate;
-   from v$session
+-- RAC
-  where machine ='머신이름'
+ALTER SYSTEM KILL SESSION '8,4093,@1' immediate;
-    and schemaname ='스키마이름'
 </source>
-=== 현재 커서 수 확인 ===
+== 현재 TOP  세션 gv$session ==
-* V$OPEN_CURSOR
+<source lang=sql>
-* v$session_wait
+select  inst_id||'_'||sid||' '||serial# inst_sid_ser,
-* v$transaction
+    username||case when regexp_substr(program,' \(...') <> ' (TNS' then regexp_substr(program,' \(.+') end username,
-* v$session_wait
+    sql_id sql_id,
+    round((sysdate-sql_exec_start)*24*3600,1) sql_dT,
+        last_call_et call_dT,
+    case state when 'WAITING' then round(wait_time_micro/1000000,2) else round(time_since_last_wait_micro/1000000,2) end W_dT,
+        decode(state,'WAITING',event,'CPU') event,
+    service_name||' '||substr(module,1,20)||' '||ACTION serv_mod_action,
+          nullif(row_wait_obj#,-1) obj#,decode(taddr,null,null,'NN') tr
+from gv$session
+where ((state='WAITING' and wait_class<>'Idle') or (state<>'WAITING' and status='ACTIVE'))
+      --and audsid != to_number(sys_context('USERENV','SESSIONID')) -- this is clean but does not work on ADG so replaced by following line
+      and (machine,port) <> (select machine,port from v$session where sid=sys_context('USERENV','SID')) --workaround for ADG
+order by inst_id,sql_id
+</source>
-<source lang="sql">
+== 오래 걸리는  세션 정보 ==
--- sid별 열린커셔
-SELECT sid, count(sid) cursor
-  FROM V$OPEN_CURSOR
-  WHERE user_name = 'ilips'
- GROUP BY sid
- ORDER BY cursor DESC;
--- sql 별 열린 커서
+== 현재 작업중인 세션 정보 V$SESSION_LONGOPS ==
-SELECT sql_text, count(sid) cnt
+* V$SESSION_LONGOPS
-   FROM v$OPEN_CURSOR
+* V$SESSION
-  GROUP BY sql_text
+<source lang=sql>
-  ORDER BY cnt DESC
+SELECT B.USERNAME, A.SID, B.OPNAME, B.TARGET
+     , ROUND(B.SOFAR*100/B.TOTALWORK,0) || '%' AS "%DONE", B.TIME_REMAINING
+     , TO_CHAR(B.START_TIME,'YYYY/MM/DD HH24:MI:SS') START_TIME
+   FROM V$SESSION_LONGOPS B
+     , V$SESSION A
+  WHERE A.SID = B.SID
+  ORDER BY 6 DESC;
+</source>
+== 세션당 리소스 사용률 gv$sessmetric ==
+* gv$sessmetric
+<source lang=sql>
+select * from (
+  select inst_id, session_id sid, to_char(begin_time,'hh24:mi:ss') begTime, round(intsize_csec/100,0) D_sec,  cpu,
+         physical_reads PhyReads, logical_reads LogicalReads, pga_memory, hard_parses, soft_parses
+  from gv$sessmetric
+  order by cpu+physical_reads desc
+)
+where rownum<20
+</source>
--- 대기 세션
+== 대기중인 세션 ==
+* v$session_wait
+<source lang="sql">
 select *
    from v$session_wait;
-select sid, serial#, username, taddr, used_ublk, used_urec
-  from v$transaction t
-     , v$session s
-where t.addr = s.taddr;
 </source>
+== 1시간 이상 idle 상태인 세션 ==
-=== 연결되어 있는 OS 사용자 및 프로그램 조회 ===
-* V$SESSION
-<source lang="sql">
-SELECT SID
-     , SERIAL#
-     , OSUSER
-     , SUBSTRB(USERNAME, 1, 10) AS USER_NAME
-     , SUBSTRB(PROGRAM, 1, 30) AS PROGRAM_NAME
-     , STATUS
-     , TO_CHAR(LOGON_TIME, 'YYYY/MM/DD HH:MI') AS LOGON_TIME
-  FROM V$SESSION
- WHERE TYPE!= ‘BACKGROUND’ AND STATUS = ‘ACTIVE’; </source>
-=== 1시간 이상 유휴 상태인 세션 ===
 * V$SESSION
@@ 77번째 줄: / 96번째 줄: @@
 </source>
-=== 프로세스 아이디를 이용하여 쿼리문 알아내기 ===
+== Active Session 중 Idle Time이 긴 작업 ==
-* v$process
+* V$SESSION
-* v$session
+* V$PROCESS
-* v$sqltext
-<source lang="sql">
-select c.sql_text
-     , b.SID
-     , b.SERIAL#
-     , b.machine
-     , b.OSUSER
-     , b.logon_time --이 쿼리를 호출한 시간
-  from v$process a, v$session b, v$sqltext c
- where a.addr = b.paddr
-   and b.sql_hash_value = c.hash_value
-   and a.spid = '1708032' --1912870/
- order by c.PIECE
-</source>
-=== Active Session 중 Idle Time이 긴 작업 ===
-* V$SESSION
-* V$PROCESS
 <source lang="sql">
 SELECT VS.SID || ',' || VS.SERIAL# " SID"
@@ 115번째 줄: / 115번째 줄: @@
 </source>
-=== DBUser 별로 Session 정보를 조회 ===
+== 사용자 세션 중에서 2시간 이상 idle 상태가 지속되는 세션 kill ==
 * V$SESSION
 * V$PROCESS
-<source lang="sql">
+<source lang="sql"> SET PAGESIZE 0 SPOOL KILLIDLE3.SQL
-SELECT S.USERNAME
+SELECT DISTINCT '!KILL -9 ' || B.SPID, 'ALTER SYSTEM KILL SESSION '''|| A.SID || ',' || A.SERIAL# || ''';'
-     , S.SID
+  FROM V$SESSION A
-     , S.SERIAL#
+      , V$PROCESS B
-      , P.SPID
+ WHERE A.PADDR IN (SELECT S.PADDR
-     , S.OSUSER
+                     FROM V$SESSION S
-     , S.MACHINE
+                    WHERE STATUS = 'INACTIVE'
-     , S.PROGRAM
+                    GROUP BY  S.PADDR
-     , TO_CHAR(S.LOGON_TIME, 'MM/DD HH24:MI') "LOGON_TIME"
+                   HAVING MIN(ROUND(LAST_CALL_ET / 60)) > 120)
-     , ROUND(S.LAST_CALL_ET / 60) "IDLE"
+   AND A.PADDR = B.ADDR AND A.STATUS = 'INACTIVE';
-  FROM V$SESSION S
+SPOOL OFF
-     , V$PROCESS P
- WHERE S.PADDR = P.ADDR
-   AND S.USERNAME LIKE UPPER('&DBUSER%')
- ORDER BY 9;
 </source>
-=== Session별 사용 명령어 ===
+== 트랜젝션 중인 세션 ==
-* V$SESSION SESS
+* v$transaction
-* V$SESSTAT STAT
+* gv$fast_start_transactions
-* V$STATNAME NAME
-* V$PROCESS PROC
+* USED_UBLK
+**  Number of undo blocks used ,  사용된 언두 블럭 수
+* USED_UREC
+**  Number of undo records used , 사용된 언두 레코드 수
 <source lang="sql">
-SELECT SESS.SID
+select sid, serial#, username, taddr, used_ublk, used_urec
-      , SESS.SERIAL#
+  from v$transaction t
-      , SUBSTR(SESS.USERNAME, 1, 10) "USER NAME"
+     , v$session s
-      , SUBSTR(OSUSER, 1, 11) "OS USER"
+where t.addr = s.taddr;
-      , SUBSTR(SESS.MACHINE, 1, 15) "MACHINE NAME"
+</source>
+<source lang="sql">
+select inst_id,addr,start_time,used_ublk,xid
+  from gv$transaction;
+</source>
+<source lang="sql">
+select inst_id,state,undoblocksdone,undoblockstotal,xid
+ from gv$fast_start_transactions;
+</source>
+<source lang="sql">
+SELECT sess.sid, sess.status, sess.username, machine, sql_id, prev_sql_id, trans.USED_UBLK, trans.start_date
+from gv$session sess
+   , gv$transaction trans
+WHERE sess.taddr=trans.addr and sess.inst_id=trans.inst_id;
+</source>
+== 연결되어 있는 OS 사용자 및 프로그램 조회 ==
+* V$SESSION
+<source lang="sql">
+SELECT SID
+      , SERIAL#
+      , OSUSER
+      , SUBSTRB(USERNAME, 1, 10) AS USER_NAME
+      , SUBSTRB(PROGRAM, 1, 30) AS PROGRAM_NAME
       , STATUS
-      , UPPER(
+     , TO_CHAR(LOGON_TIME, 'YYYY/MM/DD HH:MI') AS LOGON_TIME
-             DECODE(NVL(COMMAND, 0)
+  FROM V$SESSION
-               , 0, '---'
+ WHERE TYPE!= 'BACKGROUND' AND STATUS = 'ACTIVE';
-               , 1, 'CREATE TABLE'
+</source>
+== Session별 사용 명령어 ==
+* V$SESSION SESS
+* V$SESSTAT STAT
+* V$STATNAME NAME
+* V$PROCESS PROC
+<source lang="sql">
+SELECT SESS.SID
+     , SESS.SERIAL#
+     , SUBSTR(SESS.USERNAME, 1, 10) "USER NAME"
+     , SUBSTR(OSUSER, 1, 11) "OS USER"
+     , SUBSTR(SESS.MACHINE, 1, 15) "MACHINE NAME"
+     , STATUS
+      , UPPER(
+             DECODE(NVL(COMMAND, 0)
+               , 0, '---'
+               , 1, 'CREATE TABLE'
                , 2, 'INSERT -'
                , 3, 'SELECT -'
@@ 244번째 줄: / 288번째 줄: @@
   ORDER BY 3, 1, 2;
 </source>
+----
+= SQL 찾기 =
+== SQL 통계 정보확인하기 ==
+== Module LEVEL 통계 ==
+<source lang="sql">
+    SELECT *
+    FROM
+        (SELECT module ,
+                count(*) sql_cnt ,
+                sum(executions) executions ,
+                round(avg(buffer_gets/executions)) "lio(avg)" ,
+                round(avg(disk_reads/executions)) "pio(avg)" ,
+                round(avg(rows_processed/executions)) "rows(avg)" ,
+                round(avg(elapsed_time/executions/1000000),2) "elapsed(avg)" ,
+                count(CASE WHEN elapsed_time/executions/1000000>=3 THEN 1 END) bad_sql ,
+                round(max(elapsed_time/executions/1000000),2) "elapsed(max)"
+         FROM v$sql
+         WHERE executions > 0
+         GROUP BY module)
+ORDER BY "lio(avg)" * executions DESC ;
+</source>
-=== 사용자 session 중에서 2시간 이상 idle 상태가 지속되는 session을 kill   ===
-* V$SESSION
-* V$PROCESS
-<source lang="sql"> SET PAGESIZE 0 SPOOL KILLIDLE3.SQL
-SELECT DISTINCT '!KILL -9 ' || B.SPID, 'ALTER SYSTEM KILL SESSION '''|| A.SID || ',' || A.SERIAL# || ''';'
-  FROM V$SESSION A
-     , V$PROCESS B
- WHERE A.PADDR IN (SELECT S.PADDR
-                     FROM V$SESSION S
-                    WHERE STATUS = 'INACTIVE'
-                    GROUP BY  S.PADDR
-                   HAVING MIN(ROUND(LAST_CALL_ET / 60)) > 120)
-   AND A.PADDR = B.ADDR AND A.STATUS = 'INACTIVE';
-SPOOL OFF
-</source>
-----
-== 프로세스 정보 ==
-=== Oracle Process의 정보 ===
-* V$SESSION
-* V$PROCESS
-* SYS.V_$SESS_IO
+== SCHEMA LEVEL 통계 ==
 <source lang="sql">
-SELECT S.STATUS "STATUS"
+SELECT * --9i
-     , S.SERIAL# "SERIAL#"
+FROM
-     , S.TYPE "TYPE"
+    ( SELECT
-     , S.USERNAME "DB USER"
+         (SELECT username
-     , S.OSUSER "CLIENT USER"
+          FROM dba_users
-     , S.SERVER "SERVER"
+          WHERE user_id = parsing_schema_id) AS SCHEMA_NAME ,
-     , S.MACHINE "MACHINE"
+             count(*) sql_cnt ,
-     , S.MODULE "MODULE"
+             sum(executions) executions ,
-     , S.TERMINAL "TERMINAL"
+             round(avg(buffer_gets/executions)) "lio(avg)" ,
-     , S.PROGRAM "PROGRAM"
+             round(avg(disk_reads/executions)) "pio(avg)" ,
-     , P.PROGRAM "O.S. PROGRAM"
+             round(avg(rows_processed/executions)) "rows(avg)" ,
-     , S.LOGON_TIME "CONNECT TIME"
+             round(avg(elapsed_time/executions/1000000),2) "elapsed(avg)" ,
-     , LOCKWAIT "LOCK WAIT"
+             count(CASE WHEN elapsed_time/executions/1000000>=10 THEN 1 END) bad_sql ,
-     , SI.PHYSICAL_READS "PHYSICAL READS"
+             round(max(elapsed_time/executions/1000000),2) "elapsed(max)"
-     , SI.BLOCK_GETS "BLOCK GETS"
+      FROM v$sqlarea
-     , SI.CONSISTENT_GETS "CONSISTENT GETS"
+      WHERE executions > 0
-     , SI.BLOCK_CHANGES "BLOCK CHANGES"
+      GROUP BY parsing_schema_id)
-     , SI.CONSISTENT_CHANGES "CONSISTENT CHANGES"
+ORDER BY "lio(avg)" * executions DESC ;
-     , S.PROCESS "PROCESS"
+</source>
-      , P.SPID
-     , P.PID
-     , S.SERIAL#
-     , SI.SID
-     , S.SQL_ADDRESS "ADDRESS"
-     , S.SQL_HASH_VALUE "SQL HASH"
-     , S.ACTION
-  FROM V$SESSION S
-      , V$PROCESS P
-      , SYS.V_$SESS_IO SI
- WHERE S.PADDR = P.ADDR(+)
-   AND SI.SID(+) = S.SID
-   AND S.USERNAME IS NOT NULL
-   AND NVL(S.OSUSER, 'X') <> 'SYSTEM'
-   AND S.TYPE <> 'BACKGROUND'
- ORDER BY 3;
-</source>
-=== 오브젝트에 접속되어 있는 프로그램 조회   ===
+<source lang="sql">
-* V$SESSION
+SELECT * --10g
-* V$ACCESS
+FROM
-<source lang="sql">
+    (SELECT parsing_schema_name ,
-SELECT SUBSTR(B.OBJECT, 1, 15) AS OBJECT, SUBSTR(A.PROGRAM, 1, 15) AS PROGRAM, COUNT(*) AS CNT
+            count(*) sql_cnt ,
-  FROM V$SESSION A
+            sum(executions) executions ,
-      , V$ACCESS B
+            round(avg(buffer_gets/executions)) "lio(avg)" ,
- WHERE A.SID = B.SID
+            round(avg(disk_reads/executions)) "pio(avg)" ,
-   AND B.OWNER NOT IN ('SYS')
+            round(avg(rows_processed/executions)) "rows(avg)" ,
-   AND A.TYPE!= 'BACKGROUND'
+            round(avg(elapsed_time/executions/1000000),2) "elapsed(avg)" ,
-   AND B.OBJECT LIKE UPPER('&OBJECT_NAME') || '%'
+            count(CASE WHEN elapsed_time/executions/1000000>=10 THEN 1 END) bad_sql ,
- GROUP BY B.OBJECT, SUBSTR(A.PROGRAM, 1, 15);
+            round(max(elapsed_time/executions/1000000),2) "elapsed(max)"
-</source>
+     FROM v$sqlarea
+      WHERE executions > 0
+     GROUP BY parsing_schema_name)
+ORDER BY "lio(avg)" * executions DESC ;
+</source>
-----
-== 부하 발생 모니터링 ==
+== DB Time 분석하기 ==
-=== cpu를 많이 사용하는 쿼리문과 프로세스아이디,시리얼번호,머신 알아내기 ===
+<source lang="sql">
-* v$process
+SELECT STAT_NAME ,
-* v$session
+       VALUE,
-* v$sqltext
+       ROUND(VALUE/
+                 ( SELECT VALUE
+                  FROM V$SYS_TIME_MODEL
+                  WHERE STAT_NAME='DB TIME' ) *100 , 2) TIME_RATIO
+FROM V$SYS_TIME_MODEL
+WHERE STAT_NAME NOT IN ( 'BACKGROUND ELAPSED TIME' ,
+                         'BACKGROUND CPU TIME' ,
+                         'RMAN CPU TIME (BACKUP/RESTORE)' ,
+                         'HARD PARSE ELAPSED TIME' ,
+                         'HARD PARSE (SHARING CRITERIA) ELAPSED TIME' ,
+                         'HARD PARSE (BIND MISMATCH) ELAPSED TIME' ,
+                         'FAILED PARSE ELAPSED TIME' ,
+                         'FAILED PARSE (OUT OF SHARED MEMORY) ELAPSED TIME' ,
+                         'DB CPU')
+ORDER BY TIME_RATIO DESC ;
+</source>
-<source lang="sql">
+== SQL 추출하기 (10g) ==
-select c.sql_text
+<source lang="sql">
-     , b.SID
+    SELECT *
-     , b.SERIAL#
+    FROM
-     , b.machine
+        ( SELECT rownum cnt,
-     , b.OSUSER
+                 t1.*
-     , b.logon_time --이 쿼리를 호출한 시간
+         FROM
-  from v$process a
+             ( SELECT parsing_schema_name SCHEMA,    --> 1
-     , v$session b
+                                          module,    --> 2
-     , v$sqltext c
+                                          sql_id,    --> 3
- where a.addr = b.paddr
+                                          hash_value,--> 4
-   and b.sql_hash_value = c.hash_value
+                                          substr(sql_text,1,100) substr_sqltext, --> 5
---and a.spid = '675958'
+                                          executions,  --> 6
-  order by c.PIECE
+                                          buffer_gets, --> 7 (전체 I/O 처리량)
-</source>
+                                          disk_reads,  --> 8
+                                          rows_processed, --> 9
-=== 롤백 세그먼트 경합 조회  ===
+                                          round(buffer_gets/executions,1) lio, --> 10 (1회 수행 당 I/O)
-* V$ROLLSTAT
+                                          round(elapsed_time/executions/1000000,1) elapsed_sec, --> 11 (1회 수행 당 Elapsed)
-* V$ROLLNAME
+                                          round(cpu_time/executions/1000000,1) cpu_sec, --> 12 (1회 수행 당 CPU)
+                                          round(elapsed_time/tot_elapsed*100,1) ratio_elapsed, --> 13 (전체 대비 Elapsed )
-<source lang="sql">
+                                          round(cpu_time/tot_cpu*100,1) ratio_cpu --> 14 (전체 대비 CPU 사용률)
-SELECT NAME T0
+                            FROM v$sqlarea s,
-     , GETS T1
+                  (SELECT sum(elapsed_time) tot_elapsed,
-     , WAITS T2
+                          sum(cpu_time) tot_cpu
-     , TO_CHAR(TRUNC(WAITS / GETS * 100, 2), 099.99) || '%' T3
+                   FROM v$sqlarea) t
-     , TO_CHAR(ROUND(RSSIZE / 1024)) T4
+              WHERE s.executions > 0
-     , SHRINKS T5
+                  AND parsing_schema_name NOT IN ('SYS','SYSTEM')
-     , EXTENDS T6
+                  AND ((module NOT LIKE 'TOAD%'
-  FROM V$ROLLSTAT
+                        AND module NOT LIKE 'SQL De%'
-     , V$ROLLNAME
+                        AND module NOT LIKE 'Orange%'
- WHERE V$ROLLSTAT.USN = V$ROLLNAME.USN;
+                        AND module NOT LIKE 'PL/SQL%'
-</source>
+                        AND module NOT LIKE 'plsqldev.exe')
+                       OR (module IS NULL))
+              ORDER BY 14 DESC ) t1
+         WHERE rownum <= 50) WHERE cnt >= 1
+;
+</source>
+== 엑사 DB ==
+== Exa . Smart Scan으로 수행되지 않고,Interconnect Bytes가 높은 SQL ==
-=== Buffer Cache Hit Ratio ===
+<source lang="sql">
-* V$SYSSTAT
+WITH Cell_Mon_SQL AS
-<source lang="sql">
+        ( SELECT sql_id,
-SELECT ROUND(((1-(SUM(DECODE(name, 'physical reads', value,0))/(SUM(DECODE(name, 'db block gets', value,0))+ (SUM(DECODE(name, 'consistent gets', value, 0))))))*100),2) || '%' "Buffer Cache Hit Ratio"
+                 io_cell_offload_eligible_bytes AS coeb,
-  FROM V$SYSSTAT;
+                 io_interconnect_bytes AS ib,
+                 io_cell_offload_returned_bytes AS corb,
+                 io_cell_uncompressed_bytes AS cub,
+                 optimized_phy_read_requests AS oprr,
+                 physical_read_requests AS prr,
+                 physical_read_bytes AS prb,
+                 sql_text
+         FROM v$sqlarea)
+    SELECT /*** Smart Scan이 되지 않으면서, Interconnect Bytes가 높은 SQL ***/ *
+    FROM
+        ( SELECT sql_id ,
+                 decode(coeb,0,'No','Yes') AS "Offload",
+                 round(prb/1024/1024/1024,3) AS "Physical_Bytes(GB)",
+                 round(coeb/1024/1024/1024,3) AS "Offload_Bytes(GB)",
+                 round(ib/1024/1024/1024,3) AS "Interconnect_Bytes(GB)",
+                 round(oprr*8192/1024/1024/1024,3) AS "Cell_Flash_Cache(GB)",
+                 round(corb/1024/1024/1024,3) AS "Returned_Bytes(GB)",
+                 round(decode(coeb,0,0,100*(coeb-ib)/coeb),2) AS "Saved_IO(%)",
+                 sql_text
+         FROM Cell_Mon_SQL
+         WHERE coeb = 0
+         ORDER BY ib DESC) WHERE rownum <= 10 ;
 </source>
-=== Library Cache Hit Ratio ===
+== Smart Scan으로 수행되지 않고,Offload Returned Bytes가 높은 SQL  ==
-<source lang="sql">
+* Reverting TO Block Shipping으로 수행된 SQL로 Smart Scan으로 동작할 수 있는 SQL이지만,Direct I/O로 동작하지 않는 경우 (FULL TABLE Scan도 cell single block physical read로 수행됨.)
-SELECT (1-SUM (reloads)/SUM(pins))*100 "Library Cache Hit Ratio"
-From V$LIBRARYCACHE;
+* Chained Row가 있는 블록들을 읽을 때 발생. : Smart Scan을 통해서 Chained Row가 있는 블록을 읽을 때 각 블록이 다른 STORAGE Cell에 있는 경우 STORAGE Cell 간 통신할 수 없기 때문에 Buffer Cache로 전체 블록을 전송한 후 처리하도록 Block Shipping Mode로 변경
+* 읽기 일관성 모드로 변경 중인 데이터를 가진 Block에 대해서 Smart Scan을 중지하고,     Single Block Read를 수행하는 Block Shipping Mode로 변경
+<source lang="sql">
+WITH Cell_Mon_SQL AS
+    ( SELECT sql_id,
+             io_cell_offload_eligible_bytes AS coeb,
+             io_interconnect_bytes AS ib,
+             io_cell_offload_returned_bytes AS corb,
+             io_cell_uncompressed_bytes AS cub,
+             optimized_phy_read_requests AS oprr,
+             physical_read_requests AS prr,
+             physical_read_bytes AS prb,
+             sql_text
+     FROM v$sqlarea)
+SELECT *
+FROM
+    ( SELECT sql_id ,
+             decode(coeb,0,'No','Yes') AS "Offload",
+             round(prb/1024/1024/1024,3) AS "Physical_Bytes(GB)",
+             round(coeb/1024/1024/1024,3) AS "Offload_Bytes(GB)",
+             round(ib/1024/1024/1024,3) AS "Interconnect_Bytes(GB)",
+             round(oprr*8192/1024/1024/1024,3) AS "Cell_Flash_Cache(GB)",
+             round(corb/1024/1024/1024,3) AS "Returned_Bytes(GB)",
+             round(decode(coeb,0,0,100*(coeb-ib)/coeb),2) AS "Saved_IO(%)",
+             sql_text
+     FROM Cell_Mon_SQL
+     WHERE coeb = 0
+     ORDER BY corb DESC)
+WHERE rownum <= 10 ;
 </source>
-=== Data Dictionary Cache Hit Ratio ===
+== Full Table Scan 처리가 있는 SQL문 정보 추출하기(10g) ==
-V$ROWCACHE
+<source lang="sql">
-<source lang="sql">
+SELECT *
-SELECT (1-SUM(getmisses)/SUM(gets))*100 "Data Dictionary Hit Ratio"
+FROM
-  FROM V$ROWCACHE;
+    (SELECT rownum cnt,
-</source>
+            t1.*
+     FROM
-----
+         (SELECT parsing_schema_name,
+                 --> 1
-== DB 용량 관리 ==
+                 module,
-=== 테이블 스페이스 사용량 ===
+                 --> 2
-* dba_data_files
-* dba_free_space
-<source lang="sql">
-SELECT a.tablespace_name,
-             a.total "Total(Mb)",
-             a.total - b.free "Used(Mb)",
-             nvl(b.free,0) "Free(Mb)",
-             round((a.total - nvl(b.free,0))*100/total,0)  "Used(%)"
-from    (   select     tablespace_name,
-                            round((sum(bytes)/1024/1024),0) as total
-               from       dba_data_files
-               group by tablespace_name) a,
-         (     select     tablespace_name,
-                             round((sum(bytes)/1024/1024),0) as free
-               from        dba_free_space
-               group by  tablespace_name) b
-where      a.tablespace_name = b.tablespace_name(+)
-order by   a.tablespace_name;
-</source>
-== 오라클 서버 스펙  ==
-=== 오라클서버의 메모리 ===
-* v$sgastat
-<source lang="sql">
-select * from v$sgastat;
-select pool, sum(bytes) "SIZE"
+                 sql_id,
-  from v$sgastat
+                 --> 3
- where pool = 'shared pool'
- group by pool
-</source>
+                 hash_value,
+                 --> 4
+                 substr(sql_text,1,100) substr_sqltext,
+                 --> 5
+                 executions,
+                 --> 6
-## V$LOCK 을 사용한 잠금 경합 모니터링
+                 buffer_gets,
-<source lang="sql">
+                 --> 7 (전체 I/O 처리량)
-SELECT s.username, s.sid, s.serial#, s.logon_time,
-    DECODE(l.type, 'TM', 'TABLE LOCK',
-          'TX', 'ROW LOCK',
-       NULL) "LOCK LEVEL",
-    o.owner, o.object_name, o.object_type
-FROM v$session s, v$lock l, dba_objects o
-WHERE s.sid = l.sid
-AND o.object_id = l.id1
-AND s.username IS NOT NULL
-</source>
-## 락이 걸린 세션 자세히 알아보기
+                 disk_reads,
-<source lang="sql">
+                 --> 8
-select a.sid, a.serial#,a.username,a.process,b.object_name,
-decode(c.lmode,2,'RS',3,'RX',4,'S',5,'SRX',8,'X','NO') "TABLE LOCK",
-decode (a.command,2,'INSERT',3,'SELECT',6,'UPDATE',7,'DELETE',12,'DROP TABLE',26,'LOCK TABLE','UNknown') "SQL",
-decode(a.lockwait, NULL,'NO wait','Wait') "STATUS"
-from v$session a,dba_objects b, v$lock c
-where a.sid=c.sid and b.object_id=c.id1
-and c.type='TM'
-</source>
-## 락이 걸린 세션 간단히 알아보기
+                 rows_processed,
-<source lang="sql">
+                 --> 9
-select a.sid, a.serial#, b.type, c.object_name, a.program, a.lockwait,
-       a.logon_time, a.process, a.osuser, a.terminal
-from v$session a, v$lock b, dba_objects c
-where a.sid = b.sid
-  and b.id1 = c.object_id
-  and b.type = 'TM';
-select a.sid, a.serial#, a.username, a.process, b.object_name
+                 round(buffer_gets/executions,1) lio,
-from v$session a , dba_objects b, v$lock c
+                 --> 10 (1회 수행 당 I/O)
-where a.sid=c.sid and b.object_id = c.id1
-and c.type = 'TM'
-</source>
---락이 걸린 세션을 찾아 내어 세션을 죽이려고 해도 죽지 않는 경우
---아래 쿼리문으로 OS단의 PROCESS ID를 찾아내어 OS에서 죽인다
---kill -9 프로세스아이디
-<source lang="sql">
+                 round(elapsed_time/executions/1000000,1) elapsed_sec,
-select substr(s.username,1,11) "ORACLE USER", p.pid "PROCESS ID",
+                 --> 11 (1회 수행 당 Elapsed)
-s.sid "SESSION ID", s.serial#, osuser "OS USER",
-p.spid "PROC SPID",s.process "SESS SPID", s.lockwait "LOCK WAIT"
-from v$process p, v$session s, v$access a
-where a.sid=s.sid and
-p.addr=s.paddr and
