Adaptive Cursor Sharing(第五篇)
ACS最佳實踐
從上面的介紹我們可以看出,ACS雖然可以解決繫結變數窺探引起的問題,但是它也存在著一些缺點:
1)一旦一個SQL被標註為繫結敏感,最佳化器就要監視SQL語句處理的行數,一旦行數發生“鉅變”,就要更新v$sql_cs_histogram檢視。但是這代價似乎並不大,因為v$sql_cs_histogram檢視只在SQL執行完成後才會被更新,不佔用解析時間,因此這個更新完全可以是非同步的。
2)Cursor被標註為bind aware後,在解析階段要窺探變數的值,計算謂詞的選擇率,計算後的值要與對應的v$sql_cs_selectivity檢視中的值做比較,看是否已經在已存的選擇率範圍內,如果不在會發生硬解析。由於要窺探變數的值,計算選擇率,一定程度上加大瞭解析的時間。這裡的解析,指的是軟解析。但是這個影響可能也沒有那麼大,因為一個資料庫中並不是所有的SQL都會被標註為bind aware,只有操作的資料集發生過巨大變化的SQL才會被標註為bind aware。
3)使用ACS存在一個不穩定期,也就是SQL執行計劃要經歷先變糟再變好的過程。v$sql_cs_histogram檢視記錄了SQL處理的資料量,用3個bucket來表示,如果SQL處理的行數發生鉅變,也就是說處理的行數散落在了至少2個桶內,下一次解析時,就要窺探繫結變數的值,重新硬解析生成執行計劃。
4)一旦SQL被刷出共享池,這個SQL還需要重複經歷ACS不穩定期到穩定器的過程。
5)PL/SQL中存在bug如果不調整session_cached_cursors引數為0將不能使用到ACS特性。
在我所負責管理的生產環境下,ACS都是關閉的,雖然我本人也對ACS做過一些研究和測試,但是始終還是有點敬而遠之。那有沒有什麼辦法既能使用到ACS的特性,又能一定程度避免這些缺點呢?首先我們需要介紹一個hint-bind_aware。
bind_aware的用法和作用
使用了ACS的功能後,一個遊標從bind sensitive 到 bind aware,中間有著不穩定期,如果在遊標中使用bind_aware hint後,將會使遊標的狀態直接進入bind aware,而不會經歷bind sensitive狀態。我們來透過一個例子看一看:
SQL>CREATE TABLE test 2 AS 3 SELECT ROWNUM id, 4 DBMS_RANDOM.STRING('A', 12) name, 5 DECODE(MOD(ROWNUM, 500), 0, 'Inactive', 'Active') status 6 FROM all_objects 7 WHERE ROWNUM <= 50000;
Table created.
SQL>CREATE INDEX test_id_ind ON test(status);
Index created.
SQL>begin 2 dbms_stats.gather_table_stats(user, 3 'test', 4 method_opt => 'for columns status size 254', 5 cascade => true); 6 end; 7 /
PL/SQL procedure successfully completed.
SQL>SELECT COUNT (*) cnt, status 2 FROM test 3 GROUP BY status
4 / ---------- ---------------- 49900 Active
100 Inactive |
上面的程式碼本章已經出現過幾次,主要作用是:建立了一張表,表上有一列STATUS有資料傾斜,列上建立了索引,並在這列上收集直方圖。我們來看看在對SQL增加bind aware的hint後,ACS的表現會是什麼樣。
我們首先查詢STATUS為Inactive的情況,這個值在表裡佔少數。
SQL>alter system flush shared_pool;
System altered.
SQL>var a varchar2(100) SQL>exec :a :='Inactive';
PL/SQL procedure successfully completed.
SQL> SQL>select /*+ bind_aware */ count(name) from test where status=:a;
COUNT(NAME) ----------- 100
SQL>select * from table(dbms_xplan.display_cursor(null,null));
PLAN_TABLE_OUTPUT ------------------------------------------------------------------------------------------- SQL_ID a5fy4g63j8vzr, child number 0 ------------------------------------- select /*+ bind_aware */ count(name) from test where status=:a
Plan hash value: 2948918962
-------------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | -------------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | | | 2 (100)| | | 1 | SORT AGGREGATE | | 1 | 25 | | | | 2 | TABLE ACCESS BY INDEX ROWID| TEST | 133 | 3325 | 2 (0)| 00:00:01 | |* 3 | INDEX RANGE SCAN | TEST_ID_IND | 133 | | 1 (0)| 00:00:01 | --------------------------------------------------------------------------------------------
Predicate Information (identified by operation id): ---------------------------------------------------
3 - access("STATUS"=:A)
SQL>SELECT child_number, executions, buffer_gets, is_bind_sensitive, 2 is_bind_aware 3 FROM v$sql 4 WHERE sql_id='a5fy4g63j8vzr';
CHILD_NUMBER EXECUTIONS BUFFER_GETS IS IS ------------ ---------- ----------- -- -- 0 1 254 Y Y |
從v$sql的is_bind_aware輸出為Y可以看出,SQL僅執行了一次就已經被標註為bind aware,沒有經歷不穩定期。我們再看下執行STATUS為Active時的表現:
SQL>exec :a :='Active'
PL/SQL procedure successfully completed.