-s.username != 'SYS'
-</source>
-## 위 쿼리문의 결과가 있다면 락이 걸린 세션이 있다는것이므로 아래의 쿼리문으로 세션을 죽인다
-<source lang="sql">
-ALTER SYSTEM KILL SESSION '11,39061'
+                 round(cpu_time/executions/1000000,1) cpu_sec,
+                 --> 12 (1회 수행 당 CPU)
-</source>
+                 round(elapsed_time/tot_elapsed*100,1) ratio_elapsed,
+                 --> 13 (전체 대비 Elapsed )
+                 round(cpu_time/tot_cpu*100,1) ratio_cpu --> 14 (전체 대비 CPU 사용률)
+          FROM
+              (SELECT /*+ leading(x) no_merge(x) use_hash(x s) */ s.*
+               FROM v$sqlarea s,
+                   ( SELECT DISTINCT hash_value
+                    FROM v$sql_plan
+                    WHERE OPERATION = 'TABLE ACCESS'
+                        AND OPTIONS = 'FULL' --and object_owner like :b1||'%'
+  ) x
+               WHERE x.hash_value = s.hash_value) s,
+              (SELECT sum(elapsed_time) tot_elapsed,
+                      sum(cpu_time) tot_cpu
+               FROM v$sqlarea) t
+          WHERE executions > 0
+              AND parsing_schema_name NOT IN ('SYS',
+                                              'SYSTEM')
+              AND ((module NOT LIKE 'TOAD%'
+                    AND module NOT LIKE 'SQL De%'
+                    AND module NOT LIKE 'Orange%'
+                    AND module NOT LIKE 'PL/SQL%'
+                    AND module NOT LIKE 'plsqldev.exe')
+                   OR (module IS NULL))
+          ORDER BY 10 DESC ) t1
+     WHERE rownum <= 50)
+WHERE cnt >= 1
+</source>
-alter session으로 죽지않는 프로세스 죽이기
+== 하드 파싱(leteral,리터럴) SQL 찾기 ==
 <source lang="sql">
-.oracle이 설치된 서버에 텔넷으로 root로 접속한다
+    SELECT *
-.su -오라클계정
+    FROM
-.sqlplus '/as sysdba''
+        ( SELECT
-.connect system/sys
+             ( SELECT parsing_schema_name
-.ALTER SYSTEM KILL SESSION '137,1723'
+              FROM v$sqlarea
+              WHERE sql_id=a.max_sql_id ) SCHEMA,
+             ( SELECT MODULE
+              FROM v$sqlarea
+              WHERE sql_id=a.max_sql_id ) MODULE, a.literal_sql_cnt, a.execution_cnt, a.plan_cnt, a.max_sql_id,
+             ( SELECT sql_fulltext
+              FROM v$sqlarea
+              WHERE sql_id = a.max_sql_id ) sql_text
+         FROM
+             ( SELECT s.force_matching_signature,
+                      COUNT(s.exact_matching_signature) literal_sql_cnt,
+                      SUM(s.executions) execution_cnt,
+                      MAX(s.sql_id) max_sql_id,
+                      COUNT(DISTINCT s.plan_hash_value) plan_cnt
+              FROM v$sql s
+              GROUP BY s.force_matching_signature HAVING COUNT(s.exact_matching_signature) >= 2 ) a
+         ORDER BY 3 DESC ) WHERE ROWNUM <= 50
+;
 </source>
-== 테이블 스페이스 ==
+또는
-=== System 테이블스페이스에 비시스템 세그먼트 조회 ===
-* DBA_SEGMENTS
 <source lang="sql">
-SELECT OWNER
+SELECT /*+ leading(h) no_merge(h) use_nl(h s) */
-     , SEGMENT_NAME
+       s.parsing_schema_name        SCHEMA                  , --> 1
-     , SEGMENT_TYPE
+       s.module                                             , --> 2
-      , TABLESPACE_NAME
+       s.sql_id                                             , --> 3
-  FROM DBA_SEGMENTS
+       s.hash_value                                         , --> 4
-  WHERE OWNER NOT IN ('SYS', 'SYSTEM')
+       SUBSTR (s.sql_text, 1, 100)       substr_sqltext     , --> 5
-    AND TABLESPACE_NAME = 'SYSTEM';
+       s.executions                                         , --> 6
+       s.buffer_gets                                        , --> 7 (전체 I/O 처리량)
+       s.disk_reads                                         , --> 8
+       s.rows_processed                                     , --> 9
+       ROUND (s.buffer_gets / s.executions, 1)                lio     , --> 10 (1회 수행 당 I/O)
+       ROUND (s.elapsed_time / s.executions / 1000000, 1)     elapsed_sec     , --> 11 (1회 수행 당 Elapsed)
+       ROUND (s.cpu_time / s.executions / 1000000, 1)         cpu_sec --> 12 (1회 수행 당 CPU)
+  FROM v$sqlarea  s
+      , (SELECT *
+          FROM (  SELECT *
+                    FROM (  SELECT s.force_matching_signature
+                                 , COUNT (s.exact_matching_signature)    literal_sql_cnt
+                                 , SUM (s.executions)                    execution_cnt
+                                 , MAX (s.sql_id)                        max_sql_id
+                                 , COUNT (DISTINCT s.plan_hash_value)    plan_cnt
+                              FROM v$sql s
+                             WHERE s.executions > 0
+                                   AND parsing_schema_name NOT IN
+                                           ('SYS', 'SYSTEM')
+                                   AND (   (    module NOT LIKE 'TOAD%'
+                                            AND module NOT LIKE 'SQL De%'
+                                            AND module NOT LIKE 'Orange%'
+                                            AND module NOT LIKE 'PL/SQL%'
+                                            AND module NOT LIKE 'plsqldev.exe')
+                                        OR (module IS NULL))
+                          GROUP BY s.force_matching_signature
+                            HAVING COUNT (s.exact_matching_signature) >= 2) a
+                ORDER BY 3 DESC)
+         WHERE ROWNUM <= 50) h
+ WHERE h.max_sql_id = s.sql_id;
 </source>
-== DB Time 분석하기 ==
+{{틀:고지상자
+|제목= SQL 수행 이력 추출하기
+|내용=* DBA_HIST_SQLTEXT : SQL TEXT가 CLOB으로 저장
+* DBA_HIST_SQL_PLAN : SQL문의 수행 PLAN이 저장되어 있으므로 플랜 변경등의 이력을 조회해 볼때 용이하다.
+* DBA_HIST_SQLBIND : SQL문을 수행한 BIND VALUE를 저장하고 있는데 시간정보가 같이 있으므로 조회패턴을 분석하기 용이하다. DBA_HIST_SQLSTAT : V$SQL 정보와 같은 SQL 수행이력이 있는 VIEW로 각 SNAP_ID 마다의 Snap Shot 정보를 담고 있으므로 수행이력 패턴을 조회할 수 있다.
+}}
-SQL> DESC v$sys_time_model 이름 널? 유형
+== 특정 SQL 수행내역 확인하기 ==
-------------------------- -------- ------------
+-- SQL TEXT
- STAT_ID NUMBER STAT_NAME VARCHAR2(64) VALUE NUMBER 1) background elapsed time 2) background cpu time 3) RMAN cpu time (backup/restore) --------------------------------------------------------------------------------
-) DB time : USER LEVEL call. 2) DB CPU 2) CONNECTION management CALL elapsed time 2) sequence LOAD elapsed time 2) SQL EXECUTE elapsed time 2) parse time elapsed 3) hard parse elapsed time 4) hard parse (sharing criteria) elapsed time 5) hard parse (bind mismatch) elapsed time 3) failed parse elapsed time 4) failed parse (OUT OF shared memory) elapsed time 2) PL/SQL execution elapsed time 2) inbound PL/SQL rpc elapsed time 2) PL/SQL compilation elapsed time 2) Java execution elapsed time 2) repeated bind elapsed time
 <source lang="sql">
-SELECT stat_name ,
+SELECT module,
-        value,
+        sql_fulltext
-       ROUND(value/
+FROM v$sqlarea
-                 ( SELECT value
+WHERE hash_value = :hash_value -- 수행내역
-                  FROM v$sys_time_model
+;
-                  WHERE stat_name='DB time' ) *100 , 2) time_ratio
-FROM v$sys_time_model
-WHERE stat_name NOT IN ( 'background elapsed time' ,
-                         'background cpu time' ,
-                         'RMAN cpu time (backup/restore)' ,
-                         'hard parse elapsed time' ,
-                         'hard parse (sharing criteria) elapsed time' ,
-                         'hard parse (bind mismatch) elapsed time' ,
-                         'failed parse elapsed time' ,
-                         'failed parse (out of shared memory) elapsed time' ,
-                         'DB CPU')
-ORDER BY time_ratio DESC ;
 </source>
-=== 총 CPU Time 대비 SQL Parsing Time ===
+* v$sqlarea
 <source lang="sql">
-SELECT max(decode(y.rno,1,decode(x.name,'CPU used by this session',x.value))) AS "CPU used by this session" ,
+    SELECT EXECUTIONS "Executions(total)",
-       max(decode(y.rno,2,decode(x.name,'parse time cpu',x.value))) AS "parse time cpu" ,
+           round(DISK_READS/executions,2) "Disk_reads(one)",
-       round((max(decode(y.rno,2,decode(x.name,'parse time cpu',x.value)))/max(decode(y.rno,1,decode(x.name,'CPU used by this session',x.value))) * 100),1)||'%' AS "ParseTime/CPU(%)"
+           round(BUFFER_GETS/executions,0) "Buffer_gets(one)",
-FROM
+           round(ROWS_PROCESSED/EXECUTIONS,0) "Rows(one)",
-    ( SELECT name,
+           round((ELAPSED_TIME/EXECUTIONS)/1000000,2) "Elapsed_time(one)",
-             value
+           round((CPU_TIME/EXECUTIONS)/1000000,2) "Cpu_time(one)"
-     FROM v$sysstat
+  FROM v$sqlarea sa
-     WHERE name IN ('CPU used by this session',
+ WHERE hash_value = :hash_value -- 바인드 변수값
-                    'parse time cpu') ) x,
+;
-    (SELECT LEVEL AS rno
-     FROM dual CONNECT BY LEVEL<=2) y ;
-=== 총 Parsing Time 대비 Hard Parsing Time ===
-<source lang="sql">
-SELECT max(decode(y.rno,1,decode(x.name,'parse time elapsed',x.value))) AS "parse time elapsed" ,
-       max(decode(y.rno,2,decode(x.name,'hard parse elapsed time',x.value))) AS "hard parse elapsed time" ,
-       round((max(decode(y.rno,2,decode(x.name,'hard parse elapsed time',x.value)))/max(decode(y.rno,1,decode(x.name,'parse time elapsed',x.value))) * 100),1)||'%' AS "Hard/ParseToral(%)"
-FROM
-    ( SELECT stat_name AS name,
-             value
-     FROM v$sys_time_model
-     WHERE stat_name IN ('parse time elapsed',
-                         'hard parse elapsed time') ) x,
-    (SELECT LEVEL AS rno
-     FROM dual CONNECT BY LEVEL<=2) y ;
 </source>
-=== Log file sync 관련 대기현상 Sync Writes Time ===
+== SQL_ID를 이용하여 SQL TEXT 추출하기 ==
-- redo synch time의 단위는 centisecond이다. => v$sysstat의 stat 중 time 정보를 담고 있는 stat의 value 단위는 centisecond이다.
+* DBA_HIST_SQLTEXT
 <source lang="sql">
-SELECT max(decode(y.rno,1,decode(x.name,'user commits',x.value))) AS "user commits" ,
+SELECT *
-       max(decode(y.rno,2,decode(x.name,'redo synch writes',x.value))) AS "redo synch writes" ,
+  FROM DBA_HIST_SQLTEXT
-       max(decode(y.rno,3,decode(x.name,
+ WHERE sql_id = :sql_id
-        'redo synch time',x.value/100))) AS "redo synch time" ,
-       round((max(decode(y.rno,3,decode(x.name,'redo synch time',x.value/100)))/max(decode(y.rno,2,decode(x.name,'redo synch writes',x.value)))),3)||'초' AS "SyncTime/SyncWrites(초)"
-FROM
-    ( SELECT name,
-             value
-     FROM v$sysstat
-     WHERE name IN ('user commits' ,
-                    'redo synch time' ,
-                    'redo synch writes') ) x,
-    (SELECT LEVEL AS rno
-     FROM dual CONNECT BY LEVEL<=3) y
 ;
 </source>
-=== 인스턴스 기동 이후 I/O 관련 대기현상 누적 값 ===
-- v$system_event의 time_waited는 centisecond이다.
+== SQL_ID를 이용하여 SQL문의 수행내역 추출하기 ==
+* dba_hist_sqlstat
 <source lang="sql">
-SELECT event ,
+    SELECT snap_id,
-       total_waits ,
+           EXECUTIONS_TOTAL "Executions(total)",
-       time_waited / 100 AS "time_waited(sec)" ,
+           round(DISK_READS_TOTAL/executions_total,2) "Disk_reads(one)",
-       average_wait
+           round(BUFFER_GETS_TOTAL/executions_total,0) "Buffer_gets(one)",
-FROM v$system_event
+           round(ROWS_PROCESSED_TOTAL/EXECUTIONS_TOTAL,0) "Rows(one)",
-WHERE event IN ('db file sequential read' ,
+           round((ELAPSED_TIME_TOTAL/EXECUTIONS_TOTAL)/1000000,2) "Elapsed_time(one)",
-                'log file sync' ,
+           round((CPU_TIME_TOTAL/EXECUTIONS_TOTAL)/1000000,2) "Cpu_time(one)"
-                'db file scattered read')
+    FROM dba_hist_sqlstat WHERE sql_id = :sql_id
+ORDER BY snap_id DESC
 ;
 </source>
+== 프로세스 ID 를 이용하여 SQL 찾기 ==
+* v$process
+* v$session
+* v$sqltext
+<source lang="sql">
+select c.sql_text
+     , b.SID
+     , b.SERIAL#
+     , b.machine
+     , b.OSUSER
+     , b.logon_time --이 쿼리를 호출한 시간
+  from v$process a, v$session b, v$sqltext c
+ where a.addr = b.paddr
+   and b.sql_hash_value = c.hash_value
+   and a.spid = '1708032' --1912870/
+ order by c.PIECE
+</source>
-== SQL 통계 정보확인하기 ==
+== 바인드 변수 값 ==
+* 바인드 변수 값의 이력을 가지고 있는 뷰 조회로 SQL문의 조회 패턴 분석 가능
-=== Module LEVEL 통계 ===
+* v$sql_bind_capture
 <source lang="sql">
-     SELECT *
+     SELECT name,
-    FROM
+           sql_id,
-        (SELECT module ,
+           to_char(LAST_CAPTURED,'yyyymmdd hh24:mi:ss') LAST_CAPTURED,
-                count(*) sql_cnt ,
+           datatype,
-                sum(executions) executions ,
+           VALUE_STRING
-                round(avg(buffer_gets/executions)) "lio(avg)" ,
+    FROM v$sql_bind_capture
-                round(avg(disk_reads/executions)) "pio(avg)" ,
+   WHERE hash_value = :hash_value -- or sql_id=:sql_id;
-                round(avg(rows_processed/executions)) "rows(avg)" ,
-                round(avg(elapsed_time/executions/1000000),2) "elapsed(avg)" ,
-                count(CASE WHEN elapsed_time/executions/1000000>=3 THEN 1 END) bad_sql ,
-                round(max(elapsed_time/executions/1000000),2) "elapsed(max)"
-         FROM v$sql
-         WHERE executions > 0
-         GROUP BY module)
-ORDER BY "lio(avg)" * executions DESC ;
 </source>
-=== SCHEMA LEVEL 통계 ===
+== 바인드 변수 값2 ==
+* gv$sql_monitor
 <source lang="sql">
-SELECT * --9i
+select BINDS_XML from
-FROM
+* gv$sql_monitor
-    ( SELECT
+ where sql_id='&1';
-         (SELECT username
+</source>
-          FROM dba_users
-          WHERE user_id = parsing_schema_id) AS SCHEMA_NAME ,
+== 바인드 변수 값3 ==
-             count(*) sql_cnt ,
+* dba_hist_sqlbind
-             sum(executions) executions ,
+<source lang="sql">
-             round(avg(buffer_gets/executions)) "lio(avg)" ,
+    SELECT name,
-             round(avg(disk_reads/executions)) "pio(avg)" ,
+           to_char(LAST_CAPTURED,'yyyymmdd hh24:mi:ss') LAST_CAPTURED,
-             round(avg(rows_processed/executions)) "rows(avg)" ,
+           datatype,
-             round(avg(elapsed_time/executions/1000000),2) "elapsed(avg)" ,
+           VALUE_STRING
-             count(CASE WHEN elapsed_time/executions/1000000>=10 THEN 1 END) bad_sql ,
+    FROM dba_hist_sqlbind
-             round(max(elapsed_time/executions/1000000),2) "elapsed(max)"
+   WHERE sql_id = :sql_id -- PLAN
-     FROM v$sqlarea
+;
-     WHERE executions > 0
-     GROUP BY parsing_schema_id)
-ORDER BY "lio(avg)" * executions DESC ;
 </source>
+== 플랜 정보 조회 ==
+* V$SQL_PLAN
 <source lang="sql">
-SELECT * --10g
+    SELECT --cardinality , cost,
-FROM
+ LPAD('.', 1*(LEVEL-1),'.')|| OPERATION || decode(OPTIONS,NULL,NULL,' (') || OPTIONS || decode(OPTIONS,NULL,NULL,')') || decode(object_owner,NULL,NULL,' :')||object_owner || decode(object_name,NULL,NULL,'.') || OBJECT_NAME || decode(other_tag,NULL,NULL,'(')||other_tag||decode(other_tag,NULL,NULL,')') || DECODE(ACCESS_PREDICATES,NULL,NULL,' [AP] '||ACCESS_PREDICATES) ||DECODE(FILTER_PREDICATES,NULL,NULL,' [FP] '||FILTER_PREDICATES) "operation"
-    (SELECT parsing_schema_name ,
+    FROM
-            count(*) sql_cnt ,
+        (SELECT /*+ NO_MERGE */ *
-            sum(executions) executions ,
+         FROM V$SQL_PLAN
-            round(avg(buffer_gets/executions)) "lio(avg)" ,
+         WHERE HASH_VALUE = :hash_value
-            round(avg(disk_reads/executions)) "pio(avg)" ,
+             AND CHILD_NUMBER = 0) CONNECT BY
-            round(avg(rows_processed/executions)) "rows(avg)" ,
+    PRIOR id = parent_id START WITH id=0
-            round(avg(elapsed_time/executions/1000000),2) "elapsed(avg)" ,
+ORDER BY ID
-            count(CASE WHEN elapsed_time/executions/1000000>=10 THEN 1 END) bad_sql ,
+;
-            round(max(elapsed_time/executions/1000000),2) "elapsed(max)"
-     FROM v$sqlarea
-     WHERE executions > 0
-     GROUP BY parsing_schema_name)
-ORDER BY "lio(avg)" * executions DESC ;
 </source>
+== SQL_ID를 이용하여 SQL문의 PLAN 추출하기 ==
-=== SQL LEVEL 통계 -- SQL 추출하기 (9i) ===
+* DBA_HIST_SQL_PLAN
 <source lang="sql">
-SELECT * --9i
+    SELECT LPAD(' ',4*(depth-1))||OPERATION||DECODE(OTHER_TAG,NULL,'','*')|| DECODE(OPTIONS,NULL,'',' ('||OPTIONS||')')|| DECODE(OBJECT_NAME,NULL,'',' OF '|| OBJECT_NAME||'')|| DECODE(OBJECT_TYPE,NULL,'',' ('||OBJECT_TYPE||')')|| DECODE(ID,0,DECODE(OPTIMIZER,NULL,'',' Optimizer='||OPTIMIZER))|| DECODE(COST,NULL,'',' (Cost='||COST|| DECODE(CARDINALITY,NULL,'',' Card='||CARDINALITY)|| DECODE(BYTES,NULL,'',' Bytes='||BYTES)||')')|| DECODE(ACCESS_PREDICATES,NULL,' ',' [AP] = '||ACCESS_PREDICATES)|| DECODE(ACCESS_PREDICATES,NULL,' ',' [FP] = '||FILTER_PREDICATES) AS PLAN_OUTPUT
-FROM
+    FROM DBA_HIST_SQL_PLAN WHERE sql_id = :sql_id
-    ( SELECT rownum cnt,
+ORDER BY ID,
-             t1.*
+         POSITION
-     FROM
-         ( SELECT
-              (SELECT username
-               FROM dba_users
-               WHERE user_id=parsing_schema_id) AS SCHEMA,
-                  module,
-                  hash_value,
-                  substr(sql_text,1,100) substr_sqltext,
-                  executions,
-                  buffer_gets,
-                  disk_reads,
-                  rows_processed,
-                  round(buffer_gets/executions,1) lio,
-                  round(elapsed_time/executions/1000000,1) elapsed_sec,
-                  round(cpu_time/executions/1000000,1) cpu_sec,
-                  round(elapsed_time/tot_elapsed*100,1) ratio_elapsed,
-                  round(cpu_time/tot_cpu*100,1) ratio_cpu
-          FROM v$sqlarea s,
-              (SELECT sum(elapsed_time) tot_elapsed,
-                      sum(cpu_time) tot_cpu
-               FROM v$sqlarea) t
-          WHERE s.executions > 0
-          ORDER BY 13 DESC ) t1
-     WHERE rownum <= 50)
-WHERE cnt >= 1
 ;
 </source>
-<source lang="sql">
+----
-    SELECT * --9i
-      FROM
+= SQL 과 커서 조회 =
-        ( SELECT rownum cnt,
-                 t1.*
+== 현재 커서 수 확인 ==
-         FROM
+* V$OPEN_CURSOR
-             ( SELECT
+* v$session_wait
-                  (SELECT username
+* v$transaction
-                   FROM dba_users
-                   WHERE user_id=parsing_schema_id) AS SCHEMA,
-                      module,
-                      hash_value,
-                      substr(sql_text,1,100) substr_sqltext,
-                      executions,
-                      buffer_gets,
-                      disk_reads,
-                      rows_processed,
-                      round(buffer_gets/executions,1) lio,
-                      round(elapsed_time/executions/1000000,1) elapsed_sec,
-                      round(cpu_time/executions/1000000,1) cpu_sec,
-                      round(elapsed_time/tot_elapsed*100,1) ratio_elapsed,
-                      round(cpu_time/tot_cpu*100,1) ratio_cpu
-              FROM v$sqlarea s,
-                  (SELECT sum(elapsed_time) tot_elapsed,
-                          sum(cpu_time) tot_cpu
-                   FROM v$sqlarea) t
-              WHERE s.executions > 0
-                  AND (s.sql_text NOT LIKE 'BEGIN'||'%'
-                       AND s.sql_text NOT LIKE 'begin'||'%'
-                       AND s.sql_text NOT LIKE 'DECLARE'||'%'
-                       AND s.sql_text NOT LIKE 'declare'||'%' )
-              ORDER BY 13 DESC ) t1
-         WHERE rownum <= 50) WHERE cnt >= 1
-;
-</source>
-=== SQL 추출하기 (10g) ===
+== sid별 열린커셔 ==
 <source lang="sql">
-    SELECT *
+SELECT sid, count(sid) cursor
-    FROM
+  FROM V$OPEN_CURSOR
-        ( SELECT rownum cnt,
+ WHERE user_name = 'dbcafe'
-                 t1.*
+ GROUP BY sid
-         FROM
+ ORDER BY cursor DESC;
-             ( SELECT parsing_schema_name SCHEMA,    --> 1
+</source>
-                                          module,    --> 2
-                                          sql_id,    --> 3
-                                          hash_value,--> 4
-                                          substr(sql_text,1,100) substr_sqltext, --> 5
-                                          executions,  --> 6
-                                          buffer_gets, --> 7 (전체 I/O 처리량)
-                                          disk_reads,  --> 8
-                                          rows_processed, --> 9
-                                          round(buffer_gets/executions,1) lio, --> 10 (1회 수행 당 I/O)
-                                          round(elapsed_time/executions/1000000,1) elapsed_sec, --> 11 (1회 수행 당 Elapsed)
-                                          round(cpu_time/executions/1000000,1) cpu_sec, --> 12 (1회 수행 당 CPU)
-                                          round(elapsed_time/tot_elapsed*100,1) ratio_elapsed, --> 13 (전체 대비 Elapsed )
-                                          round(cpu_time/tot_cpu*100,1) ratio_cpu --> 14 (전체 대비 CPU 사용률)
-                            FROM v$sqlarea s,
-                  (SELECT sum(elapsed_time) tot_elapsed,
-                          sum(cpu_time) tot_cpu
-                   FROM v$sqlarea) t
-              WHERE s.executions > 0
-                  AND parsing_schema_name NOT IN ('SYS','SYSTEM')
-                  AND ((module NOT LIKE 'TOAD%'
-                        AND module NOT LIKE 'SQL De%'
-                        AND module NOT LIKE 'Orange%'
-                        AND module NOT LIKE 'PL/SQL%'
-                        AND module NOT LIKE 'plsqldev.exe')
-                       OR (module IS NULL))
-              ORDER BY 14 DESC ) t1
-         WHERE rownum <= 50) WHERE cnt >= 1
-;
-</source>
+== sql 별 열린 커서 ==
+<source lang="sql">
+SELECT sql_text, count(sid) cnt
+  FROM v$OPEN_CURSOR
+ GROUP BY sql_text
+ ORDER BY cnt DESC
+</source>
-=== Exa SQL[1]. Smart Scan으로 수행되지 않고,Interconnect Bytes가 높은 SQL ===
+----
+= 락 모니터링 =
-<source lang="sql">
+== V$LOCK 을 사용한 잠금 경합 모니터링 ==
-WITH Cell_Mon_SQL AS
+* V$SESSION
-        ( SELECT sql_id,
+* DBA_OBJECTS
-                 io_cell_offload_eligible_bytes AS coeb,
+* V$LOCK
-                 io_interconnect_bytes AS ib,
+<source lang="sql">
-                 io_cell_offload_returned_bytes AS corb,
+SELECT S.USERNAME, S.SID, S.SERIAL#, S.LOGON_TIME,
-                 io_cell_uncompressed_bytes AS cub,
+    DECODE(L.TYPE, 'TM', 'TABLE LOCK',
-                 optimized_phy_read_requests AS oprr,
+                   'TX', 'ROW LOCK',
-                 physical_read_requests AS prr,
+             NULL) "LOCK LEVEL",
-                 physical_read_bytes AS prb,