SQL>select /*+ bind_aware */ count(name) from test where status=:a;
COUNT(NAME) ----------- 49900
SQL>select * from table(dbms_xplan.display_cursor(null,null));
PLAN_TABLE_OUTPUT ------------------------------------------------------------------------------- SQL_ID a5fy4g63j8vzr, child number 1 ------------------------------------- select /*+ bind_aware */ count(name) from test where status=:a
Plan hash value: 1950795681
--------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | --------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | | | 51 (100)| | | 1 | SORT AGGREGATE | | 1 | 25 | | | |* 2 | TABLE ACCESS FULL| TEST | 49862 | 1217K| 51 (2)| 00:00:01 | ---------------------------------------------------------------------------
Predicate Information (identified by operation id): ---------------------------------------------------
2 - filter("STATUS"=:A)
19 rows selected.
SQL>SELECT child_number, executions, buffer_gets, is_bind_sensitive, 2 is_bind_aware 3 FROM v$sql 4 WHERE sql_id='a5fy4g63j8vzr';
CHILD_NUMBER EXECUTIONS BUFFER_GETS IS IS ------------ ---------- ----------- -- -- 0 1 1486 Y Y 1 1 210 Y Y |
非常棒!我們更換繫結變數的值為Active後,第一次執行就已經得到了正確的執行計劃,在v$sql中也已經新增了一個entry,用來記錄新產生的遊標的執行計劃。因此使用bind aware這個hint後,遊標將不會經歷不穩定期,SQL每次解析的時候都要窺探繫結變數的值,然後計算選擇率,如果計算選擇率與現有的遊標的選擇率不符,就會基於窺探到的繫結變數的值硬解析重新產生了一個新的遊標。如果你確認一個SQL需要使用ACS功能,但是又不想讓它經歷不穩定期,那麼你可以透過bind aware這個hint做到這一點。還有著一些手段可以嘗試,例如我們可以關閉ACS的功能,對有需要的SQL單獨開啟ACS的功能。可能是從阿里做DBA沿襲來的習慣,喜歡直接關閉繫結變數窺探,繫結變數窺探被關閉後,ACS也就自動關閉了。然後對有需要使用ACS的SQL,透過增加hint,OPT_PARAM('_optim_peek_user_binds' 'true') bind_aware來使用到ACS的特性,OPT_PARAM('_optim_peek_user_binds' 'true')用來在SQL語句級別開啟繫結變數窺探的功能。繫結變數窺探和ACS關閉後,就規避了上面提到的ACS的缺點。DBA可以有選擇性的對某些SQL使用ACS。當然這個對DBA要求較高,需要了解應用,瞭解表的資料分佈特點,瞭解表上的SQL的查詢特點。有些大公司已經配備了應用DBA的角色,負責開發的SQL REVIEW等工作,可以在SQL REVIEW階段裡DBA透過了解應用的SQL,對有需要的SQL增加ACS功能。如果不能第一時間增加hint進去,也可以透過sql profile,sql patch的方式在不修改SQL語句的情況下增加這些hint繫結到SQL語句上去。(SPM baseline無效在這裡)。
如我們可以透過sql profile來對一個SQL增加ACS的功能:
SQL>show parameter binds
NAME TYPE VALUE ------------------------------------ ---------------------- ------------------- _optim_peek_user_binds boolean FALSE
SQL>exec :a :='Active'
PL/SQL procedure successfully completed.