+    O.OWNER, O.OBJECT_NAME, O.OBJECT_TYPE
-                 sql_text
+FROM V$SESSION S, V$LOCK L, DBA_OBJECTS O
-         FROM v$sqlarea)
+WHERE S.SID = L.SID
-    SELECT /*** Smart Scan이 되지 않으면서, Interconnect Bytes가 높은 SQL ***/ *
+AND O.OBJECT_ID = L.ID1
-    FROM
+AND S.USERNAME IS NOT NULL
-        ( SELECT sql_id ,
+</source>
-                 decode(coeb,0,'No','Yes') AS "Offload",
-                 round(prb/1024/1024/1024,3) AS "Physical_Bytes(GB)",
+== 락이 걸린 세션 간단히 보기 ==
-                 round(coeb/1024/1024/1024,3) AS "Offload_Bytes(GB)",
+<source lang="sql">
-                 round(ib/1024/1024/1024,3) AS "Interconnect_Bytes(GB)",
+SELECT A.SID, A.SERIAL#, B.TYPE, C.OBJECT_NAME, A.PROGRAM, A.LOCKWAIT,
-                 round(oprr*8192/1024/1024/1024,3) AS "Cell_Flash_Cache(GB)",
+       A.LOGON_TIME, A.PROCESS, A.OSUSER, A.TERMINAL
-                 round(corb/1024/1024/1024,3) AS "Returned_Bytes(GB)",
+  FROM V$SESSION   A
-                 round(decode(coeb,0,0,100*(coeb-ib)/coeb),2) AS "Saved_IO(%)",
+     , V$LOCK      B
-                 sql_text
+     , DBA_OBJECTS C
-         FROM Cell_Mon_SQL
+ WHERE A.SID = B.SID
-         WHERE coeb = 0
+   AND B.ID1 = C.OBJECT_ID
-         ORDER BY ib DESC) WHERE rownum <= 10 ;
+   AND B.TYPE = 'TM';
+</source>
-  SQL[2]. Smart Scan으로 수행되지 않고,Offload Returned Bytes가 높은 SQL :
-Reverting TO Block Shipping으로 수행된 SQL로 Smart Scan으로 동작할 수 있는 SQL이지만,                                                                                                           Direct I/O로 동작하지 않는 경우 (FULL TABLE Scan도 cell single block physical read로 수행됨.)
+== 락이 걸린 세션 상세 보기 ==
-- Chained Row가 있는 블록들을 읽을 때 발생. : Smart Scan을 통해서 Chained Row가 있는 블록을 읽을 때 각 블록이 다른
+* V$SESSION
-STORAGE Cell에 있는 경우 STORAGE Cell 간 통신할 수 없기 때문에 Buffer Cache로 전체 블록을 전송한 후 처리하도록 Block Shipping Mode로 변경
+* DBA_OBJECTS
-- 읽기 일관성 모드로 변경 중인 데이터를 가진 Block에 대해서 Smart Scan을 중지하고,                                                                                                                                Single Block Read를 수행하는 Block Shipping Mode로 변경
+* V$LOCK
+<source lang="sql">
+SELECT A.SID, A.SERIAL#,A.USERNAME,A.PROCESS,B.OBJECT_NAME
+     , DECODE(C.LMODE,2,'RS',3,'RX',4,'S',5,'SRX',8,'X','NO') "TABLE LOCK"
+     , DECODE (A.COMMAND,2,'INSERT',3,'SELECT',6,'UPDATE',7,'DELETE',12,'DROP TABLE',26,'LOCK TABLE','UNKNOWN') "SQL"
+     , DECODE(A.LOCKWAIT, NULL,'NO WAIT','WAIT') "STATUS"
+  FROM V$SESSION A
+     , DBA_OBJECTS B
+     , V$LOCK C
+ WHERE A.SID=C.SID
+   AND B.OBJECT_ID=C.ID1
+   AND C.TYPE='TM'
+</source>
+* 락이 걸린 세션을 찾아 내어 세션을 죽이려고 해도 죽지 않는 경우
+* 아래 쿼리문으로 OS단의 PROCESS ID를 찾아내어 OS에서 죽인다
+* kill -9 프로세스아이디
+<source lang="sql">
+SELECT SUBSTR(S.USERNAME,1,11) "ORACLE USER", P.PID "PROCESS ID"
+     , S.SID "SESSION ID", S.SERIAL#, OSUSER "OS USER"
+     , P.SPID "PROC SPID",S.PROCESS "SESS SPID", S.LOCKWAIT "LOCK WAIT"
+  FROM V$PROCESS P
+     , V$SESSION S
+     , V$ACCESS  A
+ WHERE A.SID=S.SID
+   AND P.ADDR=S.PADDR
+   AND S.USERNAME != 'SYS'
+</source>
+* 위 쿼리문의 결과가 있다면 락이 걸린 세션이 있다는것이므로 아래의 쿼리문으로 세션을 죽인다
+<source lang="sql">
+ALTER SYSTEM KILL SESSION '11,39061'
+</source>
+* alter session으로 죽지않는 프로세스 죽이기
+<source lang="sql">
+.oracle이 설치된 서버에 텔넷으로 root로 접속한다
+.su -오라클계정
+.sqlplus '/as sysdba''
+.connect system/sys
+.ALTER SYSTEM KILL SESSION '137,1723'
+</source>
-<source lang="sql">
-WITH Cell_Mon_SQL AS
-    ( SELECT sql_id,
-             io_cell_offload_eligible_bytes AS coeb,
-             io_interconnect_bytes AS ib,
-             io_cell_offload_returned_bytes AS corb,
-             io_cell_uncompressed_bytes AS cub,
-             optimized_phy_read_requests AS oprr,
-             physical_read_requests AS prr,
-             physical_read_bytes AS prb,
-             sql_text
-     FROM v$sqlarea)
-SELECT *
-FROM
-    ( SELECT sql_id ,
-             decode(coeb,0,'No','Yes') AS "Offload",
-             round(prb/1024/1024/1024,3) AS "Physical_Bytes(GB)",
-             round(coeb/1024/1024/1024,3) AS "Offload_Bytes(GB)",
-             round(ib/1024/1024/1024,3) AS "Interconnect_Bytes(GB)",
-             round(oprr*8192/1024/1024/1024,3) AS "Cell_Flash_Cache(GB)",
-             round(corb/1024/1024/1024,3) AS "Returned_Bytes(GB)",
-             round(decode(coeb,0,0,100*(coeb-ib)/coeb),2) AS "Saved_IO(%)",
-             sql_text
-     FROM Cell_Mon_SQL
-     WHERE coeb = 0
-     ORDER BY corb DESC)
-WHERE rownum <= 10 ;
-</source>
+== 1일 이상 존재 하는 조회 / kill / AWS_KILL_CMD ==
+<source lang=sql>
+-- DROP VIEW RTIS_DBA.V_DBA_KILL_SESS_OVER_1DAY;
-=== Full Table Scan 처리가 있는 SQL문 정보 추출하기(10g) ===
+CREATE OR REPLACE FORCE VIEW V_DBA_KILL_SESS_OVER_1DAY
-<source lang="sql">
+(
-SELECT *
+    STATUS
-FROM
+  , USERNAME
-    (SELECT rownum cnt,
+  , OSUSER
-            t1.*
+  , MACHINE
-     FROM
+  , EXEC_TIME
-         (SELECT parsing_schema_name,
+  , SID
-                 --> 1
+  , SERIAL#
+  , CMD_KILL
+)
+BEQUEATH DEFINER
+AS
+      SELECT STATUS
+           , USERNAME
+           , OSUSER
+           , MACHINE
+           , NVL (A.SQL_EXEC_START, A.PREV_EXEC_START)    EXEC_TIME
+           , SID
+           , SERIAL#
+           ,    'EXEC RDSADMIN.RDSADMIN_UTIL.KILL('
+             || TO_CHAR (SID)
+             || ','
+             || TO_CHAR (SERIAL#)
+             || ');'                                      CMD_KILL
+        FROM V$SESSION A
+       WHERE     STATUS = 'INACTIVE'
+             AND TYPE = 'USER'
+             AND MACHINE NOT LIKE 'ip%'
+             AND (SYSDATE - NVL (A.SQL_EXEC_START, A.PREV_EXEC_START)) * 24 > 1
+    ORDER BY EXEC_TIME DESC;
+</source>
-                 module,
-                 --> 2
-                 sql_id,
-                 --> 3
-                 hash_value,
+----
-                 --> 4
-                 substr(sql_text,1,100) substr_sqltext,
+= 프로세스 정보 =
-                 --> 5
+== Oracle Process의 정보 ==
+* V$SESSION
+* V$PROCESS
+* SYS.V_$SESS_IO
-                 executions,
+<source lang="sql">
-                 --> 6
+SELECT S.STATUS "STATUS"
+     , S.SERIAL# "SERIAL#"
-                 buffer_gets,
+     , S.TYPE "TYPE"
-                 --> 7 (전체 I/O 처리량)
+     , S.USERNAME "DB USER"
+     , S.OSUSER "CLIENT USER"
-                 disk_reads,
+     , S.SERVER "SERVER"
-                 --> 8
+     , S.MACHINE "MACHINE"
+     , S.MODULE "MODULE"
-                 rows_processed,
+     , S.TERMINAL "TERMINAL"
-                 --> 9
+     , S.PROGRAM "PROGRAM"
+     , P.PROGRAM "O.S. PROGRAM"
-                 round(buffer_gets/executions,1) lio,
+     , S.LOGON_TIME "CONNECT TIME"
-                 --> 10 (1회 수행 당 I/O)
+     , LOCKWAIT "LOCK WAIT"
+     , SI.PHYSICAL_READS "PHYSICAL READS"
-                 round(elapsed_time/executions/1000000,1) elapsed_sec,
+     , SI.BLOCK_GETS "BLOCK GETS"
-                 --> 11 (1회 수행 당 Elapsed)
+     , SI.CONSISTENT_GETS "CONSISTENT GETS"
+     , SI.BLOCK_CHANGES "BLOCK CHANGES"
-                 round(cpu_time/executions/1000000,1) cpu_sec,
+     , SI.CONSISTENT_CHANGES "CONSISTENT CHANGES"
-                 --> 12 (1회 수행 당 CPU)
+     , S.PROCESS "PROCESS"
+     , P.SPID
-                 round(elapsed_time/tot_elapsed*100,1) ratio_elapsed,
+     , P.PID
-                 --> 13 (전체 대비 Elapsed )
+     , S.SERIAL#
+     , SI.SID
+     , S.SQL_ADDRESS "ADDRESS"
+     , S.SQL_HASH_VALUE "SQL HASH"
+     , S.ACTION
+  FROM V$SESSION S
+     , V$PROCESS P
+     , SYS.V_$SESS_IO SI
+ WHERE S.PADDR = P.ADDR(+)
+   AND SI.SID(+) = S.SID
+   AND S.USERNAME IS NOT NULL
+   AND NVL(S.OSUSER, 'X') <> 'SYSTEM'
+   AND S.TYPE <> 'BACKGROUND'
+ ORDER BY 3;
+</source>
-                 round(cpu_time/tot_cpu*100,1) ratio_cpu --> 14 (전체 대비 CPU 사용률)
+== 오브젝트에 접속되어 있는 프로그램 조회   ==
+* V$SESSION
-          FROM
+* V$ACCESS
-              (SELECT /*+ leading(x) no_merge(x) use_hash(x s) */ s.*
+<source lang="sql">
-               FROM v$sqlarea s,
+SELECT SUBSTR(B.OBJECT, 1, 15) AS OBJECT, SUBSTR(A.PROGRAM, 1, 15) AS PROGRAM, COUNT(*) AS CNT
-                   ( SELECT DISTINCT hash_value
+  FROM V$SESSION A
-                    FROM v$sql_plan
+     , V$ACCESS B
-                    WHERE OPERATION = 'TABLE ACCESS'
+  WHERE A.SID = B.SID
-                        AND OPTIONS = 'FULL' --and object_owner like :b1||'%'
+   AND B.OWNER NOT IN ('SYS')
-  ) x
+   AND A.TYPE!= 'BACKGROUND'
-               WHERE x.hash_value = s.hash_value) s,
+   AND B.OBJECT LIKE UPPER('&OBJECT_NAME') || '%'
-              (SELECT sum(elapsed_time) tot_elapsed,
+ GROUP BY B.OBJECT, SUBSTR(A.PROGRAM, 1, 15);
-                      sum(cpu_time) tot_cpu
+</source>
-               FROM v$sqlarea) t
-          WHERE executions > 0
+----
-              AND parsing_schema_name NOT IN ('SYS',
-                                              'SYSTEM')
-              AND ((module NOT LIKE 'TOAD%'
-                    AND module NOT LIKE 'SQL De%'
-                    AND module NOT LIKE 'Orange%'
-                    AND module NOT LIKE 'PL/SQL%'
-                    AND module NOT LIKE 'plsqldev.exe')
-                   OR (module IS NULL))
-          ORDER BY 10 DESC ) t1
-     WHERE rownum <= 50)
-WHERE cnt >= 1
-=== 하드파싱 찾기 ===
+= 부하 발생 모니터링 =
+== cpu를 많이 사용하는 쿼리문과 프로세스아이디,시리얼번호,머신 알아내기 ==
+* V$PROCESS
+* V$SESSION
+* V$SQLTEXT
 <source lang="sql">
-    SELECT *
+SELECT C.SQL_TEXT
-    FROM
+     , B.SID
-        ( SELECT
+     , B.SERIAL#
-             ( SELECT parsing_schema_name
+     , B.MACHINE
-              FROM v$sqlarea
+     , B.OSUSER
-              WHERE sql_id=a.max_sql_id ) SCHEMA,
+     , B.LOGON_TIME --이 쿼리를 호출한 시간
-             ( SELECT MODULE
+  FROM V$PROCESS A
-              FROM v$sqlarea
+     , V$SESSION B
-              WHERE sql_id=a.max_sql_id ) MODULE, a.literal_sql_cnt, a.execution_cnt, a.plan_cnt, a.max_sql_id,
+     , V$SQLTEXT C
-             ( SELECT sql_fulltext
+ WHERE A.ADDR = B.PADDR
-              FROM v$sqlarea
+   AND B.SQL_HASH_VALUE = C.HASH_VALUE
-              WHERE sql_id = a.max_sql_id ) sql_text
+--AND A.SPID = '675958'
-         FROM
+ ORDER BY C.PIECE
-             ( SELECT s.force_matching_signature,
+</source>
-                      COUNT(s.exact_matching_signature) literal_sql_cnt,
-                      SUM(s.executions) execution_cnt,
+== 롤백 세그먼트 경합 조회  ==
-                      MAX(s.sql_id) max_sql_id,
+* V$ROLLSTAT
-                      COUNT(DISTINCT s.plan_hash_value) plan_cnt
+* V$ROLLNAME
-              FROM v$sql s
-              GROUP BY s.force_matching_signature HAVING COUNT(s.exact_matching_signature) >= 2 ) a
-         ORDER BY 3 DESC ) WHERE ROWNUM <= 50
-    SELECT /*+ leading(h) no_merge(h) use_nl(h s) */ s.parsing_schema_name SCHEMA,
-                                                                           --> 1
-                                                                           s.module,
+<source lang="sql">
-                                                                           --> 2
+SELECT NAME T0
+     , GETS T1
-                                                                           s.sql_id,
+     , WAITS T2
-                                                                           --> 3
+     , TO_CHAR(TRUNC(WAITS / GETS * 100, 2), 099.99) || '%' T3
+     , TO_CHAR(ROUND(RSSIZE / 1024)) T4
-                                                                           s.hash_value,
+     , SHRINKS T5
-                                                                           --> 4
+     , EXTENDS T6
+  FROM V$ROLLSTAT
+     , V$ROLLNAME
+ WHERE V$ROLLSTAT.USN = V$ROLLNAME.USN;
+</source>
-                                                                           s.substr(s.sql_text,1,100) substr_sqltext,
+== 버퍼 캐시 히트율(Buffer Cache Hit Ratio) ==
-                                                                           --> 5
+* V$SYSSTAT
+<source lang="sql">
+SELECT ROUND(((1-(SUM(DECODE(name, 'physical reads', value,0))/(SUM(DECODE(name, 'db block gets', value,0))+ (SUM(DECODE(name, 'consistent gets', value, 0))))))*100),2) || '%' "Buffer Cache Hit Ratio"
+  FROM V$SYSSTAT;
+</source>
-                                                                           s.executions,
+== 라이브러리 캐시 히트율(Library Cache Hit Ratio) ==
-                                                                           --> 6
+* V$LIBRARYCACHE
+<source lang="sql">
+SELECT (1-SUM (reloads)/SUM(pins))*100 "Library Cache Hit Ratio"
+From V$LIBRARYCACHE;
+</source>
-                                                                           s.buffer_gets,
+== 데이터 딕셔너리 캐시 히트율(Data Dictionary Cache Hit Ratio) ==
-                                                                           --> 7 (전체 I/O 처리량)
+* V$ROWCACHE
+<source lang="sql">
+SELECT (1-SUM(getmisses)/SUM(gets))*100 "Data Dictionary Hit Ratio"
+  FROM V$ROWCACHE;
+</source>
+----
-                                                                           s.disk_reads,
+= I/O 부하 찾기 =
-                                                                           --> 8
+* V$IOFUNCMETRIC
+<source lang=sql>
+-- iometric values in 11g
-                                                                           s.rows_processed,
+select min(begin_time) b_time, min(end_time) e_time, function_name,
-                                                                           --> 9
+       round(sum(small_read_iops+large_read_iops)) read_TOT_iops, round(sum(small_write_iops+large_write_iops)) write_TOT_iops, round(sum(large_read_mbps+small_read_mbps)) read_TOT_mbps, round(sum(large_write_mbps+small_write_mbps)) write_TOT_mbps
+       from GV$IOFUNCMETRIC
+       group by rollup(function_name)
+       having round(sum(small_read_iops+large_read_iops)) + round(sum(large_read_mbps+small_read_mbps)) + round(sum(small_write_iops+large_write_iops)) + round(sum(large_write_mbps+small_write_mbps))  >0
+       order by function_name;
+</source>
-                                                                           round(s.buffer_gets/s.executions,1) lio,
+<source lang=sql>
-                                                                           --> 10 (1회 수행 당 I/O)
+-- query to iofuncmetric view in 11g
-                                                                           round(s.elapsed_time/s.executions/1000000,1) elapsed_sec,
+select inst_id,begin_time,function_name,
-                                                                           --> 11 (1회 수행 당 Elapsed)
+       round(small_read_iops) RD_IOPS_sm, round(large_read_iops) RD_IOPS_lg,
+       round(small_read_mbps) RD_MBPS_sm, round(large_read_mbps) RD_MBPS_lg,
+       round(small_write_iops) WT_IOPS_sm, round(large_write_iops) WT_IOPS_lg,
+       round(small_write_mbps) WT_MBPS_sm, round(large_write_mbps) WT_MBPS_lg
+from GV$IOFUNCMETRIC
+--where function_name in ('Buffer Cache Reads','LGWR','DBWR','Direct Reads','Direct Writes','RMAN')
+--where round(small_read_iops+large_read_iops+small_write_iops+large_write_iops) >0
+order by function_name,inst_id;
+</source>
+<source lang=sql>
+select min(begin_time) b_time, min(end_time) e_time, round(sum(small_read_iops+large_read_iops)) read_TOT_iops, round(sum(large_read_mbps+small_read_mbps)) read_TOT_mbps, round(sum(small_write_iops+large_write_iops)) write_TOT_iops, round(sum(large_write_mbps+small_write_mbps)) write_TOT_mbps from GV$IOFUNCMETRIC;
+</source>
-                                                                           round(s.cpu_time/s.executions/1000000,1) cpu_sec --> 12 (1회 수행 당 CPU)
+== 인스턴스 기동 이후 I/O 관련 대기현상 누적 값 ==
+- v$system_event의 time_waited는 centisecond (1/100 초)
-    FROM v$sqlarea s,
+<source lang="sql">
-        (SELECT *
+SELECT EVENT ,
-         FROM
+       TOTAL_WAITS ,
-             (SELECT *
+       TIME_WAITED / 100 AS "TIME_WAITED(SEC)" ,
-              FROM
+       AVERAGE_WAIT
-                  (SELECT s.force_matching_signature,
+FROM V$SYSTEM_EVENT
-                          count(s.exact_matching_signature) literal_sql_cnt,
+WHERE EVENT IN ('DB FILE SEQUENTIAL READ' ,
-                          sum(s.executions) execution_cnt,
+                'LOG FILE SYNC' ,
-                          max(s.sql_id) max_sql_id,
+                'DB FILE SCATTERED READ')
-                          count(DISTINCT s.plan_hash_value) plan_cnt
+;
-                   FROM v$sql s
+</source>
-                   WHERE s.executions > 0
-                       AND parsing_schema_name NOT IN ('SYS',
-                                                       'SYSTEM')
-                       AND ((module NOT LIKE 'TOAD%'
-                             AND module NOT LIKE 'SQL De%'
-                             AND module NOT LIKE 'Orange%'
-                             AND module NOT LIKE 'PL/SQL%'
-                             AND module NOT LIKE 'plsqldev.exe')
-                            OR (module IS NULL))
-                   GROUP BY s.force_matching_signature HAVING count(s.exact_matching_signature) >= 2) a
-              ORDER BY 3 DESC)
-         WHERE rownum <= 50) h WHERE h.max_sql_id = s.sql_id ;
-</source>
+== 전체 IO량(100%) 대비 사용자별 IO량  ==
-=== 특정 SQL 수행내역 확인하기 ===
--- SQL TEXT
 <source lang="sql">
-SELECT module,
+WITH io AS (
-        sql_fulltext
+  SELECT /*+ materialize */ decode(statistic_name,'logical reads', 'lio','pio') io_type
-FROM v$sqlarea
+        , sum(value) io_value
-WHERE hash_value = :hash_value -- 수행내역
+  FROM v$segment_statistics
-;
+ WHERE statistic_name IN ( 'logical reads' ,
-</source>
+                           'physical reads' ,
+                           'physical reads direct' )
-<source lang="sql">
+ GROUP BY decode(statistic_name,'logical reads', 'lio','pio')),
-    SELECT EXECUTIONS "Executions(total)",
+per_io AS
-           round(DISK_READS/executions,2) "Disk_reads(one)",
+( SELECT OWNER,
-           round(BUFFER_GETS/executions,0) "Buffer_gets(one)",
+         object_name,
-           round(ROWS_PROCESSED/EXECUTIONS,0) "Rows(one)",
+         object_type,
-           round((ELAPSED_TIME/EXECUTIONS)/1000000,2) "Elapsed_time(one)",
+         seg_io,
-           round((CPU_TIME/EXECUTIONS)/1000000,2) "Cpu_time(one)"
+         seg_value,
-  FROM v$sqlarea sa
+         round(seg_value / decode(seg_io,'lio',
- WHERE hash_value = :hash_value -- 바인드 변수값
+                                      (SELECT io_value
-;
+                                       FROM io
-;
+                                       WHERE io_type='lio'),
-</source>
+                                      (SELECT io_value
+                                       FROM io
-<source lang="sql">
+                                       WHERE io_type='pio') )*100,5) per_seg_io
-    SELECT name,
+ FROM
-           sql_id,
+     ( SELECT ss.OWNER,
-           to_char(LAST_CAPTURED,'yyyymmdd hh24:mi:ss') LAST_CAPTURED,
+                 ss.object_name,
-           datatype,
+                 ss.object_type,
-           VALUE_STRING
+                 decode(ss.statistic_name,'logical reads', 'lio','pio') seg_io,
-    FROM v$sql_bind_capture WHERE hash_value = :hash_value
+                                                                        sum(ss.value) seg_value
--- 바인드 변수 값의 이력을 가지고 있는 뷰 조회로 SQL문의 조회 패턴을 분석할 수 있다.
+      FROM v$segment_statistics ss
-;
+      WHERE ss.statistic_name IN ( 'logical reads' ,
-</source>
+                                   'physical reads' ,
+                                   'physical reads direct' )
-<source lang="sql">
+      GROUP BY ss.OWNER,
-    SELECT name,
+                  ss.object_name,
-           to_char(LAST_CAPTURED,'yyyymmdd hh24:mi:ss') LAST_CAPTURED,
+                  ss.object_type,
-           datatype,
+                  decode(ss.statistic_name,'logical reads', 'lio','pio') ))
-           VALUE_STRING
+SELECT *
-    FROM dba_hist_sqlbind WHERE sql_id = :sql_id -- PLAN
+FROM
-;
+( SELECT OWNER,
-;
+         sum(decode(seg_io,'lio',per_seg_io)) AS lio,
-</source>
+         sum(decode(seg_io,'pio',per_seg_io)) AS pio
+ FROM per_io
-<source lang="sql">
+ GROUP BY OWNER
-    SELECT --cardinality , cost,
+ ORDER BY sum(decode(seg_io,'lio',per_seg_io)) DESC)
- LPAD('.', 1*(LEVEL-1),'.')|| OPERATION || decode(OPTIONS,NULL,NULL,' (') || OPTIONS || decode(OPTIONS,NULL,NULL,')') || decode(object_owner,NULL,NULL,' :')||object_owner || decode(object_name,NULL,NULL,'.') || OBJECT_NAME || decode(other_tag,NULL,NULL,'(')||other_tag||decode(other_tag,NULL,NULL,')') || DECODE(ACCESS_PREDICATES,NULL,NULL,' [AP] '||ACCESS_PREDICATES) ||DECODE(FILTER_PREDICATES,NULL,NULL,' [FP] '||FILTER_PREDICATES) "operation"
+WHERE rownum <= 50 - Segment별 전체 IO량(100%) 대비 IO량 (TABLE/INDEX 분리) WITH io AS
-    FROM
+( SELECT /*+ materialize */ decode(statistic_name,'logical reads', 'lio','pio') io_type,
-        (SELECT /*+ NO_MERGE */ *
+                                                                                sum(value) io_value
-         FROM V$SQL_PLAN
+ FROM v$segment_statistics
-         WHERE HASH_VALUE = :hash_value
+ WHERE statistic_name IN ( 'logical reads' ,
-             AND CHILD_NUMBER = 0) CONNECT BY
+                           'physical reads' ,
-    PRIOR id = parent_id START WITH id=0
+                           'physical reads direct' )
-ORDER BY ID
+ GROUP BY decode(statistic_name,'logical reads', 'lio','pio')),
-;
+                                                                        per_io AS
-</source>
+( SELECT OWNER,
+         object_name,
---SQL 수행 이력 추출하기
+         object_type,
-DBA_HIST_SQLTEXT : SQL TEXT가 CLOB으로 저장
+         seg_io,
-DBA_HIST_SQL_PLAN : SQL문의 수행 PLAN이 저장되어 있으므로 플랜 변경등의 이력을 조회해 볼때 용이하다.
+         seg_value,
-DBA_HIST_SQLBIND : SQL문을 수행한 BIND VALUE를 저장하고 있는데 시간정보가 같이 있으므로 조회패턴을 분석하기 용이하다. DBA_HIST_SQLSTAT : V$SQL 정보와 같은 SQL 수행이력이 있는 VIEW로 각 SNAP_ID 마다의 Snap Shot 정보를 담고 있으므로 수행이력 패턴을 조회할 수 있다.
+         round(seg_value / decode(seg_io,'lio',
+                                      (SELECT io_value
-=== SQL_ID를 이용하여 SQL TEXT 추출하기 ===
+                                       FROM io