SQL>select count(name) from test where status=:a;
COUNT(NAME) ----------- 49900
SQL>select * from table(dbms_xplan.display_cursor(null,null));
PLAN_TABLE_OUTPUT ------------------------------------------------------------------------------- SQL_ID 7yjf9wt1rt8a6, child number 0 ------------------------------------- select count(name) from test where status=:a
Plan hash value: 1950795681
--------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | --------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | | | 51 (100)| | | 1 | SORT AGGREGATE | | 1 | 25 | | | |* 2 | TABLE ACCESS FULL| TEST | 25000 | 610K| 51 (2)| 00:00:01 | ---------------------------------------------------------------------------
Predicate Information (identified by operation id): ---------------------------------------------------
2 - filter("STATUS"=:A)
SQL>SELECT child_number, executions, buffer_gets, is_bind_sensitive, 2 is_bind_aware 3 FROM v$sql 4 WHERE sql_id='7yjf9wt1rt8a6';
CHILD_NUMBER EXECUTIONS BUFFER_GETS IS IS ------------ ---------- ----------- -- -- 0 2 540 N N |
關閉繫結變數窺探後,也就關閉了ACS的功能,SQL的bind sensitive語句被標註為N。我們透過sql profile增加hint看看。
SQL>@profile Enter value for sql_id: 7yjf9wt1rt8a6
PLAN_TABLE_OUTPUT ------------------------------------------------------------------------------- SQL_ID 7yjf9wt1rt8a6, child number 0 ------------------------------------- select count(name) from test where status=:a
Plan hash value: 1950795681
--------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | --------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | | | 51 (100)| | | 1 | SORT AGGREGATE | | 1 | 25 | | | |* 2 | TABLE ACCESS FULL| TEST | 25000 | 610K| 51 (2)| 00:00:01 | ---------------------------------------------------------------------------
Outline Data -------------
/*+ BEGIN_OUTLINE_DATA IGNORE_OPTIM_EMBEDDED_HINTS OPTIMIZER_FEATURES_ENABLE('11.2.0.3') DB_VERSION('11.2.0.3') OPT_PARAM('_optim_peek_user_binds' 'false') OPT_PARAM('_optimizer_skip_scan_enabled' 'false') ALL_ROWS OUTLINE_LEAF(@"SEL$1") FULL(@"SEL$1" "TEST"@"SEL$1") END_OUTLINE_DATA */
Predicate Information (identified by operation id): ---------------------------------------------------
2 - filter("STATUS"=:A)
35 rows selected.
Enter value for hint_text: OPT_PARAM('_optim_peek_user_binds' 'true') bind_aware
Profile profile_7yjf9wt1rt8a6_dwrose created.
SQL>select count(name) from test where status=:a;
COUNT(NAME) ----------- 49900
1 row selected.
SQL>SELECT child_number, executions, buffer_gets, is_bind_sensitive, 2 is_bind_aware 3 FROM v$sql 4 WHERE sql_id='7yjf9wt1rt8a6';
CHILD_NUMBER EXECUTIONS BUFFER_GETS IS IS ------------ ---------- ----------- -- -- 0 1 270 Y Y
1 row selected.
SQL>select * from table(dbms_xplan.display_cursor('7yjf9wt1rt8a6',null));
PLAN_TABLE_OUTPUT ------------------------------------------------------------------------------- SQL_ID 7yjf9wt1rt8a6, child number 0 ------------------------------------- select count(name) from test where status=:a
Plan hash value: 1950795681
--------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | --------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | | | 51 (100)| | | 1 | SORT AGGREGATE | | 1 | 25 | | | |* 2 | TABLE ACCESS FULL| TEST | 49862 | 1217K| 51 (2)| 00:00:01 | ---------------------------------------------------------------------------
Predicate Information (identified by operation id): ---------------------------------------------------
2 - filter("STATUS"=:A)
Note ----- - SQL profile profile_7yjf9wt1rt8a6_dwrose used for this statement
23 rows selected.
SQL>exec :a :='Inactive'
PL/SQL procedure successfully completed.
SQL>select count(name) from test where status=:a;
COUNT(NAME) ----------- 100
1 row selected.