+                                       WHERE io_type='lio'),
-<source lang="sql">
+                                      (SELECT io_value
-SELECT *
+                                       FROM io
-   FROM DBA_HIST_SQLTEXT
+                                       WHERE io_type='pio') )*100,5) per_seg_io
- WHERE sql_id = :sql_id
+ FROM
-;
+     ( SELECT ss.OWNER,
+                 ss.object_name,
+                 ss.object_type,
+                 decode(ss.statistic_name,'logical reads', 'lio','pio') seg_io,
+                                                                        sum(ss.value) seg_value
+      FROM v$segment_statistics ss
+      WHERE ss.statistic_name IN ( 'logical reads' ,
+                                   'physical reads' ,
+                                   'physical reads direct' )
+      GROUP BY ss.OWNER,
+                  ss.object_name,
+                  ss.object_type,
+                  decode(ss.statistic_name,'logical reads', 'lio','pio') ))
+SELECT *
+   FROM per_io
+ORDER BY per_seg_io DESC
+;
 </source>
-=== SQL_ID를 이용하여 SQL문의 수행내역 추출하기 ===
-<source lang="sql">
-    SELECT snap_id,
-           EXECUTIONS_TOTAL "Executions(total)",
-           round(DISK_READS_TOTAL/executions_total,2) "Disk_reads(one)",
-           round(BUFFER_GETS_TOTAL/executions_total,0) "Buffer_gets(one)",
-           round(ROWS_PROCESSED_TOTAL/EXECUTIONS_TOTAL,0) "Rows(one)",
-           round((ELAPSED_TIME_TOTAL/EXECUTIONS_TOTAL)/1000000,2) "Elapsed_time(one)",
-           round((CPU_TIME_TOTAL/EXECUTIONS_TOTAL)/1000000,2) "Cpu_time(one)"
-    FROM dba_hist_sqlstat WHERE sql_id = :sql_id
-ORDER BY snap_id DESC
-;
-</source>
-=== SQL_ID를 이용하여 바인드 변수 값을 추출하기 ===
+= 테이블스페이스 =
+== 테이블스페이스 사용량 ==
+* dba_data_files
+* dba_free_space
+<source lang="sql">
+SELECT A.TABLESPACE_NAME,
+             A.TOTAL "TOTAL(MB)",
+             A.TOTAL - B.FREE "USED(MB)",
+             NVL(B.FREE,0) "FREE(MB)",
+             ROUND((A.TOTAL - NVL(B.FREE,0))*100/TOTAL,0)  "USED(%)"
+  FROM ( SELECT TABLESPACE_NAME
+                , ROUND((SUM(BYTES)/1024/1024),0) AS TOTAL
+             FROM DBA_DATA_FILES
+            GROUP BY TABLESPACE_NAME
+       ) A
+     , ( SELECT TABLESPACE_NAME
+              , ROUND((SUM(BYTES)/1024/1024),0) AS FREE
+           FROM DBA_FREE_SPACE
+          GROUP BY  TABLESPACE_NAME
+       ) B
+ WHERE A.TABLESPACE_NAME = B.TABLESPACE_NAME(+)
+ ORDER BY A.TABLESPACE_NAME;
+</source>
+== System 테이블스페이스에 비시스템 세그먼트 조회 ==
+* DBA_SEGMENTS
 <source lang="sql">
-SELECT name,
+SELECT OWNER
-       to_char(LAST_CAPTURED,'yyyymmdd hh24:mi:ss') LAST_CAPTURED,
+     , SEGMENT_NAME
-       datatype,
+     , SEGMENT_TYPE
-       VALUE_STRING
+     , TABLESPACE_NAME
-FROM dba_hist_sqlbind
+  FROM DBA_SEGMENTS
-WHERE sql_id = :sql_id
+ WHERE OWNER NOT IN ('SYS', 'SYSTEM')
--- dba_hist_sqlbind를 이용하면 바인드 변수 값의 이력을 가지고 있는 뷰이므로 SQL문의 조회 패턴을 분석할 수 있다.
+   AND TABLESPACE_NAME = 'SYSTEM';
-;
+</source>
-</source>
+= CPU/메모리/사양  =
+== 오라클서버의 메모리 ==
+* v$sgastat
+<source lang="sql">
+SELECT * FROM V$SGASTAT;
-== SQL_ID를 이용하여 SQL문의 PLAN 추출하기 ==
+SELECT POOL, SUM(BYTES) "SIZE"
-<source lang="sql">
+  FROM V$SGASTAT
-    SELECT LPAD(' ',4*(depth-1))||OPERATION||DECODE(OTHER_TAG,NULL,'','*')|| DECODE(OPTIONS,NULL,'',' ('||OPTIONS||')')|| DECODE(OBJECT_NAME,NULL,'',' OF '|| OBJECT_NAME||'')|| DECODE(OBJECT_TYPE,NULL,'',' ('||OBJECT_TYPE||')')|| DECODE(ID,0,DECODE(OPTIMIZER,NULL,'',' Optimizer='||OPTIMIZER))|| DECODE(COST,NULL,'',' (Cost='||COST|| DECODE(CARDINALITY,NULL,'',' Card='||CARDINALITY)|| DECODE(BYTES,NULL,'',' Bytes='||BYTES)||')')|| DECODE(ACCESS_PREDICATES,NULL,' ',' [AP] = '||ACCESS_PREDICATES)|| DECODE(ACCESS_PREDICATES,NULL,' ',' [FP] = '||FILTER_PREDICATES) AS PLAN_OUTPUT
+ WHERE POOL = 'SHARED POOL'
-    FROM DBA_HIST_SQL_PLAN WHERE sql_id = :sql_id
+ GROUP BY POOL
-ORDER BY ID,
+</source>
-         POSITION
+== 총 CPU Time 대비 SQL Parsing Time ==
+* V$SYSSTAT
+<source lang="sql">
+SELECT MAX(DECODE(Y.RNO,1,DECODE(X.NAME,'CPU USED BY THIS SESSION',X.VALUE))) AS "CPU USED BY THIS SESSION" ,
+       MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'PARSE TIME CPU',X.VALUE))) AS "PARSE TIME CPU" ,
+       ROUND((MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'PARSE TIME CPU',X.VALUE)))/MAX(DECODE(Y.RNO,1,DECODE(X.NAME,'CPU USED BY THIS SESSION',X.VALUE))) * 100),1)||'%' AS "PARSETIME/CPU(%)"
+FROM
+    ( SELECT NAME,
+             VALUE
+     FROM V$SYSSTAT
+     WHERE NAME IN ('CPU USED BY THIS SESSION',
+                    'PARSE TIME CPU') ) X,
+    (SELECT LEVEL AS RNO
+     FROM DUAL CONNECT BY LEVEL<=2) Y ;
+</source>
+== 총 Parsing Time 대비 Hard Parsing Time ==
+* V$SYS_TIME_MODEL
+<source lang="sql">
+SELECT MAX(DECODE(Y.RNO,1,DECODE(X.NAME,'PARSE TIME ELAPSED',X.VALUE))) AS "PARSE TIME ELAPSED" ,
+       MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'HARD PARSE ELAPSED TIME',X.VALUE))) AS "HARD PARSE ELAPSED TIME" ,
+       ROUND((MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'HARD PARSE ELAPSED TIME',X.VALUE)))/MAX(DECODE(Y.RNO,1,DECODE(X.NAME,'PARSE TIME ELAPSED',X.VALUE))) * 100),1)||'%' AS "HARD/PARSETORAL(%)"
+FROM
+    ( SELECT STAT_NAME AS NAME,
+             VALUE
+     FROM V$SYS_TIME_MODEL
+     WHERE STAT_NAME IN ('PARSE TIME ELAPSED',
+                         'HARD PARSE ELAPSED TIME') ) X,
+    (SELECT LEVEL AS RNO
+     FROM DUAL CONNECT BY LEVEL<=2) Y ;
+</source>
+== Log file sync 관련 대기현상 Sync Writes Time ==
+* redo synch time의 단위는 centisecond (1/100 초)
+* v$sysstat의 stat 중 time 정보를 담고 있는 stat의 value 단위는 centisecond이다.
+<source lang="sql">
+SELECT MAX(DECODE(Y.RNO,1,DECODE(X.NAME,'USER COMMITS',X.VALUE))) AS "USER COMMITS" ,
+       MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'REDO SYNCH WRITES',X.VALUE))) AS "REDO SYNCH WRITES" ,
+       MAX(DECODE(Y.RNO,3,DECODE(X.NAME,
+        'REDO SYNCH TIME',X.VALUE/100))) AS "REDO SYNCH TIME" ,
+       ROUND((MAX(DECODE(Y.RNO,3,DECODE(X.NAME,'REDO SYNCH TIME',X.VALUE/100)))/MAX(DECODE(Y.RNO,2,DECODE(X.NAME,'REDO SYNCH WRITES',X.VALUE)))),3)||'초' AS "SYNCTIME/SYNCWRITES(초)"
+FROM
+    ( SELECT NAME,
+             VALUE
+     FROM V$SYSSTAT
+     WHERE NAME IN ('USER COMMITS' ,
+                    'REDO SYNCH TIME' ,
+                    'REDO SYNCH WRITES') ) X,
+    (SELECT LEVEL AS RNO
+     FROM DUAL CONNECT BY LEVEL<=3) Y
 ;
 </source>
-== INDEX 정보 추출하기 ==
+----
+= INDEX 정보 추출하기 =
-===  Function-based Index 컬럼 찾기 ===
+==  Function-based Index 컬럼 찾기 ==
 <source lang="sql">
@@ 1,112번째 줄: / 1,272번째 줄: @@
 </source>
-=== 중복 인덱스 찾기(불필요 인덱스) ===
+== 중복 인덱스 찾기(불필요 인덱스) ==
 <source lang="sql">
 SELECT di.table_owner "OWNER" ,
         di.TABLE_NAME ,
-           dic1.index_name || chr(10) || ' (' || replace(dic1.index_cols,' ',',') || decode(sign(dic1.cnt-6),1,'...') || ')' "삭제대상 INDEX" ,
+           dic1.index_name || chr(10) || ' (' || replace(dic1.index_cols,' ',',') || decode(sign(dic1.cnt-6),1,'...') || ')' "삭제대상 INDEX" , dic2.index_name || chr(10) || ' (' || replace(dic2.index_cols,' ',',') || decode(sign(dic1.cnt-7),1,'...') || ')' "삭제원인 INDEX"
-                                                                                                                            dic2.index_name || chr(10) || ' (' || replace(dic2.index_cols,' ',',') || decode(sign(dic1.cnt-7),1,'...') || ')' "삭제원인 INDEX"
 FROM dba_indexes di ,
      ( SELECT table_owner,
@@ 1,158번째 줄: / 1,317번째 줄: @@
 </source>
-=== CONSTRAINTS 확인하기 ===
+== CONSTRAINTS 확인하기 ==
 -- 해당 테이블에 걸린 Constraints 확인하기
 <source lang="sql">
@@ 1,183번째 줄: / 1,342번째 줄: @@
 </source>
-=== 해당 테이블을 참조하는 테이블(자식) Constraints 확인하기 ===
+== 해당 테이블을 참조하는 테이블(자식) Constraints 확인하기 ==
 <source lang="sql">
 SELECT t.owner ,
@@ 1,209번째 줄: / 1,368번째 줄: @@
-=== v$segment_statistics 활용하기 ===
+== v$segment_statistics 활용하기 ==
+== Parameter 확인하기 ==
--- User별 전체 IO량(100%) 대비 IO량
 <source lang="sql">
-WITH io AS (
+SET linesize 200
-  SELECT /*+ materialize */ decode(statistic_name,'logical reads', 'lio','pio') io_type
+SET pagesize 100 col name
-        , sum(value) io_value
+FOR a30 col value
-  FROM v$segment_statistics
+FOR a10 col display_value
- WHERE statistic_name IN ( 'logical reads' ,
+FOR a10 col isdefault
-                           'physical reads' ,
+FOR a10 col ismodified
-                           'physical reads direct' )
+FOR a10
- GROUP BY decode(statistic_name,'logical reads', 'lio','pio')),
-per_io AS
+SELECT name,
-( SELECT OWNER,
+       value,
-         object_name,
+       isdefault,
-         object_type,
+        ismodified
-         seg_io,
+FROM v$parameter
-         seg_value,
+WHERE name IN ( 'lock_sga' ,
-         round(seg_value / decode(seg_io,'lio',
+                'db_cache_advice' ,
-                                      (SELECT io_value
+                'optimizer_dynamic_sampling' ,
-                                       FROM io
+                'session_cached_cursors' ,
-                                       WHERE io_type='lio'),
+                'sga_max_size' ,
-                                      (SELECT io_value
+                'sga_target' ,
-                                       FROM io
+                'db_cache_size' ,
-                                       WHERE io_type='pio') )*100,5) per_seg_io
+                'shared_pool_size' ,
- FROM
+                'shared_pool_reserved_size' ,
-     ( SELECT ss.OWNER,
+                'log_buffer' ,
-                 ss.object_name,
+                'skip_unusable_indexes' ,
-                 ss.object_type,
+                'pga_aggregate_target' ,
-                 decode(ss.statistic_name,'logical reads', 'lio','pio') seg_io,
+                'workarea_size_policy' ,
-                                                                        sum(ss.value) seg_value
+                'cpu_count' ,
-      FROM v$segment_statistics ss
+                'statistics_level') ;
-      WHERE ss.statistic_name IN ( 'logical reads' ,
+</source>
-                                   'physical reads' ,
-                                   'physical reads direct' )
+<source lang="sql">
-      GROUP BY ss.OWNER,
+SELECT ksppinm AS name,
-                  ss.object_name,
+       ksppstvl AS value
-                  ss.object_type,
+FROM sys.x$ksppi x
-                  decode(ss.statistic_name,'logical reads', 'lio','pio') ))
+   , sys.x$ksppcv y
-SELECT *
+WHERE (x.indx = y.indx)
-FROM
+AND (translate(ksppinm,'_','#') LIKE '%_optim_peek_user_binds%'
-( SELECT OWNER,
+     OR translate(ksppinm,'_','#') LIKE '%_kks_use_mutex_pin%'
-         sum(decode(seg_io,'lio',per_seg_io)) AS lio,
+     OR translate(ksppinm,'_','#') LIKE '%_gby_hash_aggregation_enabled%'
-         sum(decode(seg_io,'pio',per_seg_io)) AS pio
+     OR translate(ksppinm,'_','#') LIKE '%_gc_affinity_time%'
- FROM per_io
+     OR translate(ksppinm,'_','#') LIKE '%_optimizer_skip_scan_enabled%'
- GROUP BY OWNER
+     OR translate(ksppinm,'_','#') LIKE '%_pga_max_size%'
- ORDER BY sum(decode(seg_io,'lio',per_seg_io)) DESC)
+     OR translate(ksppinm,'_','#') LIKE '%_smm_max_size%'
-WHERE rownum <= 50 - Segment별 전체 IO량(100%) 대비 IO량 (TABLE/INDEX 분리) WITH io AS
+     OR translate(ksppinm,'_','#') LIKE '%_b_tree_bitmap_plans%'
-( SELECT /*+ materialize */ decode(statistic_name,'logical reads', 'lio','pio') io_type,
+     OR translate(ksppinm,'_','#') LIKE '%_undo_autotune%' ) ;
-                                                                                sum(value) io_value
+</source>
- FROM v$segment_statistics
- WHERE statistic_name IN ( 'logical reads' ,
+== Bind Peeked 확인하기 ==
-                           'physical reads' ,
-                           'physical reads direct' )
+<source lang="sql">
- GROUP BY decode(statistic_name,'logical reads', 'lio','pio')),
+SELECT p.plan_table_output
-                                                                        per_io AS
+  FROM
-( SELECT OWNER,
+( SELECT sql_id ,
-         object_name,
+         child_number
-         object_type,
+ FROM v$sql
-         seg_io,
+ WHERE hash_value=:hash_value ) s
-         seg_value,
+      , TABLE(dbms_xplan.display_cursor(s.sql_id , s.child_number , 'typical +peeked_binds')) p
-         round(seg_value / decode(seg_io,'lio',
+;
-                                      (SELECT io_value
+</source>
-                                       FROM io
-                                       WHERE io_type='lio'),
-                                      (SELECT io_value
+-- Group By 처리하는 방식 변경
-                                       FROM io
+<source lang="sql">
-                                       WHERE io_type='pio') )*100,5) per_seg_io
+파라미터 : _gby_hash_aggregation_enabled VERSION : 10g에서 신규 추가된 파라미터 해석방법 : FALSE
- FROM
+-
-      ( SELECT ss.OWNER,
+GROUP BY +
-                 ss.object_name,
+ORDER BY TRUE -
-                 ss.object_type,
+GROUP BY 주의사항 : 9i에서 10g로 업그레이드 수행 시 해당 파라미터가 True이면 기존
-                 decode(ss.statistic_name,'logical reads', 'lio','pio') seg_io,
+GROUP BY 절이 있는 SQL문의 정렬이 되지 않기 때문에 도출되는 결과가 바뀔 수 있다. 이럴 경우에는 해당 프로그램들에
-                                                                        sum(ss.value) seg_value
+ORDER by를 추가하여야 한다. _b_tree_bitmap_plans=TRUE
-      FROM v$segment_statistics ss
+</source>
-      WHERE ss.statistic_name IN ( 'logical reads' ,
-                                   'physical reads' ,
+== JOB 확인하기 ==
-                                   'physical reads direct' )
-      GROUP BY ss.OWNER,
-                  ss.object_name,
-                  ss.object_type,
-                  decode(ss.statistic_name,'logical reads', 'lio','pio') ))
+<source lang="sql">
 SELECT *
-   FROM per_io
+   FROM dba_jobs
-ORDER BY per_seg_io DESC
+ WHERE last_date >= to_date('20110906 14:00:00' , 'yyyymmdd hh24:mi:ss')
-;
+ORDER BY last_date
 </source>
-=== Parameter 확인하기 ===
 <source lang="sql">
+SELECT *
+  FROM dba_objects
+ WHERE object_name = upper('proc_olap_summ_measure_run_02')
+</source>
-SET linesize 200
+<source lang="sql">
-SET pagesize 100 col name
+SELECT *
-FOR a30 col value
+FROM v$sqlarea WHERE program_id = 100286
-FOR a10 col display_value
+;
-FOR a10 col isdefault
+</source>
-FOR a10 col ismodified
-FOR a10
+<source lang="sql">
+SELECT *
+FROM TABLE(DBMS_XPLAN.DISPLAY_CURSOR(NULL,NULL,’ALLSTATS COST LAST’));
+</source>
-SELECT name,
+== DB Link SQL CURSOR 확인하기 ==
-        value,
+<source lang="sql">
-        isdefault,
+SELECT inst_id ,
-        ismodified
+        kglnaown ,
-FROM v$parameter
+        kglnaobj ,
-WHERE name IN ( 'lock_sga' ,
+        kglnadlk , -- DB Link
-                'db_cache_advice' ,
+       decode(kglhdnsp , 0 , 'CURSOR' , 1 , 'TABLE/PROCEDURE' , 2 , 'BODY' , 3 , 'TRIGGER' , 4 , 'INDEX' , 5 , 'CLUSTER' , 6 , 'OBJECT' , 13 , 'JAVA SOURCE' , 14 , 'JAVA RESOURCE' , 15 , 'REPLICATED TABLE OBJECT' , 16 , 'REPLICATION INTERNAL PACKAGE' , 17 , 'CONTEXT POLICY' , 18 , 'PUB_SUB' , 19 , 'SUMMARY' , 20 , 'DIMENSION' , 21 , 'APP CONTEXT' , 22 , 'STORED OUTLINE' , 23 , 'RULESET' , 24 , 'RSRC PLAN' , 25 , 'RSRC CONSUMER GROUP' , 26 , 'PENDING RSRC PLAN' , 27 , 'PENDING RSRC CONSUMER GROUP' , 28 , 'SUBSCRIPTION' , 29 , 'LOCATION' , 30 , 'REMOTE OBJECT' , 31 , 'SNAPSHOT METADATA' , 32 , 'JAVA SHARED DATA' , 33 , 'SECURITY PROFILE' , 'INVALID NAMESPACE') ,
-                'optimizer_dynamic_sampling' ,
+       decode(bitand(kglobflg , 3) , 0 , 'NOT LOADED' , 2 , 'NON-EXISTENT' , 3 , 'INVALID STATUS' , decode(kglobtyp , 0 , 'CURSOR' , 1 , 'INDEX' , 2 , 'TABLE' , 3 , 'CLUSTER' , 4 , 'VIEW' , 5 , 'SYNONYM' , 6 , 'SEQUENCE' , 7 , 'PROCEDURE' , 8 , 'FUNCTION' , 9 , 'PACKAGE' , 10 , 'NON-EXISTENT' , 11 , 'PACKAGE BODY' , 12 , 'TRIGGER' , 13 , 'TYPE' , 14 , 'TYPE BODY' , 15 , 'OBJECT' , 16 , 'USER' , 17 , 'DBLINK' , 18 , 'PIPE' , 19 , 'TABLE PARTITION' , 20 , 'INDEX PARTITION' , 21 , 'LOB' , 22 , 'LIBRARY' , 23 , 'DIRECTORY' , 24 , 'QUEUE' , 25 , 'INDEX-ORGANIZED TABLE' , 26 , 'REPLICATION OBJECT GROUP' , 27 , 'REPLICATION PROPAGATOR' , 28 , 'JAVA SOURCE' , 29 , 'JAVA CLASS' , 30 , 'JAVA RESOURCE' , 31 , 'JAVA JAR' , 32 , 'INDEX TYPE' , 33 , 'OPERATOR' , 34 , 'TABLE SUBPARTITION' , 35 , 'INDEX SUBPARTITION' , 36 , 'REPLICATED TABLE OBJECT' , 37 , 'REPLICATION INTERNAL PACKAGE' , 38 , 'CONTEXT POLICY' , 39 , 'PUB_SUB' , 40 , 'LOB PARTITION' , 41 , 'LOB SUBPARTITION' , 42 , 'SUMMARY' , 43 , 'DIMENSION' , 44 , 'APP CONTEXT' , 45 , 'STORED OUTLINE' , 46 , 'RULESET' , 47 , 'RSRC PLAN' , 48 , 'RSRC CONSUMER GROUP' , 49 , 'PENDING RSRC PLAN' , 50 , 'PENDING RSRC CONSUMER GROUP' , 51 , 'SUBSCRIPTION' , 52 , 'LOCATION' , 53 , 'REMOTE OBJECT' , 54 , 'SNAPSHOT METADATA' , 55 , 'IFS' , 56 , 'JAVA SHARED DATA' , 57 , 'SECURITY PROFILE' , 'INVALID TYPE')) ,
-                'session_cached_cursors' ,
+       kglobhs0+kglobhs1+kglobhs2+kglobhs3+kglobhs4+kglobhs5+kglobhs6 ,
-                'sga_max_size' ,
+       kglhdldc ,
-                'sga_target' ,
+       kglhdexc ,
-                'db_cache_size' ,
+       kglhdlkc ,
-                'shared_pool_size' ,
+       kglobpc0 ,
-                'shared_pool_reserved_size' ,
+       decode(kglhdkmk , 0 , 'NO' , 'YES') ,
-                'log_buffer' ,
+       kglhdclt ,
-                'skip_unusable_indexes' ,
+       kglhdivc ,
-                'pga_aggregate_target' ,
+       kglhdpar AS paddress,
-                'workarea_size_policy' ,
+       ------->>> 추가한 부분
-                'cpu_count' ,
-                'statistics_level') ;
-</source>
-<source lang="sql">
+       kglhdadr AS address,
-SELECT ksppinm AS name,
+       ------->>> 추가한 부분
-        ksppstvl AS value
-FROM sys.x$ksppi x
+       kglnahsh AS hash_value,
-   , sys.x$ksppcv y
+        ------->>> 추가한 부분
-WHERE (x.indx = y.indx)
-AND (translate(ksppinm,'_','#') LIKE '%_optim_peek_user_binds%'
+       kglobt03 AS sql_id ------->>> 추가한 부분
-     OR translate(ksppinm,'_','#') LIKE '%_kks_use_mutex_pin%'
+FROM x$kglob
-     OR translate(ksppinm,'_','#') LIKE '%_gby_hash_aggregation_enabled%'
+WHERE kglhdnsp = 0 -- NAMESPACE='CURSOR'
-     OR translate(ksppinm,'_','#') LIKE '%_gc_affinity_time%'
+AND kglnadlk IS NOT NULL ;
-     OR translate(ksppinm,'_','#') LIKE '%_optimizer_skip_scan_enabled%'
-     OR translate(ksppinm,'_','#') LIKE '%_pga_max_size%'
-     OR translate(ksppinm,'_','#') LIKE '%_smm_max_size%'
-     OR translate(ksppinm,'_','#') LIKE '%_b_tree_bitmap_plans%'
-     OR translate(ksppinm,'_','#') LIKE '%_undo_autotune%' ) ;
 </source>
-=== Bind Peeked 확인하기 ===
+== INDEX Deleted ROWS CHECK [1] ==
+- Validate + index_stats
 <source lang="sql">
-SELECT p.plan_table_output
+SET serveroutput ON
-  FROM
+SET pagesize 0
-( SELECT sql_id ,
+SET linesize 200
-         child_number
+SET feedback OFF
- FROM v$sql
+SET timing OFF
- WHERE hash_value=:hash_value ) s
-     , TABLE(dbms_xplan.display_cursor(s.sql_id , s.child_number , 'typical +peeked_binds')) p
-;
-</source>
-<source lang="sql">
+--------------------------------------------------------------------------------
--- Group By 처리하는 방식 변경
+-- Work Time
-파라미터 : _gby_hash_aggregation_enabled VERSION : 10g에서 신규 추가된 파라미터 해석방법 : FALSE
+--------------------------------------------------------------------------------
--
-GROUP BY +
-ORDER BY TRUE -
-GROUP BY 주의사항 : 9i에서 10g로 업그레이드 수행 시 해당 파라미터가 True이면 기존
-GROUP BY 절이 있는 SQL문의 정렬이 되지 않기 때문에 도출되는 결과가 바뀔 수 있다. 이럴 경우에는 해당 프로그램들에
-ORDER by를 추가하여야 한다. _b_tree_bitmap_plans=TRUE
-=== JOB 확인하기 ===
+COLUMN the_date new_value run_date
+SELECT to_char(sysdate,'yyyymmddhh24mi') the_date
+FROM dual;
-<source lang="sql">
+--------------------------------------------------------------------------------
-SELECT *
+-- File Creation - Index Size Check Script
-  FROM dba_jobs
+--------------------------------------------------------------------------------
- WHERE last_date >= to_date('20110906 14:00:00' , 'yyyymmdd hh24:mi:ss')
+spool index_validate_check.SQL DECLARE
-ORDER BY last_date
+CURSOR cur_index_name IS
+SELECT OWNER,
+       index_name
+FROM dba_indexes
+WHERE OWNER = 'SYSTEM' ;
+BEGIN dbms_output.enable(buffer_size=>2000000);
+FOR index_rec IN cur_index_name LOOP dbms_output.put_line('Validate index '||index_rec.OWNER||'.'||index_rec.index_name||';');
+dbms_output.put_line(' ');
+dbms_output.put_line('select  '''||index_rec.OWNER||''',');
+dbms_output.put_line('        name index_name,');
+dbms_output.put_line('        lf_rows Tot_rows,');
+dbms_output.put_line('        del_lf_rows Deleted_Rows,');
+dbms_output.put_line('        (del_lf_rows * 100)/decode(lf_rows,0,0.1,lf_rows) "deleted(%)",');
+dbms_output.put_line('        round(btree_space/1024/1024/1024,3) "tot_space(GB)",');
+dbms_output.put_line('        round(used_space/1024/1024/1024,3) "used_space(GB)",');
+dbms_output.put_line('        pct_used "used_pct(%)",');
+dbms_output.put_line('        100 - pct_used "empty_pct(%)",');
+dbms_output.put_line('        decode( sign((del_lf_rows * 100)/decode(lf_rows,0,0.1,lf_rows) - 20),1,');
+dbms_output.put_line('                ''--->> Rebuild required'',');
+dbms_output.put_line('                ''--->> Good'') Check_Result');
+dbms_output.put_line(' from index_stats ;');
+END LOOP;
+END;
+/
 </source>
 <source lang="sql">
+--------------------------------------------------------------------------------
+-- Run - Index Size Check Script
+--------------------------------------------------------------------------------
+col OWNER format a10;
+col index_name format a33;
+col tot_rows format 99999999 col deleted_rows format 99999999 spool c:\total_index_size_check_&run_date..txt prompt OWNER INDEX name total ROWS deleted ROWS deleted(%) tot_space used_space used_pct(%) empty_pct(%) prompt
+---------------------------------------------------------------------------------------------------------------------------------
+@index_validate_check spool OFF
+SET feedback ON
+SET serveroutput OFF
+################################################################################
+## INDEX Deleted ROWS CHECK [2] - INDEX STATISTICS