SQL>select * from table(dbms_xplan.display_cursor(null,null));
PLAN_TABLE_OUTPUT ------------------------------------------------------------------------------- SQL_ID 7yjf9wt1rt8a6, child number 5 ------------------------------------- select count(name) from test where status=:a
Plan hash value: 2948918962
-------------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | -------------------------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | | | 2 (100)| | | 1 | SORT AGGREGATE | | 1 | 25 | | | | 2 | TABLE ACCESS BY INDEX ROWID| TEST | 133 | 3325 | 2 (0)| 00:00:01 | |* 3 | INDEX RANGE SCAN | TEST_ID_IND | 133 | | 1 (0)| 00:00:01 | --------------------------------------------------------------------------------------------
Predicate Information (identified by operation id): ---------------------------------------------------
3 - access("STATUS"=:A)
Note ----- - SQL profile profile_7yjf9wt1rt8a6_dwrose used for this statement
24 rows selected. |
但是可惜的是,11.2.0.3版本存在BUG(其他版本沒做測試),在session或system級關閉繫結變數窺探的情況下,如果使用了hint OPT_PARAM('_optim_peek_user_binds' 'true') bind_aware來使用ACS功能,每執行一次SQL,就會在共享池中新生成一個執行計劃,之前產生的計劃被標註為不能共享,不能共享的原因是:user_bind_peek_mismatch。此BUG在12.0.1版本已經被修復。因此如果使用筆者所說的方式,使用前一定要做好測試,防止產生過多的子游標。如果對於有資料傾斜的列,唯一值非常少,可以考慮直接使用文字變數,放棄繫結變數的使用。
SQL>select child_number,user_bind_peek_mismatch from v$sql_shared_cursor where sql_id='7yjf9wt1rt8a6';
CHILD_NUMBER US ------------ -- 0 N 1 Y 2 Y 3 Y 4 Y 5 Y |
如何關閉ACS的特性:
alter system set "_optimizer_extended_cursor_sharing_rel"=none scope=both;
如果你的系統關閉了繫結變數窺探的功能也會自動關閉ACS。
alter system set "_optim_peek_user_binds"=false scope=both;
來自 “ ITPUB部落格 ” ,連結:http://blog.itpub.net/22034023/viewspace-2153800/,如需轉載,請註明出處,否則將追究法律責任。
相關文章
- Adaptive Cursor SharingAPT
- Adaptive Cursor Sharing(第二篇)APT
- Adaptive Cursor Sharing(第三篇)APT
- Adaptive Cursor Sharing(第四篇)APT
- Adaptive Cursor Sharing (第一篇)APT
- Adaptive Cursor Sharing: Worked Example (Doc ID 836256.1)APT
- 11G Adaptive Cursor Sharing(ACS)的研究APT
- 優化器革命之- adaptive cursor sharing (三)優化APT
- 優化器革命之- adaptive cursor sharing (二)優化APT
- 11GR1 新特性 adaptive cursor sharingAPT
- 最佳化器革命之- adaptive cursor sharing (三)APT
- 最佳化器革命之- adaptive cursor sharing (四)APT
- 11G Adaptive Cursor Sharing(ACS)的研究 (轉)APT
- 驗證11g的adaptive cursor sharing功能APT
- Oracle自適應共享遊標——Adaptive Cursor Sharing(上)OracleAPT
- Oracle自適應共享遊標——Adaptive Cursor Sharing(下)OracleAPT
- [20120327]Adaptive Cursor Sharing 的問題APT
- 11gR2 新特性之(一)Adaptive Cursor Sharing(ACS)APT
- 揭秘自適應遊標共享技術(Adaptive Cursor Sharing)APT
- Oracle 11g SQL效能的新特性(一)- Adaptive Cursor SharingOracleSQLAPT
- 揭祕自適應遊標共享技術(Adaptive Cursor Sharing)APT
- [20221227]Adaptive Cursor Sharing & 直方圖.txtAPT直方圖
- 繫結變數之自適應遊標共享(Adaptive Cursor Sharing)變數APT
- 11G Adaptive Cursor Sharing(ACS)自適應遊標測試APT
- 繫結變數的正確使用,bind peeking,Adaptive Cursor Sharing變數APT
- 【DBA】Oracle 11g 針對SQL效能的新特性(一)- Adaptive Cursor SharingOracleSQLAPT
- V$SQL_遊標、adaptive cursor sharing、bind peeking、直方圖的概念SQLAPT直方圖
- oracle11g中SQL最佳化新特性之Adaptive Cursor Sharing (ACS)OracleSQLAPT
- [20120220]Adaptive Cursor Sharing 與hints.txtAPT
- CURSOR_SHARING=SIMILARMILA
- Postgresql的CURSOR SHARINGSQL
- oracle cursor_sharing [轉]Oracle
- 關於 cursor_sharing = similarMILA
- oracle引數-cursor_sharingOracle
- cursor_sharing=similar深度剖析MILA
- cursor_sharing : exact , force , similarMILA
- 關於cursor_sharing=similarMILA
- 關於cursor_sharing = similar(ZT)MILA