+################################################################################
+col blevel heading "INDEX|BLEVEL"
+col leaf_blocks heading "LEAF|BLOCKS"
+col num_rows heading "NUM|ROWS"
+col usable_key_count_per_block heading "USABLE_KEY|COUNT|PER_BLOCK"
+col current_key_count_per_block heading "CURRENT_KEY|COUNT|PER_BLOCK"
+col used_percent heading "USED|PERCENT|(%)"
+col empty_percent heading "EMPTY|PERCENT|(%)"
 SELECT *
-  FROM dba_objects
+FROM
- WHERE object_name = upper('proc_olap_summ_measure_run_02')
+( SELECT /*+ leading(i) use_hash(i s) */ i.index_name,
-</source>
+                                         max(i.blevel) AS blevel,
+                                         max(i.leaf_blocks) AS leaf_blocks,
-<source lang="sql">
+                                         max(i.num_rows) AS num_rows,
-SELECT *
+                                         sum(s.blocks) AS "BLOCKS",
-FROM v$sqlarea WHERE program_id = 100286
+                                         sum(s.bytes)/1024/1024 AS "SIZE(MB)",
-;
+                                         max(i.usable_key_count_per_block) AS usable_key_count_per_block,
-</source>
+                                         max(i.current_key_count_per_block) AS current_key_count_per_block,
+                                         max(i.used_percent) AS used_percent,
-<source lang="sql">
+                                         max(i.empty_percent) AS empty_percent
-SELECT *
+ FROM
-FROM TABLE(DBMS_XPLAN.DISPLAY_CURSOR(NULL,NULL,’ALLSTATS COST LAST’));
+     ( SELECT i.index_name,
-</source>
+              i.blevel,
+              i.leaf_blocks,
-=== DB Link SQL CURSOR 확인하기 ===
+              i.num_rows,
-<source lang="sql">
+              round(i.usable_block_size / (i.avg_key_length),0) AS usable_key_count_per_block,
-SELECT inst_id ,
+              round(i.num_rows / i.leaf_blocks,0) AS current_key_count_per_block,
-       kglnaown ,
+              round((i.num_rows / i.leaf_blocks) / (i.usable_block_size / i.avg_key_length) * 100,0) AS used_percent,
-       kglnaobj ,
+              round(100 -((i.num_rows / i.leaf_blocks) / (i.usable_block_size / i.avg_key_length) * 100)) AS empty_percent
-       kglnadlk , -- DB Link
+      FROM
-       decode(kglhdnsp , 0 , 'CURSOR' , 1 , 'TABLE/PROCEDURE' , 2 , 'BODY' , 3 , 'TRIGGER' , 4 , 'INDEX' , 5 , 'CLUSTER' , 6 , 'OBJECT' , 13 , 'JAVA SOURCE' , 14 , 'JAVA RESOURCE' , 15 , 'REPLICATED TABLE OBJECT' , 16 , 'REPLICATION INTERNAL PACKAGE' , 17 , 'CONTEXT POLICY' , 18 , 'PUB_SUB' , 19 , 'SUMMARY' , 20 , 'DIMENSION' , 21 , 'APP CONTEXT' , 22 , 'STORED OUTLINE' , 23 , 'RULESET' , 24 , 'RSRC PLAN' , 25 , 'RSRC CONSUMER GROUP' , 26 , 'PENDING RSRC PLAN' , 27 , 'PENDING RSRC CONSUMER GROUP' , 28 , 'SUBSCRIPTION' , 29 , 'LOCATION' , 30 , 'REMOTE OBJECT' , 31 , 'SNAPSHOT METADATA' , 32 , 'JAVA SHARED DATA' , 33 , 'SECURITY PROFILE' , 'INVALID NAMESPACE') ,
+          ( SELECT i.index_name,
-       decode(bitand(kglobflg , 3) , 0 , 'NOT LOADED' , 2 , 'NON-EXISTENT' , 3 , 'INVALID STATUS' , decode(kglobtyp , 0 , 'CURSOR' , 1 , 'INDEX' , 2 , 'TABLE' , 3 , 'CLUSTER' , 4 , 'VIEW' , 5 , 'SYNONYM' , 6 , 'SEQUENCE' , 7 , 'PROCEDURE' , 8 , 'FUNCTION' , 9 , 'PACKAGE' , 10 , 'NON-EXISTENT' , 11 , 'PACKAGE BODY' , 12 , 'TRIGGER' , 13 , 'TYPE' , 14 , 'TYPE BODY' , 15 , 'OBJECT' , 16 , 'USER' , 17 , 'DBLINK' , 18 , 'PIPE' , 19 , 'TABLE PARTITION' , 20 , 'INDEX PARTITION' , 21 , 'LOB' , 22 , 'LIBRARY' , 23 , 'DIRECTORY' , 24 , 'QUEUE' , 25 , 'INDEX-ORGANIZED TABLE' , 26 , 'REPLICATION OBJECT GROUP' , 27 , 'REPLICATION PROPAGATOR' , 28 , 'JAVA SOURCE' , 29 , 'JAVA CLASS' , 30 , 'JAVA RESOURCE' , 31 , 'JAVA JAR' , 32 , 'INDEX TYPE' , 33 , 'OPERATOR' , 34 , 'TABLE SUBPARTITION' , 35 , 'INDEX SUBPARTITION' , 36 , 'REPLICATED TABLE OBJECT' , 37 , 'REPLICATION INTERNAL PACKAGE' , 38 , 'CONTEXT POLICY' , 39 , 'PUB_SUB' , 40 , 'LOB PARTITION' , 41 , 'LOB SUBPARTITION' , 42 , 'SUMMARY' , 43 , 'DIMENSION' , 44 , 'APP CONTEXT' , 45 , 'STORED OUTLINE' , 46 , 'RULESET' , 47 , 'RSRC PLAN' , 48 , 'RSRC CONSUMER GROUP' , 49 , 'PENDING RSRC PLAN' , 50 , 'PENDING RSRC CONSUMER GROUP' , 51 , 'SUBSCRIPTION' , 52 , 'LOCATION' , 53 , 'REMOTE OBJECT' , 54 , 'SNAPSHOT METADATA' , 55 , 'IFS' , 56 , 'JAVA SHARED DATA' , 57 , 'SECURITY PROFILE' , 'INVALID TYPE')) ,
+                   sum(c1.avg_col_len)+9 AS avg_key_length,
-       kglobhs0+kglobhs1+kglobhs2+kglobhs3+kglobhs4+kglobhs5+kglobhs6 ,
+                   max(t.block_size*0.98) usable_block_size,
-       kglhdldc ,
+                   max(i.blevel) blevel,
-       kglhdexc ,
+                   max(i.leaf_blocks) leaf_blocks,
-       kglhdlkc ,
+                   max(i.num_rows) num_rows
-       kglobpc0 ,
+           FROM dba_indexes i,
-       decode(kglhdkmk , 0 , 'NO' , 'YES') ,
+                dba_tab_columns c1,
-       kglhdclt ,
+                dba_ind_columns c2,
-       kglhdivc ,
+                dba_tablespaces t
-       kglhdpar AS paddress,
+           WHERE i.OWNER=upper('GCKDPROD') --and    i.table_name = upper('TSD_PTSEQ_I')
-       ------->>> 추가한 부분
+ --and    i.index_name LIKE '%'
-       kglhdadr AS address,
+               AND i.OWNER=c1.OWNER
-       ------->>> 추가한 부분
+               AND i.TABLE_NAME=c1.TABLE_NAME
+               AND c1.OWNER=c2.table_owner
-       kglnahsh AS hash_value,
+               AND c1.TABLE_NAME=c2.TABLE_NAME
-       ------->>> 추가한 부분
+               AND c1.COLUMN_NAME=c2.COLUMN_NAME
+               AND i.OWNER=c2.index_owner
-       kglobt03 AS sql_id ------->>> 추가한 부분
+               AND i.index_name=c2.index_name
-FROM x$kglob
+               AND i.tablespace_name=t.tablespace_name
-WHERE kglhdnsp = 0 -- NAMESPACE='CURSOR'
+           GROUP BY i.index_name HAVING max(i.leaf_blocks) >= 10000 ) i ) i,
-AND kglnadlk IS NOT NULL ;
+                                                                          dba_segments s
+ WHERE s.OWNER = 'GCKDPROD'
+     AND i.index_name = s.segment_name
+ GROUP BY i.index_name
+ ORDER BY max(i.empty_percent) DESC)
+WHERE rownum <= 50 ;
 </source>
-=== INDEX Deleted ROWS CHECK [1] ===
-- Validate + index_stats
 <source lang="sql">
-SET serveroutput ON
+################################################################################
-SET pagesize 0
+## Locking Contention Query
-SET linesize 200
+################################################################################
-SET feedback OFF
+SELECT OS_USER_NAME AS OSUSER,
-SET timing OFF
+       s.serial# AS "SERIAL NO.",
+       PROCESS AS "PID",
+       ORACLE_USERNAME AS "USERNAME",
+       l.SID AS "SID",
+       DECODE(l.TYPE -- Long locks, 'TM', 'dml/data enq (TM)', 'TX', 'transac enq (TX) ', 'UL', 'pls usr lock (UL)'
+-- Short locks, 'BL', 'buf hash tbl (BL)', 'CF', 'control file (CF)', 'CI', 'cross inst f (CI)', 'CU', 'cursor bind (CU) ', 'DF', 'data file (CF) ', 'DL', 'direct load (DL) ', 'DM', 'mount/strtup (DM)', 'DR', 'reco lock (DR) ', 'DX', 'distrib tran (DX)', 'FI', 'sga opn file (FI)', 'FS', 'file set (FS) ', 'IN', 'instance num (IN)', 'IR', 'instce recvr (IR)', 'IS', 'get state (IS) ', 'IV', 'libcache inv (IV)', 'JQ', 'job queue (JQ) ', 'KK', 'log sw kick (KK) ', 'LS', 'log switch (LS) ', 'MM', 'mount def (MM) ', 'MR', 'media recvry (MR)', 'PF', 'pwfile enq (PF) ', 'PR', 'process strt (PR)', 'RW', 'row wait (RW) ', 'RT', 'redo thread (RT) ', 'SC', 'scn enq (SC) ', 'SM', 'smon lock (SM) ', 'SN', 'seqno instce (SN)', 'SQ', 'seqno enq (SQ) ', 'ST', 'space transc (ST)', 'SV', 'seqno value (SV) ', 'TA', 'generic enq (TA) ', 'TD', 'dll enq (TD) ', 'TE', 'extend seg (TE) ', 'TS', 'temp segment (TS)', 'TT', 'temp table (TT) ', 'UN', 'user name (UN) ', 'WL', 'write redo (WL) ', 'TYPE = ' || l.TYPE) AS type,
+ decode(l.lmode, 0, 'none', 1, 'null', 2, 'RS', 3, 'RX', 4, 'S', 5, 'SRX', 6, 'Exclusive', TO_CHAR(l.lmode)) AS lmode,
+ decode(l.request, 0, 'none', 1, 'null', 2, 'RS', 3, 'RX', 4, 'S', 5, 'SRX', 6, 'X', TO_CHAR(l.request)) AS lrequest,
+ decode(BLOCK, 0, 'Not Blocking', 1, 'Blocking', 2, 'Global', block) AS "DETAIL",
+ OWNER,
+ object_name
+FROM sys.v_$locked_object lo,
+                          dba_objects DO,
+                                      sys.v_$lock l,
+(SELECT a.sid,
+        a.serial#
+ FROM v$session a,
+                v$bgprocess b
+ WHERE a.paddr = b.paddr(+)) s
+WHERE lo.object_id = DO.object_id
+AND l.sid = lo.session_id
+AND s.sid = l.sid
+;
+</source>
---------------------------------------------------------------------------------
+= ASH를 이용한 모니터링 =
--- Work Time
+* 최근 가장 많이 수행 된 SQL과 수행 점유율(수행 횟수)
---------------------------------------------------------------------------------
+* v$active_session_history
+<source lang="sql">
+SELECT sql_id ,
+       COUNT(*) ,
+       COUNT(*) *100/sum(COUNT(*)) over() pctload
+  FROM v$active_session_history
+ WHERE sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
+AND sample_time < to_date(:to_time,'yyyymmdd hh24miss')
+GROUP BY sql_id
+ORDER BY COUNT(*) DESC ;
+</source>
+== 특정 Session이 가장 많이 수행 된 SQL과 수행 점유율(수행 횟수) ==
+* v$active_session_history
+<source lang="sql">
+SELECT sql_id ,
+       COUNT(*) ,
+       COUNT(*) *100/sum(COUNT(*)) over() pctload
+FROM v$active_session_history
+WHERE sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
+AND sample_time < to_date(:to_time,'yyyymmdd hh24miss')
+AND session_id = :b1
+GROUP BY sql_id
+ORDER BY COUNT(*) DESC ;
+</source>
+== 주요한 이벤트 메트릭 표시 gv$eventmetric v$event_name ==
+* sqlplus 스크립트
+<source lang="sql">
+/*
+   eventmetric.sql - sqlplus script - displays significant event metrics
+*/
+col "Time /Delta" for a14
+col name for a40
+col INST_ID for 999
+set linesize 140
+set pagesize 1000
-COLUMN the_date new_value run_date
+set wrap off
-SELECT to_char(sysdate,'yyyymmddhh24mi') the_date
+select "Time /Delta",inst_id,name,
-FROM dual;
+        T_per_wait_fg*10 "Avg_FG_wait_ms", round(T_waited_fg/100,1) "Waited_FG_sec", W_count_fg "W_count_FG",
+        round(T_waited/100,1) "Waited_tot_sec", W_count "W_count_tot"
+from (
+  select to_char(min(begin_time),'hh24:mi:ss')||' /'||round(avg(intsize_csec/100),0)||'s' "Time /Delta",
+       em.inst_id,en.name,
+       sum(em.time_waited_fg) T_waited_fg, sum(em.time_waited) T_waited,sum(wait_count) W_count, sum(wait_count_fg) W_count_fg,
+       sum(decode(em.wait_count, 0,0,round(em.time_waited/em.wait_count,2))) T_per_wait,
+       sum(decode(em.wait_count_fg, 0,0,round(em.time_waited_fg/em.wait_count_fg,2))) T_per_wait_fg
+  from gv$eventmetric em
+     , v$event_name en
+  where em.event#=en.event#
+      and en.wait_class <>'Idle'
+  group by em.inst_id,en.name,em.event_id
+  order by T_waited_fg desc
+  )
+where rownum<=20;
+set wrap on
---------------------------------------------------------------------------------
+</source>
--- File Creation - Index Size Check Script
---------------------------------------------------------------------------------
-spool index_validate_check.SQL DECLARE
-CURSOR cur_index_name IS
-SELECT OWNER,
-       index_name
-FROM dba_indexes
-WHERE OWNER = 'SYSTEM' ;
-BEGIN dbms_output.enable(buffer_size=>2000000);
+== 중요한 시스템 메트릭 표시 gv$sysmetric ==
-FOR index_rec IN cur_index_name LOOP dbms_output.put_line('Validate index '||index_rec.OWNER||'.'||index_rec.index_name||';');
+<source lang="sql">
-dbms_output.put_line(' ');
+select "Time+Delta", "Metric",
-dbms_output.put_line('select  '''||index_rec.OWNER||''',');
+       case when "Total" >10000000 then '* '||round("Total"/1024/1024,0)||' M'
-dbms_output.put_line('        name index_name,');
+            when "Total" between 10000 and 10000000 then '+ '||round("Total"/1024,0)||' K'
-dbms_output.put_line('        lf_rows Tot_rows,');
+            when "Total" between 10 and 1024 then '  '||to_char(round("Total",0))
-dbms_output.put_line('        del_lf_rows Deleted_Rows,');
+            else '  '||to_char("Total")
-dbms_output.put_line('        (del_lf_rows * 100)/decode(lf_rows,0,0.1,lf_rows) "deleted(%)",');
+       end "Total"
-dbms_output.put_line('        round(btree_space/1024/1024/1024,3) "tot_space(GB)",');
+from (
-dbms_output.put_line('        round(used_space/1024/1024/1024,3) "used_space(GB)",');
+ select to_char(min(begin_time),'hh24:mi:ss')||' /'||round(avg(intsize_csec/100),0)||'s' "Time+Delta",
-dbms_output.put_line('        pct_used "used_pct(%)",');
+       metric_name||' - '||metric_unit "Metric",
-dbms_output.put_line('        100 - pct_used "empty_pct(%)",');
+       nvl(sum(value_inst1),0)+nvl(sum(value_inst2),0)+nvl(sum(value_inst3),0)+nvl(sum(value_inst4),0)+
-dbms_output.put_line('        decode( sign((del_lf_rows * 100)/decode(lf_rows,0,0.1,lf_rows) - 20),1,');
+       nvl(sum(value_inst5),0)+nvl(sum(value_inst6),0)+nvl(sum(value_inst7),0)+nvl(sum(value_inst8),0) "Total",
-dbms_output.put_line('                ''--->> Rebuild required'',');
+       sum(value_inst1) inst1, sum(value_inst2) inst2, sum(value_inst3) inst3, sum(value_inst4) inst4,
-dbms_output.put_line('                ''--->> Good'') Check_Result');
+        sum(value_inst5) inst5, sum(value_inst6) inst6, sum(value_inst7) inst7, sum(value_inst8) inst8
-dbms_output.put_line(' from index_stats ;');
+ from
-END LOOP;
+  ( select begin_time,intsize_csec,metric_name,metric_unit,metric_id,group_id,
-END;
+       case inst_id when 1 then round(value,1) end value_inst1,
-/
+        case inst_id when 2 then round(value,1) end value_inst2,
-</source>
+       case inst_id when 3 then round(value,1) end value_inst3,
+       case inst_id when 4 then round(value,1) end value_inst4,
+       case inst_id when 5 then round(value,1) end value_inst5,
+       case inst_id when 6 then round(value,1) end value_inst6,
+        case inst_id when 7 then round(value,1) end value_inst7,
+       case inst_id when 8 then round(value,1) end value_inst8
+  from gv$sysmetric
+  where metric_name in ('Host CPU Utilization (%)','Current OS Load', 'Physical Write Total IO Requests Per Sec',
+        'Physical Write Total Bytes Per Sec', 'Physical Write IO Requests Per Sec', 'Physical Write Bytes Per Sec',
+         'I/O Requests per Second', 'I/O Megabytes per Second',
+        'Physical Read Total Bytes Per Sec', 'Physical Read Total IO Requests Per Sec', 'Physical Read IO Requests Per Sec',
+        'CPU Usage Per Sec','Network Traffic Volume Per Sec','Logons Per Sec','Redo Generated Per Sec','Redo Writes Per Sec',
+        'User Transaction Per Sec','Average Active Sessions','Average Synchronous Single-Block Read Latency',
+        'Logical Reads Per Sec','DB Block Changes Per Sec')
+  )
+ group by metric_id,group_id,metric_name,metric_unit
+ order by metric_name
+);
+</source>
+== RAC 인스턴스별로 상세한 중요한 시스템 메트릭 gv$sysmetric ==
+<source lang="sql">
+select to_char(min(begin_time),'hh24:mi:ss')||' /'||round(avg(intsize_csec/100),0)||'s' "Time+Delta",
+       metric_name||' - '||metric_unit "Metric",
+       sum(value_inst1) inst1, sum(value_inst2) inst2, sum(value_inst3) inst3, sum(value_inst4) inst4,
+       sum(value_inst5) inst5, sum(value_inst6) inst6
+ from
+  ( select begin_time,intsize_csec,metric_name,metric_unit,metric_id,group_id,
+       case inst_id when 1 then round(value,1) end value_inst1,
+       case inst_id when 2 then round(value,1) end value_inst2,
+       case inst_id when 3 then round(value,1) end value_inst3,
+       case inst_id when 4 then round(value,1) end value_inst4,
+       case inst_id when 5 then round(value,1) end value_inst5,
+       case inst_id when 6 then round(value,1) end value_inst6
+  from gv$sysmetric
+  where metric_name in ('Host CPU Utilization (%)','Current OS Load', 'Physical Write Total IO Requests Per Sec',
+        'Physical Write Total Bytes Per Sec', 'Physical Write IO Requests Per Sec', 'Physical Write Bytes Per Sec',
+         'I/O Requests per Second', 'I/O Megabytes per Second',
+        'Physical Read Total Bytes Per Sec', 'Physical Read Total IO Requests Per Sec', 'Physical Read IO Requests Per Sec',
+        'CPU Usage Per Sec','Network Traffic Volume Per Sec','Logons Per Sec','Redo Generated Per Sec',
+        'User Transaction Per Sec','Average Active Sessions','Average Synchronous Single-Block Read Latency',
+        'Logical Reads Per Sec','DB Block Changes Per Sec')
+  )
+ group by metric_id,group_id,metric_name,metric_unit
+ order by metric_name;
+</source>
+== 특정 구간 이벤트 별 대기 시간 v$active_session_history ==
 <source lang="sql">
---------------------------------------------------------------------------------
+SELECT NVL(a.event, 'ON CPU') AS event,
--- Run - Index Size Check Script
+       COUNT(*) AS total_wait_time
---------------------------------------------------------------------------------
+FROM v$active_session_history a
-col OWNER format a10;
+WHERE sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
-col index_name format a33;
+AND sample_time < to_date(:to_time,'yyyymmdd hh24miss')
-col tot_rows format 99999999 col deleted_rows format 99999999 spool c:\total_index_size_check_&run_date..txt prompt OWNER INDEX name total ROWS deleted ROWS deleted(%) tot_space used_space used_pct(%) empty_pct(%) prompt
+GROUP BY a.event
----------------------------------------------------------------------------------------------------------------------------------
+ORDER BY total_wait_time DESC;
-@index_validate_check spool OFF
+</source>
-SET feedback ON
-SET serveroutput OFF
+== 특정 구간 CPU 점유율 순 - TOP SQL ==
-################################################################################
+<source lang="sql">
-## INDEX Deleted ROWS CHECK [2] - INDEX STATISTICS
+SELECT ash.session_id ,
-################################################################################
+       ash.session_serial# ,
-col blevel heading "INDEX|BLEVEL"
+       ash.user_id ,
-col leaf_blocks heading "LEAF|BLOCKS"
+       ash.program ,
-col num_rows heading "NUM|ROWS"
+       MAX(en.name) event_name ,
-col usable_key_count_per_block heading "USABLE_KEY|COUNT|PER_BLOCK"
+       SUM(decode(ash.session_state , 'ON CPU' , 1 , 0)) "CPU" ,
-col current_key_count_per_block heading "CURRENT_KEY|COUNT|PER_BLOCK"
+       SUM(decode(ash.session_state , 'WAITING' , 1 , 0)) - SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "WAITING" ,
-col used_percent heading "USED|PERCENT|(%)"
+       SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "IO" ,
-col empty_percent heading "EMPTY|PERCENT|(%)"
+       SUM(decode(session_state , 'ON CPU' , 1 , 1)) "TOTAL"
-SELECT *
+FROM v$active_session_history ash
-FROM
+   , v$event_name en
-( SELECT /*+ leading(i) use_hash(i s) */ i.index_name,
+WHERE en.event# = ash.event#
-                                         max(i.blevel) AS blevel,
+AND ash.sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
-                                         max(i.leaf_blocks) AS leaf_blocks,
+AND ash.sample_time < to_date(:to_time,'yyyymmdd hh24miss')
-                                         max(i.num_rows) AS num_rows,
+GROUP BY session_id ,
-                                         sum(s.blocks) AS "BLOCKS",
+         user_id ,
-                                         sum(s.bytes)/1024/1024 AS "SIZE(MB)",
+         session_serial# ,
-                                         max(i.usable_key_count_per_block) AS usable_key_count_per_block,
+         program
-                                         max(i.current_key_count_per_block) AS current_key_count_per_block,
+ORDER BY SUM(decode(session_state , 'ON CPU' , 1 , 1))
-                                         max(i.used_percent) AS used_percent,
+;
-                                         max(i.empty_percent) AS empty_percent
+</source>
- FROM
-     ( SELECT i.index_name,
+== 특정 구간 CPU 점유율 순 - TOP SESSION v$active_session_history ==
-              i.blevel,
+<source lang="sql">
-              i.leaf_blocks,
+SELECT ash.session_id ,
-              i.num_rows,
+       ash.session_serial# ,
-              round(i.usable_block_size / (i.avg_key_length),0) AS usable_key_count_per_block,
+       ash.user_id ,
-              round(i.num_rows / i.leaf_blocks,0) AS current_key_count_per_block,
+       ash.program ,
-              round((i.num_rows / i.leaf_blocks) / (i.usable_block_size / i.avg_key_length) * 100,0) AS used_percent,
+       SUM(decode(ash.session_state , 'ON CPU' , 1 , 0)) "CPU" ,
-              round(100 -((i.num_rows / i.leaf_blocks) / (i.usable_block_size / i.avg_key_length) * 100)) AS empty_percent
+       SUM(decode(ash.session_state , 'WAITING' , 1 , 0)) - SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "WAITING" ,
-      FROM
+       SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "IO" ,
-          ( SELECT i.index_name,
+       SUM(decode(session_state , 'ON CPU' , 1 , 1)) "TOTAL"
-                   sum(c1.avg_col_len)+9 AS avg_key_length,
+FROM v$active_session_history ash
-                   max(t.block_size*0.98) usable_block_size,
+   , v$event_name en
-                   max(i.blevel) blevel,
+WHERE en.event# = ash.event#
-                   max(i.leaf_blocks) leaf_blocks,
+AND ash.sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
-                   max(i.num_rows) num_rows
+AND ash.sample_time < to_date(:to_time,'yyyymmdd hh24miss')
-           FROM dba_indexes i,
+GROUP BY session_id ,
-                dba_tab_columns c1,
+         user_id ,
-                dba_ind_columns c2,
+         session_serial# ,
-                dba_tablespaces t
+         program
-           WHERE i.OWNER=upper('GCKDPROD') --and    i.table_name = upper('TSD_PTSEQ_I')
+ORDER BY SUM(decode(session_state , 'ON CPU' , 1 , 1))
- --and    i.index_name LIKE '%'
+;
+</source>
-               AND i.OWNER=c1.OWNER
-               AND i.TABLE_NAME=c1.TABLE_NAME
-               AND c1.OWNER=c2.table_owner
-               AND c1.TABLE_NAME=c2.TABLE_NAME
-               AND c1.COLUMN_NAME=c2.COLUMN_NAME
-               AND i.OWNER=c2.index_owner
-               AND i.index_name=c2.index_name
-               AND i.tablespace_name=t.tablespace_name
-           GROUP BY i.index_name HAVING max(i.leaf_blocks) >= 10000 ) i ) i,
-                                                                          dba_segments s
- WHERE s.OWNER = 'GCKDPROD'
-     AND i.index_name = s.segment_name
- GROUP BY i.index_name
- ORDER BY max(i.empty_percent) DESC)
-WHERE rownum <= 50 ;
-</source>
+== 특정 구간 수행 이력 v$active_session_history ==
 <source lang="sql">
-################################################################################
+SELECT ash.sample_time TIME ,
-## Locking Contention Query
+        ash.session_id ,
-################################################################################
+        ash.session_serial# ,
-SELECT OS_USER_NAME AS OSUSER,
+        ash.user_id ,
-        s.serial# AS "SERIAL NO.",
+        ash.program ,
-        PROCESS AS "PID",
+        ash.module ,
-        ORACLE_USERNAME AS "USERNAME",
+       ash.client_id ,
-        l.SID AS "SID",
+       ash.machine ,
-        DECODE(l.TYPE -- Long locks, 'TM', 'dml/data enq (TM)', 'TX', 'transac enq (TX) ', 'UL', 'pls usr lock (UL)'
+       ash.session_state ,
--- Short locks, 'BL', 'buf hash tbl (BL)', 'CF', 'control file (CF)', 'CI', 'cross inst f (CI)', 'CU', 'cursor bind (CU) ', 'DF', 'data file (CF) ', 'DL', 'direct load (DL) ', 'DM', 'mount/strtup (DM)', 'DR', 'reco lock (DR) ', 'DX', 'distrib tran (DX)', 'FI', 'sga opn file (FI)', 'FS', 'file set (FS) ', 'IN', 'instance num (IN)', 'IR', 'instce recvr (IR)', 'IS', 'get state (IS) ', 'IV', 'libcache inv (IV)', 'JQ', 'job queue (JQ) ', 'KK', 'log sw kick (KK) ', 'LS', 'log switch (LS) ', 'MM', 'mount def (MM) ', 'MR', 'media recvry (MR)', 'PF', 'pwfile enq (PF) ', 'PR', 'process strt (PR)', 'RW', 'row wait (RW) ', 'RT', 'redo thread (RT) ', 'SC', 'scn enq (SC) ', 'SM', 'smon lock (SM) ', 'SN', 'seqno instce (SN)', 'SQ', 'seqno enq (SQ) ', 'ST', 'space transc (ST)', 'SV', 'seqno value (SV) ', 'TA', 'generic enq (TA) ', 'TD', 'dll enq (TD) ', 'TE', 'extend seg (TE) ', 'TS', 'temp segment (TS)', 'TT', 'temp table (TT) ', 'UN', 'user name (UN) ', 'WL', 'write redo (WL) ', 'TYPE = ' || l.TYPE) AS type,
+       ash.event ,
- decode(l.lmode, 0, 'none', 1, 'null', 2, 'RS', 3, 'RX', 4, 'S', 5, 'SRX', 6, 'Exclusive', TO_CHAR(l.lmode)) AS lmode,
+       ash.sql_id ,
- decode(l.request, 0, 'none', 1, 'null', 2, 'RS', 3, 'RX', 4, 'S', 5, 'SRX', 6, 'X', TO_CHAR(l.request)) AS lrequest,
+       ash.blocking_session ,
- decode(BLOCK, 0, 'Not Blocking', 1, 'Blocking', 2, 'Global', block) AS "DETAIL",
+       ash.current_obj# ,
- OWNER,
+       ash.current_file# ,
- object_name
+       ash.pga_allocated ,
-FROM sys.v_$locked_object lo,
+       ash.temp_space_allocated
-                          dba_objects DO,
+FROM v$active_session_history ash
-                                      sys.v_$lock l,
+WHERE ash.sample_time >= TO_DATE(:from_time , 'yyyymmdd hh24miss')
-(SELECT a.sid,
+AND ash.sample_time < TO_DATE(:to_time , 'yyyymmdd hh24miss')
-        a.serial#
+ORDER BY ash.sample_time DESC
- FROM v$session a,
-                v$bgprocess b
- WHERE a.paddr = b.paddr(+)) s
-WHERE lo.object_id = DO.object_id
-AND l.sid = lo.session_id
-AND s.sid = l.sid
 ;
 </source>
-== ASH를 이용한 모니터링 ==
+= AWR을 이용한 모니터링 =
-* 최근 가장 많이 수행 된 SQL과 수행 점유율(수행 횟수) *
+----
+== AWR을 이용한 literal SQL 추출 방법 ==
+* get_literal_sql.SQL
 <source lang="sql">
-SELECT sql_id ,
+SET linesize 120 col SCHEMA_NAME
-       COUNT(*) ,
+FOR a8 col module_name
-       COUNT(*) *100/sum(COUNT(*)) over() pctload
+FOR a15 col sql_text
-  FROM v$active_session_history
+FOR a35 col SCHEMA_NAME heading "SCHEMA|NAME" col module_name heading "MODULE|NAME" col literal_sql_cnt heading "LITERAL|CNT"
- WHERE sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
+FOR 99999 col execution_cnt heading "EXECUTION|CNT"
-AND sample_time < to_date(:to_time,'yyyymmdd hh24miss')
+FOR 99999 col plan_cnt heading "PLAN|CNT"
-GROUP BY sql_id
+FOR 999 col literal_sql_ratio heading "LITERAL|RATIO(%)"
-ORDER BY COUNT(*) DESC ;
+FOR 999 accept i_begin_time prompt 'Enter begin time[YYYYMMDDHH24]: ' accept i_end_time prompt 'Enter end   time[YYYYMMDDHH24]: ' VARIABLE v_begin_time char(10) VARIABLE v_end_time char(10) EXEC :v_begin_time:=&i_begin_time EXEC :v_end_time :=&i_end_time
-</source>
-=== 특정 Session이 가장 많이 수행 된 SQL과 수행 점유율(수행 횟수) ===
+SELECT a.SCHEMA_NAME , a.module_name , a.literal_sql_cnt , a.execution_cnt , round(a.literal_sql_cnt/a.execution_cnt*100,1) literal_sql_ratio, a.plan_cnt,
-<source lang="sql">
+    ( SELECT sql_text
-SELECT sql_id ,
+     FROM dba_hist_sqltext
-       COUNT(*) ,
+     WHERE dbid = a.dbid
-       COUNT(*) *100/sum(COUNT(*)) over() pctload
+         AND sql_id = a.max_sql_id ) sql_text
-FROM v$active_session_history
+FROM
-WHERE sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
+    ( SELECT s.force_matching_signature,
-AND sample_time < to_date(:to_time,'yyyymmdd hh24miss')
+             s.dbid ,
-AND session_id = :b1
+             s.parsing_schema_name SCHEMA_NAME ,
-GROUP BY sql_id
+                                   s.module module_name ,
-ORDER BY COUNT(*) DESC ;
+                                   count(*) literal_sql_cnt ,
-</source>
+                                   sum(s.executions_total) execution_cnt ,
+                                   max(s.sql_id) max_sql_id,
-=== 특정 구간 이벤트 별 대기 시간 ===
+                                   count(DISTINCT s.plan_hash_value) plan_cnt
-<source lang="sql">
+     FROM sys.wrm$_snapshot w ,
-SELECT NVL(a.event, 'ON CPU') AS event,
+                            dba_hist_sqlstat s
-       COUNT(*) AS total_wait_time
+     WHERE w.begin_interval_time >= to_timestamp(:v_begin_time , 'yyyymmddhh24')
-FROM v$active_session_history a
+         AND w.end_interval_time <= to_timestamp(:v_end_time , 'yyyymmddhh24')
-WHERE sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
+         AND w.snap_id = s.snap_id
-AND sample_time < to_date(:to_time,'yyyymmdd hh24miss')
+         AND w.dbid = s.dbid
-GROUP BY a.event
+         AND w.instance_number = s.instance_number
-ORDER BY total_wait_time DESC;
+         AND s.parsing_schema_name <> 'SYS'
+     GROUP BY s.dbid ,
+              s.force_matching_signature ,
+              s.parsing_schema_name ,
+              s.module
+     ORDER BY count(*) DESC ) a
+WHERE rownum<=10;
 </source>
-=== 특정 구간 CPU 점유율 순 - TOP SQL ===
+== AWR SQL ordered BY Elapsed Time ==
 <source lang="sql">
-SELECT ash.session_id ,
-       ash.session_serial# ,
-       ash.user_id ,
-       ash.program ,
-       MAX(en.name) event_name ,
-       SUM(decode(ash.session_state , 'ON CPU' , 1 , 0)) "CPU" ,
-       SUM(decode(ash.session_state , 'WAITING' , 1 , 0)) - SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "WAITING" ,
-       SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "IO" ,
-       SUM(decode(session_state , 'ON CPU' , 1 , 1)) "TOTAL"
-FROM v$active_session_history ash ,
-                              v$event_name en
-WHERE en.event# = ash.event#
-AND ash.sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
-AND ash.sample_time < to_date(:to_time,'yyyymmdd hh24miss')
-GROUP BY session_id ,
-         user_id ,
-         session_serial# ,
-         program
-ORDER BY SUM(decode(session_state , 'ON CPU' , 1 , 1))
-;
-</source>
-=== 특정 구간 CPU 점유율 순 - TOP SESSION ===
+def dbid="&1"
-<source lang="sql">
+def inst_num="&2"
-SELECT ash.session_id ,
+def snap_start="&3"
-       ash.session_serial# ,
+def snap_end="&4"
-       ash.user_id ,
+def top_num="&5"
-       ash.program ,
+col ela_Time format 999990.00 heading "ElaTime|(s)"
-       SUM(decode(ash.session_state , 'ON CPU' , 1 , 0)) "CPU" ,
+col EXEC format 9999990 heading "Exec"
-       SUM(decode(ash.session_state , 'WAITING' , 1 , 0)) - SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "WAITING" ,
+col elapexec format 999990.00 heading "ElaTimeperexec (s)"
-       SUM(decode(ash.session_state , 'WAITING' , decode(en.wait_class , 'User I/O' , 1 , 0) , 0)) "IO" ,
+--col Total               format 9999990.00         heading "DBTime%"
-       SUM(decode(session_state , 'ON CPU' , 1 , 1)) "TOTAL"
+col CPU format 999990.00 heading "CPU%"
-FROM v$active_session_history ash ,
+col IO format 999990.00 heading "IO%"
-                              v$event_name en
+col SQLId format a15 heading "SQL_ID"
-WHERE en.event# = ash.event#
+col SQLModule format a50 heading "SQL_Module"
-AND ash.sample_time >= to_date(:from_time,'yyyymmdd hh24miss')
+col SQLText format a50 heading "SQL_Text"
-AND ash.sample_time < to_date(:to_time,'yyyymmdd hh24miss')
-GROUP BY session_id ,
+WITH sqt AS
-         user_id ,
+(SELECT elap,
-         session_serial# ,
+        cput,
-         program
+        EXEC,
-ORDER BY SUM(decode(session_state , 'ON CPU' , 1 , 1))
+        norm_val,
+        iowt,
+        sql_id,
+        module,
+        rnum
+ FROM
+     (SELECT sql_id,
+             module,
+             elap,
+             norm_val,
+             cput,
+             EXEC,
+             iowt,
+             rownum rnum
+      FROM
+          (SELECT sql_id,
+                  MAX(module) module,
+                              SUM(elapsed_time_delta) elap, (100 * (SUM(elapsed_time_delta) / NULLIF(
+                                                                                                         (SELECT (SUM(e.VALUE) - SUM(b.VALUE)) / 1000000 / 60
+                                                                                                          FROM DBA_HIST_SYS_TIME_MODEL e, DBA_HIST_SYS_TIME_MODEL b
+                                                                                                          WHERE e.SNAP_ID = &snap_end
+                                                                                                              AND e.DBID = &dbid
+                                                                                                              AND e.INSTANCE_NUMBER = &inst_num
+                                                                                                              AND e.STAT_NAME = 'DB time'
+                                                                                                              AND b.SNAP_ID = &snap_start
+                                                                                                              AND b.DBID = &dbid
+                                                                                                              AND b.INSTANCE_NUMBER = &inst_num
+                                                                                                              AND b.STAT_NAME = 'DB time'), 0))) norm_val,
+                                                                                                                                                 SUM(cpu_time_delta) cput,
+                                                                                                                                                                     SUM(executions_delta) EXEC,
+                                                                                                                                                                                           SUM(iowait_delta) iowt
+           FROM dba_hist_sqlstat
+           WHERE dbid = &dbid
+               AND instance_number = &inst_num
+               AND &snap_start < snap_id
+               AND snap_id <= &snap_end
+           GROUP BY sql_id
+           ORDER BY nvl(SUM(elapsed_time_delta), -1) DESC, sql_id))
+ WHERE rnum < &top_num --  and (rnum <= :tsql_min or norm_val > :top_pct_sql)
+ )
 ;
 </source>
-=== 특정 구간 수행 이력 ===
-<source lang="sql">
-SELECT ash.sample_time TIME ,
-       ash.session_id ,
-       ash.session_serial# ,
-       ash.user_id ,
-       ash.program ,
-       ash.module ,
-       ash.client_id ,
-       ash.machine ,
-       ash.session_state ,
-       ash.event ,
-       ash.sql_id ,
-       ash.blocking_session ,
-       ash.current_obj# ,
-       ash.current_file# ,
-       ash.pga_allocated ,
-       ash.temp_space_allocated
-FROM v$active_session_history ash
-WHERE ash.sample_time >= TO_DATE(:from_time , 'yyyymmdd hh24miss')
-AND ash.sample_time < TO_DATE(:to_time , 'yyyymmdd hh24miss')
-ORDER BY ash.sample_time DESC
-;
-</source>
-== AWR을 이용한 모니터링 ==
-----
-=== AWR을 이용한 literal SQL 추출 방법 ===
-* get_literal_sql.SQL
 <source lang="sql">
-SET linesize 120 col SCHEMA_NAME
+SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.elap / 1000000), to_number(NULL)) ela_Time,
-FOR a8 col module_name
+                            sqt.EXEC EXEC,
-FOR a15 col sql_text
+                                     decode(sqt.EXEC, 0, to_number(NULL), (sqt.elap / sqt.EXEC / 1000000)) elapexec,
-FOR a35 col SCHEMA_NAME heading "SCHEMA|NAME" col module_name heading "MODULE|NAME" col literal_sql_cnt heading "LITERAL|CNT"
+                                     sqt.norm_val Total ,
-FOR 99999 col execution_cnt heading "EXECUTION|CNT"
+                                     --(100 * (sum(elap) /nullif(dbtime.dbtime,0))) norm_val
-FOR 99999 col plan_cnt heading "PLAN|CNT"
+ --null Total,
-FOR 999 col literal_sql_ratio heading "LITERAL|RATIO(%)"
-FOR 999 accept i_begin_time prompt 'Enter begin time[YYYYMMDDHH24]: ' accept i_end_time prompt 'Enter end   time[YYYYMMDDHH24]: ' VARIABLE v_begin_time char(10) VARIABLE v_end_time char(10) EXEC :v_begin_time:=&i_begin_time EXEC :v_end_time :=&i_end_time
-SELECT a.SCHEMA_NAME , a.module_name , a.literal_sql_cnt , a.execution_cnt , round(a.literal_sql_cnt/a.execution_cnt*100,1) literal_sql_ratio, a.plan_cnt,
-    ( SELECT sql_text
-     FROM dba_hist_sqltext
-     WHERE dbid = a.dbid
-         AND sql_id = a.max_sql_id ) sql_text
-FROM
-    ( SELECT s.force_matching_signature,
-             s.dbid ,
-             s.parsing_schema_name SCHEMA_NAME ,
-                                   s.module module_name ,
-                                   count(*) literal_sql_cnt ,
-                                   sum(s.executions_total) execution_cnt ,
-                                   max(s.sql_id) max_sql_id,
-                                   count(DISTINCT s.plan_hash_value) plan_cnt
-     FROM sys.wrm$_snapshot w ,
-                            dba_hist_sqlstat s
-     WHERE w.begin_interval_time >= to_timestamp(:v_begin_time , 'yyyymmddhh24')
-         AND w.end_interval_time <= to_timestamp(:v_end_time , 'yyyymmddhh24')
-         AND w.snap_id = s.snap_id
-         AND w.dbid = s.dbid
-         AND w.instance_number = s.instance_number
-         AND s.parsing_schema_name <> 'SYS'
-     GROUP BY s.dbid ,
-              s.force_matching_signature ,
-              s.parsing_schema_name ,
-              s.module
-     ORDER BY count(*) DESC ) a
-WHERE rownum<=10;
-</source>
+                                     decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))) CPU,
+                                     decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.iowt / sqt.elap))) IO,
+                                     sqt.sql_id SQLId,
+                                     to_clob(decode(sqt.module, NULL, NULL, 'Module: ' || sqt.module)) SQLModule,
+                                     nvl(substr(st.sql_text, 0, 50), to_clob('** SQL Text Not Available **')) SQLText
+FROM sqt,
+     dba_hist_sqltext st
+WHERE st.sql_id(+) = sqt.sql_id
+AND st.dbid(+) = &dbid
+ORDER BY sqt.rnum
+/
+</source>
-=== AWR SQL ordered BY Elapsed Time ===
+== AWR SQL ordered BY Elapsed Time ==
 <source lang="sql">
+undef dbid
+undef inst_num
+undef snap_start
+undef snap_end
+undef sql_id
+undef top_num
-def dbid="&1"
+WITH sqt AS
-def inst_num="&2"
+(SELECT elap,
-def snap_start="&3"
-def snap_end="&4"
-def top_num="&5"
-col ela_Time format 999990.00 heading "ElaTime|(s)"
-col EXEC format 9999990 heading "Exec"
-col elapexec format 999990.00 heading "ElaTimeperexec (s)"
---col Total               format 9999990.00         heading "DBTime%"
-col CPU format 999990.00 heading "CPU%"
-col IO format 999990.00 heading "IO%"
-col SQLId format a15 heading "SQL_ID"
-col SQLModule format a50 heading "SQL_Module"
-col SQLText format a50 heading "SQL_Text"
-WITH sqt AS
-(SELECT elap,
          cput,
          EXEC,
+        iowt,
          norm_val,
-        iowt,
          sql_id,
          module,
@@ 1,775번째 줄: / 2,030번째 줄: @@
               rownum rnum
        FROM
-           (SELECT sql_id,
+           (SELECT sql_id ,
-                   MAX(module) module,
+                   MAX(module) module ,
-                               SUM(elapsed_time_delta) elap, (100 * (SUM(elapsed_time_delta) / NULLIF(
+                               SUM(elapsed_time_delta) elap , (100 * (SUM(elapsed_time_delta) / NULLIF(:dbtime,0))) norm_val ,
-                                                                                                         (SELECT (SUM(e.VALUE) - SUM(b.VALUE)) / 1000000 / 60
+                                                                                                                   SUM(cpu_time_delta) cput ,                                                                                                                                       SUM(executions_delta) EXEC ,                                                                                                                                                             SUM(iowait_delta) iowt
-                                                                                                          FROM DBA_HIST_SYS_TIME_MODEL e, DBA_HIST_SYS_TIME_MODEL b
+            FROM dba_hist_sqlstat
-                                                                                                          WHERE e.SNAP_ID = &snap_end
+            WHERE dbid = :dbid
-                                                                                                              AND e.DBID = &dbid
+                AND instance_number = :inst_num
-                                                                                                              AND e.INSTANCE_NUMBER = &inst_num
+                AND :bid < snap_id
-                                                                                                              AND e.STAT_NAME = 'DB time'
+                AND snap_id <= :eid
-                                                                                                              AND b.SNAP_ID = &snap_start
-                                                                                                              AND b.DBID = &dbid
-                                                                                                              AND b.INSTANCE_NUMBER = &inst_num
-                                                                                                              AND b.STAT_NAME = 'DB time'), 0))) norm_val,
-                                                                                                                                                 SUM(cpu_time_delta) cput,
-                                                                                                                                                                     SUM(executions_delta) EXEC,
-                                                                                                                                                                                           SUM(iowait_delta) iowt
-            FROM dba_hist_sqlstat
-            WHERE dbid = &dbid
-                AND instance_number = &inst_num
-                AND &snap_start < snap_id
-                AND snap_id <= &snap_end
             GROUP BY sql_id
             ORDER BY nvl(SUM(elapsed_time_delta), -1) DESC, sql_id))
-  WHERE rnum < &top_num --  and (rnum <= :tsql_min or norm_val > :top_pct_sql)
+  WHERE rnum < :tsql_max
- )
+     AND (rnum <= :tsql_min
+          OR norm_val > :top_pct_sql))
 ;
 </source>
 <source lang="sql">
-SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.elap / 1000000), to_number(NULL)) ela_Time,
+SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.elap/1000000), to_number(NULL)),
-                             sqt.EXEC EXEC,
+                             sqt.EXEC,
-                                     decode(sqt.EXEC, 0, to_number(NULL), (sqt.elap / sqt.EXEC / 1000000)) elapexec,
+                                decode(sqt.EXEC, 0, to_number(NULL),(sqt.elap / sqt.EXEC / 1000000)),
-                                     sqt.norm_val Total ,
+                                sqt.norm_val,
-                                     --(100 * (sum(elap) /nullif(dbtime.dbtime,0))) norm_val
+                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))),
- --null Total,
+                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.iowt / sqt.elap))),
+                                sqt.sql_id,
-                                     decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))) CPU,
+                                to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
-                                     decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.iowt / sqt.elap))) IO,
+                                nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
-                                     sqt.sql_id SQLId,
-                                     to_clob(decode(sqt.module, NULL, NULL, 'Module: ' || sqt.module)) SQLModule,
-                                     nvl(substr(st.sql_text, 0, 50), to_clob('** SQL Text Not Available **')) SQLText
 FROM sqt,
       dba_hist_sqltext st
 WHERE st.sql_id(+) = sqt.sql_id
-AND st.dbid(+) = &dbid
+AND st.dbid(+) = :dbid
 ORDER BY sqt.rnum
-/
+;
 </source>
-=== AWR SQL ordered BY Elapsed Time ===
+== AWR SQL ordered BY CPU Time ==
 <source lang="sql">
-undef dbid
-undef inst_num
-undef snap_start
-undef snap_end
-undef sql_id
-undef top_num
 WITH sqt AS
 (SELECT elap,
          cput,
          EXEC,
-         iowt,
+         uiot,
          norm_val,
          sql_id,
@@ 1,850번째 줄: / 2,084번째 줄: @@
               cput,
               EXEC,
-              iowt,
+              uiot,
               rownum rnum
        FROM
            (SELECT sql_id ,
                    MAX(module) module ,
-                               SUM(elapsed_time_delta) elap , (100 * (SUM(elapsed_time_delta) / NULLIF(:dbtime,0))) norm_val ,
+                               SUM(elapsed_time_delta) elap , (100 * (SUM(cpu_time_delta) / NULLIF(:tcpu,0))) norm_val ,
-                                                                                                                   SUM(cpu_time_delta) cput ,                                                                                                                                       SUM(executions_delta) EXEC ,                                                                                                                                                             SUM(iowait_delta) iowt
+                                                                                                             SUM(cpu_time_delta) cput ,                                                                                                                                 SUM(executions_delta) EXEC ,                                                                                                                                                       SUM(iowait_delta) uiot
             FROM dba_hist_sqlstat
             WHERE dbid = :dbid
@@ 1,863번째 줄: / 2,097번째 줄: @@
                 AND snap_id <= :eid
             GROUP BY sql_id
-            ORDER BY nvl(SUM(elapsed_time_delta), -1) DESC, sql_id))
+            ORDER BY nvl(SUM(cpu_time_delta), -1) DESC, sql_id))
   WHERE rnum < :tsql_max
       AND (rnum <= :tsql_min
@@ 1,871번째 줄: / 2,105번째 줄: @@
 <source lang="sql">
-SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.elap/1000000), to_number(NULL)),
+SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.cput/1000000), to_number(NULL)),
                              sqt.EXEC,
-                                 decode(sqt.EXEC, 0, to_number(NULL),(sqt.elap / sqt.EXEC / 1000000)),
+                                 decode(sqt.EXEC, 0, to_number(NULL), (sqt.cput / sqt.EXEC /1000000)),
                                  sqt.norm_val,
+                                nvl((sqt.elap/1000000), to_number(NULL)),
                                  decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))),
-                                 decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.iowt / sqt.elap))),
+                                 decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.uiot / sqt.elap))),
                                  sqt.sql_id,
                                  to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
@@ 1,888번째 줄: / 2,123번째 줄: @@
 </source>
-=== AWR SQL ordered BY CPU Time ===
+== AWR SQL ordered BY USER I/O Wait Time ==
 <source lang="sql">
 WITH sqt AS
@@ 1,912번째 줄: / 2,146번째 줄: @@
            (SELECT sql_id ,
                    MAX(module) module ,
-                               SUM(elapsed_time_delta) elap , (100 * (SUM(cpu_time_delta) / NULLIF(:tcpu,0))) norm_val ,
+                               SUM(elapsed_time_delta) elap , (100 * (SUM(iowait_delta) / NULLIF(:uiowt,0))) norm_val ,                                                                                                            SUM(cpu_time_delta) cput ,                                                                                                                                SUM(executions_delta) EXEC ,                                                                                                                                                      SUM(iowait_delta) uiot
-                                                                                                             SUM(cpu_time_delta) cput ,                                                                                                                                 SUM(executions_delta) EXEC ,                                                                                                                                                       SUM(iowait_delta) uiot
             FROM dba_hist_sqlstat
             WHERE dbid = :dbid
@@ 1,920번째 줄: / 2,153번째 줄: @@
                 AND snap_id <= :eid
             GROUP BY sql_id
-            ORDER BY nvl(SUM(cpu_time_delta), -1) DESC, sql_id))
+            ORDER BY nvl(SUM(iowait_delta), -1) DESC, sql_id))
   WHERE rnum < :tsql_max
       AND (rnum <= :tsql_min
@@ 1,928번째 줄: / 2,161번째 줄: @@
 <source lang="sql">
-SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.cput/1000000), to_number(NULL)),
+SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.uiot/1000000), to_number(NULL)),
                              sqt.EXEC,
-                                 decode(sqt.EXEC, 0, to_number(NULL), (sqt.cput / sqt.EXEC /1000000)),
+                                 decode(sqt.EXEC, 0, to_number(NULL), (sqt.uiot / sqt.EXEC /1000000)),
                                  sqt.norm_val,
                                  nvl((sqt.elap/1000000), to_number(NULL)),
@@ 1,946번째 줄: / 2,179번째 줄: @@
 </source>
-=== AWR SQL ordered BY USER I/O Wait Time ===
+== AWR SQL ordered BY Gets ==
 <source lang="sql">
 WITH sqt AS
@@ 1,953번째 줄: / 2,187번째 줄: @@
          EXEC,
          uiot,
+        bget,
          norm_val,
          sql_id,
@@ 1,965번째 줄: / 2,200번째 줄: @@
               EXEC,
               uiot,
+             bget,
               rownum rnum
        FROM
            (SELECT sql_id ,
                    MAX(module) module ,
-                               SUM(elapsed_time_delta) elap , (100 * (SUM(iowait_delta) / NULLIF(:uiowt,0))) norm_val ,
+                               SUM(elapsed_time_delta) elap , (100 * (SUM(buffer_gets_delta) / NULLIF(:slr,0))) norm_val ,                                                                                                               SUM(cpu_time_delta) cput ,                                                                                                                                   SUM(executions_delta) EXEC ,                                                                                                                                                         SUM(iowait_delta) uiot ,                                                                                                                                                                           SUM(buffer_gets_delta) bget
-                                                                                                            SUM(cpu_time_delta) cput ,                                                                                                                                SUM(executions_delta) EXEC ,                                                                                                                                                      SUM(iowait_delta) uiot
             FROM dba_hist_sqlstat
             WHERE dbid = :dbid
@@ 1,977번째 줄: / 2,212번째 줄: @@
                 AND snap_id <= :eid
             GROUP BY sql_id
-            ORDER BY nvl(SUM(iowait_delta), -1) DESC, sql_id))
+            ORDER BY nvl(SUM(buffer_gets_delta), -1) DESC, sql_id))
   WHERE rnum < :tsql_max
       AND (rnum <= :tsql_min
@@ 1,985번째 줄: / 2,220번째 줄: @@
 <source lang="sql">
-SELECT /*+ NO_MERGE(sqt) */ nvl((sqt.uiot/1000000), to_number(NULL)),
+SELECT /*+ NO_MERGE(sqt) */ sqt.bget,
                              sqt.EXEC,
-                                 decode(sqt.EXEC, 0, to_number(NULL), (sqt.uiot / sqt.EXEC /1000000)),
+                                 decode(sqt.EXEC, 0, to_number(NULL), (sqt.bget / sqt.EXEC)),
                                  sqt.norm_val,
                                  nvl((sqt.elap/1000000), to_number(NULL)),
-                                 decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))),
+                                 decode(sqt.elap, 0, '     ', lpad(to_char(round((100 * (sqt.cput / sqt.elap)),1), 'TM9'),5)),
-                                 decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.uiot / sqt.elap))),
+                                 decode(sqt.elap, 0, '     ', lpad(to_char(round((100 * (sqt.uiot / sqt.elap)),1), 'TM9'),5)),
                                  sqt.sql_id,
                                  to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
@@ 2,003번째 줄: / 2,238번째 줄: @@
 </source>
-=== AWR SQL ordered BY Gets ===
+== AWR SQL ordered BY READS ==
 <source lang="sql">
@@ 2,011번째 줄: / 2,246번째 줄: @@
          EXEC,
          uiot,
-         bget,
+         dskr,
          norm_val,
          sql_id,
@@ 2,024번째 줄: / 2,259번째 줄: @@
               EXEC,
               uiot,
-              bget,
+              dskr,
               rownum rnum
        FROM
            (SELECT sql_id ,
                    MAX(module) module ,
-                               SUM(elapsed_time_delta) elap , (100 * (SUM(buffer_gets_delta) / NULLIF(:slr,0))) norm_val ,
+                               SUM(elapsed_time_delta) elap , (100 * (SUM(disk_reads_delta) / NULLIF(:phyr,0))) norm_val ,                                                                                                               SUM(cpu_time_delta) cput ,                                                                                                                                   SUM(executions_delta) EXEC ,                                                                                                                                                         SUM(iowait_delta) uiot ,                                                                                                                                                                           SUM(disk_reads_delta) dskr
-                                                                                                               SUM(cpu_time_delta) cput ,                                                                                                                                   SUM(executions_delta) EXEC ,                                                                                                                                                         SUM(iowait_delta) uiot ,                                                                                                                                                                           SUM(buffer_gets_delta) bget
             FROM dba_hist_sqlstat
             WHERE dbid = :dbid
@@ 2,037번째 줄: / 2,271번째 줄: @@
                 AND snap_id <= :eid
             GROUP BY sql_id
-            ORDER BY nvl(SUM(buffer_gets_delta), -1) DESC, sql_id))
+            ORDER BY nvl(SUM(disk_reads_delta), -1) DESC, sql_id))
   WHERE rnum < :tsql_max
       AND (rnum <= :tsql_min
@@ 2,045번째 줄: / 2,279번째 줄: @@
 <source lang="sql">
-SELECT /*+ NO_MERGE(sqt) */ sqt.bget,
+SELECT /*+ NO_MERGE(sqt) */ sqt.dskr,
                              sqt.EXEC,
-                                 decode(sqt.EXEC, 0, to_number(NULL), (sqt.bget / sqt.EXEC)),
+                                 decode(sqt.EXEC, 0, to_number(NULL), (sqt.dskr / sqt.EXEC)),
                                  sqt.norm_val,
-                                 nvl((sqt.elap/1000000), to_number(NULL)),
+                                 nvl((sqt.elap / 1000000), to_number(NULL)),
-                                 decode(sqt.elap, 0, '     ', lpad(to_char(round((100 * (sqt.cput / sqt.elap)),1), 'TM9'),5)),
+                                 decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))),
-                                 decode(sqt.elap, 0, '     ', lpad(to_char(round((100 * (sqt.uiot / sqt.elap)),1), 'TM9'),5)),
+                                 decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.uiot / sqt.elap))),
                                  sqt.sql_id,
                                  to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
@@ 2,063번째 줄: / 2,297번째 줄: @@
 </source>
-=== AWR SQL ordered BY READS ===
+== AWR SQL ordered BY READS ==
 <source lang="sql">
 WITH sqt AS
-(SELECT elap,
+(SELECT EXEC,
-         cput,
+         pdr,
-        EXEC,
+         prq,
-         uiot,
+         porq,
-         dskr,
          norm_val,
          sql_id,
@@ 2,079번째 줄: / 2,312번째 줄: @@
       (SELECT sql_id,
               module,
-             elap,
               norm_val,
-             cput,
               EXEC,
-              uiot,
+              pdr,
-              dskr,
+              prq,
+             porq,
               rownum rnum
        FROM
            (SELECT sql_id ,
-                   MAX(module) module ,
+                   MAX(module) module , (100 * ((SUM(physical_read_requests_delta) - SUM(optimized_physical_reads_delta)) / NULLIF(:phydrq,0))) norm_val ,                                                                                                                                               SUM(executions_delta) EXEC ,                                                                                                                                                                     SUM(physical_read_requests_delta) - SUM(optimized_physical_reads_delta) pdr ,
-                              SUM(elapsed_time_delta) elap , (100 * (SUM(disk_reads_delta) / NULLIF(:phyr,0))) norm_val ,
+SUM(physical_read_requests_delta) prq ,
-                                                                                                               SUM(cpu_time_delta) cput ,                                                                                                                                   SUM(executions_delta) EXEC ,                                                                                                                                                         SUM(iowait_delta) uiot ,                                                                                                                                                                           SUM(disk_reads_delta) dskr
+                                  SUM(optimized_physical_reads_delta) porq
             FROM dba_hist_sqlstat
-           WHERE dbid = :dbid
+WHERE dbid = :dbid
-               AND instance_number = :inst_num
+AND instance_number = :inst_num
-               AND :bid < snap_id
+AND :bid < snap_id
-               AND snap_id <= :eid
+AND snap_id <= :eid
             GROUP BY sql_id
-            ORDER BY nvl(SUM(disk_reads_delta), -1) DESC, sql_id))
+            ORDER BY nvl(SUM(physical_read_requests_delta) - SUM(optimized_physical_reads_delta), -1) DESC, sql_id))
- WHERE rnum < :tsql_max
+WHERE rnum < :tsql_max
-     AND (rnum <= :tsql_min
+AND (rnum <= :tsql_min
-          OR norm_val > :top_pct_sql))
+ OR norm_val > :top_pct_sql))
 ;
 </source>
 <source lang="sql">
-SELECT /*+ NO_MERGE(sqt) */ sqt.dskr,
+SELECT /*+ NO_MERGE(sqt) */ sqt.pdr,
-                            sqt.EXEC,
+                sqt.prq,
-                                decode(sqt.EXEC, 0, to_number(NULL), (sqt.dskr / sqt.EXEC)),
+                sqt.EXEC,
-                                sqt.norm_val,
+        decode(sqt.EXEC, 0, to_number(NULL), (sqt.prq/ sqt.EXEC)),
-                                nvl((sqt.elap / 1000000), to_number(NULL)),
+        decode(sqt.prq, 0, to_number(NULL), (100 * (sqt.porq / sqt.prq))),
-                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.cput / sqt.elap))),
+        sqt.norm_val,
-                                decode(sqt.elap, 0, to_number(NULL), (100 * (sqt.uiot / sqt.elap))),
+        sqt.sql_id,
-                                sqt.sql_id,
+        to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
-                                to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
+        nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
-                                nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
 FROM sqt,
-     dba_hist_sqltext st
+dba_hist_sqltext st
 WHERE st.sql_id(+) = sqt.sql_id
 AND st.dbid(+) = :dbid
@@ 2,122번째 줄: / 2,353번째 줄: @@
 ;
 </source>
-=== AWR SQL ordered BY READS ===
+== AWR SQL ordered BY Sharable Memory ==
 <source lang="sql">
 WITH sqt AS
 (SELECT EXEC,
-        pdr,
+sharable_mem,
-        prq,
+sql_id,
-        porq,
+module,
-        norm_val,
+rnum
-        sql_id,
+FROM
-        module,
+(SELECT sql_id,
-        rnum
+    module,
- FROM
+    EXEC,
-     (SELECT sql_id,
+    sharable_mem,
-             module,
+    rownum rnum
-             norm_val,
+FROM
-             EXEC,
+ (SELECT sql_id,
-             pdr,
+         module,
-             prq,
+         EXEC,
-             porq,
+         sharable_mem
-             rownum rnum
+  FROM
-      FROM
+      (SELECT sharable_mem,
-          (SELECT sql_id ,
+              sql_id
-                  MAX(module) module , (100 * ((SUM(physical_read_requests_delta) - SUM(optimized_physical_reads_delta)) / NULLIF(:phydrq,0))) norm_val ,                                                                                                                                               SUM(executions_delta) EXEC ,                                                                                                                                                                     SUM(physical_read_requests_delta) - SUM(optimized_physical_reads_delta) pdr ,
+       FROM dba_hist_sqlstat
-SUM(physical_read_requests_delta) prq ,
+       WHERE dbid = :dbid
-                                  SUM(optimized_physical_reads_delta) porq
+           AND snap_id = :eid
-           FROM dba_hist_sqlstat
+           AND instance_number = :inst_num
-WHERE dbid = :dbid
+           AND sharable_mem > :shmem_thres) y
-AND instance_number = :inst_num
+  LEFT OUTER JOIN
-AND :bid < snap_id
+      (SELECT sql_id ,
-AND snap_id <= :eid
+              MAX(module) module ,
-           GROUP BY sql_id
+                          SUM(executions_delta) EXEC
-           ORDER BY nvl(SUM(physical_read_requests_delta) - SUM(optimized_physical_reads_delta), -1) DESC, sql_id))
+       FROM dba_hist_sqlstat
-WHERE rnum < :tsql_max
+       WHERE dbid = :dbid
-AND (rnum <= :tsql_min
+           AND instance_number = :inst_num
- OR norm_val > :top_pct_sql))
+           AND :bid < snap_id
+           AND snap_id <= :eid
+       GROUP BY sql_id) x USING (sql_id)
+  ORDER BY nvl(sharable_mem, -1) DESC, sql_id))
+WHERE rnum <= :tsql_max)
 ;
 </source>
 <source lang="sql">
-SELECT /*+ NO_MERGE(sqt) */ sqt.pdr,
+SELECT /*+ NO_MERGE(sqt) */ sqt.sharable_mem,
-                sqt.prq,
                  sqt.EXEC,
-         decode(sqt.EXEC, 0, to_number(NULL), (sqt.prq/ sqt.EXEC)),
+         decode(:espm, 0, 0, 100 * sqt.sharable_mem/:espm),
-        decode(sqt.prq, 0, to_number(NULL), (100 * (sqt.porq / sqt.prq))),
-        sqt.norm_val,
          sqt.sql_id,
          to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
@@ 2,179번째 줄: / 2,411번째 줄: @@
 </source>
-=== AWR SQL ordered BY Sharable Memory ===
+== AWR SQL ordered BY VERSION COUNT ==
 <source lang="sql">
 WITH sqt AS
 (SELECT EXEC,
-sharable_mem,
+version_count,
 sql_id,
 module,
@@ 2,191번째 줄: / 2,423번째 줄: @@
      module,
      EXEC,
-     sharable_mem,
+     version_count,
      rownum rnum
 FROM
@@ 2,197번째 줄: / 2,429번째 줄: @@
           module,
           EXEC,
-          sharable_mem
+          version_count
    FROM
-       (SELECT sharable_mem,
+       (SELECT version_count,
                sql_id
         FROM dba_hist_sqlstat
@@ 2,205번째 줄: / 2,437번째 줄: @@
             AND snap_id = :eid
             AND instance_number = :inst_num
-            AND sharable_mem > :shmem_thres) y
+            AND version_count > :vcnt_thres) y
    LEFT OUTER JOIN
        (SELECT sql_id ,
@@ 2,216번째 줄: / 2,448번째 줄: @@
             AND snap_id <= :eid
         GROUP BY sql_id) x USING (sql_id)
-   ORDER BY nvl(sharable_mem, -1) DESC, sql_id))
+   ORDER BY nvl(y.version_count, -1) DESC, sql_id))
 WHERE rnum <= :tsql_max)
 ;
@@ 2,222번째 줄: / 2,454번째 줄: @@
 <source lang="sql">
-SELECT /*+ NO_MERGE(sqt) */ sqt.sharable_mem,
+SELECT /*+ NO_MERGE(sqt) */ sqt.version_count,
                  sqt.EXEC,
-        decode(:espm, 0, 0, 100 * sqt.sharable_mem/:espm),
          sqt.sql_id,
          to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
@@ 2,232번째 줄: / 2,463번째 줄: @@
 WHERE st.sql_id(+) = sqt.sql_id
 AND st.dbid(+) = :dbid
 ORDER BY sqt.rnum
-;
-</source>
-=== AWR SQL ordered BY VERSION COUNT ===
-<source lang="sql">
-WITH sqt AS
-(SELECT EXEC,
-version_count,
-sql_id,
-module,
-rnum
-FROM
-(SELECT sql_id,
-    module,
-    EXEC,
-    version_count,
-    rownum rnum
-FROM
- (SELECT sql_id,
-         module,
-         EXEC,
-         version_count
-  FROM
-      (SELECT version_count,
-              sql_id
-       FROM dba_hist_sqlstat
-       WHERE dbid = :dbid
-           AND snap_id = :eid
-           AND instance_number = :inst_num
-           AND version_count > :vcnt_thres) y
-  LEFT OUTER JOIN
-      (SELECT sql_id ,
-              MAX(module) module ,
-                          SUM(executions_delta) EXEC
-       FROM dba_hist_sqlstat
-       WHERE dbid = :dbid
-           AND instance_number = :inst_num
-           AND :bid < snap_id
-           AND snap_id <= :eid
-       GROUP BY sql_id) x USING (sql_id)
-  ORDER BY nvl(y.version_count, -1) DESC, sql_id))
-WHERE rnum <= :tsql_max)
-;
 </source>
-<source lang="sql">
+[[Category:oracle]]
-SELECT /*+ NO_MERGE(sqt) */ sqt.version_count,
-                sqt.EXEC,
-        sqt.sql_id,
-        to_clob(decode(sqt.module, NULL,NULL, 'Module: ' || sqt.module)),
-        nvl(st.sql_text,to_clob('** SQL Text Not Available **'))
-FROM sqt,
-dba_hist_sqltext st
-WHERE st.sql_id(+) = sqt.sql_id
-AND st.dbid(+) = :dbid
-ORDER BY sqt.rnum
-</source>

"ORACLE 모니터링"의 두 판 사이의 차이

DB CAFE

2024년 1월 1일 (월) 23:41 기준 최신판

목차

1 세션 정보[편집]

1.1 User 별 Session 정보 조회[편집]

1.2 세션 죽이기(SID,SERAIL#)[편집]

1.3 현재 TOP 세션 gv$session[편집]

1.4 오래 걸리는 세션 정보[편집]

1.5 현재 작업중인 세션 정보 V$SESSION_LONGOPS[편집]

1.6 세션당 리소스 사용률 gv$sessmetric[편집]

1.7 대기중인 세션[편집]

1.8 1시간 이상 idle 상태인 세션[편집]

1.9 Active Session 중 Idle Time이 긴 작업[편집]

1.10 사용자 세션 중에서 2시간 이상 idle 상태가 지속되는 세션 kill[편집]

1.11 트랜젝션 중인 세션[편집]

1.12 연결되어 있는 OS 사용자 및 프로그램 조회[편집]

1.13 Session별 사용 명령어[편집]

2 SQL 찾기[편집]

2.1 SQL 통계 정보확인하기[편집]

2.2 Module LEVEL 통계[편집]

2.3 SCHEMA LEVEL 통계[편집]

2.4 DB Time 분석하기[편집]

2.5 SQL 추출하기 (10g)[편집]

2.6 엑사 DB[편집]

2.7 Exa . Smart Scan으로 수행되지 않고,Interconnect Bytes가 높은 SQL[편집]

2.8 Smart Scan으로 수행되지 않고,Offload Returned Bytes가 높은 SQL[편집]

2.9 Full Table Scan 처리가 있는 SQL문 정보 추출하기(10g)[편집]

2.10 하드 파싱(leteral,리터럴) SQL 찾기[편집]

2.11 특정 SQL 수행내역 확인하기[편집]

2.12 SQL_ID를 이용하여 SQL TEXT 추출하기[편집]

2.13 SQL_ID를 이용하여 SQL문의 수행내역 추출하기[편집]

2.14 프로세스 ID 를 이용하여 SQL 찾기[편집]

2.15 바인드 변수 값[편집]

2.16 바인드 변수 값2[편집]

2.17 바인드 변수 값3[편집]

2.18 플랜 정보 조회[편집]

2.19 SQL_ID를 이용하여 SQL문의 PLAN 추출하기[편집]

3 SQL 과 커서 조회[편집]

3.1 현재 커서 수 확인[편집]

3.2 sid별 열린커셔[편집]

3.3 sql 별 열린 커서[편집]

4 락 모니터링[편집]

4.1 V$LOCK 을 사용한 잠금 경합 모니터링[편집]

4.2 락이 걸린 세션 간단히 보기[편집]

4.3 락이 걸린 세션 상세 보기[편집]

4.4 1일 이상 존재 하는 조회 / kill / AWS_KILL_CMD[편집]

5 프로세스 정보[편집]

5.1 Oracle Process의 정보[편집]

5.2 오브젝트에 접속되어 있는 프로그램 조회[편집]

6 부하 발생 모니터링[편집]

6.1 cpu를 많이 사용하는 쿼리문과 프로세스아이디,시리얼번호,머신 알아내기[편집]

6.2 롤백 세그먼트 경합 조회[편집]

6.3 버퍼 캐시 히트율(Buffer Cache Hit Ratio)[편집]

6.4 라이브러리 캐시 히트율(Library Cache Hit Ratio)[편집]

6.5 데이터 딕셔너리 캐시 히트율(Data Dictionary Cache Hit Ratio)[편집]

7 I/O 부하 찾기[편집]

7.1 인스턴스 기동 이후 I/O 관련 대기현상 누적 값[편집]

7.2 전체 IO량(100%) 대비 사용자별 IO량[편집]

8 테이블스페이스[편집]

8.1 테이블스페이스 사용량[편집]

8.2 System 테이블스페이스에 비시스템 세그먼트 조회[편집]

9 CPU/메모리/사양[편집]

9.1 오라클서버의 메모리[편집]

9.2 총 CPU Time 대비 SQL Parsing Time[편집]

9.3 총 Parsing Time 대비 Hard Parsing Time[편집]

9.4 Log file sync 관련 대기현상 Sync Writes Time[편집]

10 INDEX 정보 추출하기[편집]

10.1 Function-based Index 컬럼 찾기[편집]

10.2 중복 인덱스 찾기(불필요 인덱스)[편집]

10.3 CONSTRAINTS 확인하기[편집]

10.4 해당 테이블을 참조하는 테이블(자식) Constraints 확인하기[편집]

10.5 v$segment_statistics 활용하기[편집]

10.6 Parameter 확인하기[편집]

10.7 Bind Peeked 확인하기[편집]

10.8 JOB 확인하기[편집]

10.9 DB Link SQL CURSOR 확인하기[편집]

10.10 INDEX Deleted ROWS CHECK [1][편집]

11 ASH를 이용한 모니터링[편집]

11.1 특정 Session이 가장 많이 수행 된 SQL과 수행 점유율(수행 횟수)[편집]