技術標籤：oracle

產生原因

官方定義：

This wait happens when a session wants to access a database block in the buffer cache but it cannot as the buffer is “busy”. The two main cases where this can occur are:

1. Another session is reading the block into the buffer
2. Another session holds the buffer in an incompatible mode to our request

模擬等待

建立測試表並插入資料：

SQL> create table tb1 (id int ,name varchar2(10));

Table created.

SQL> insert into tb1 values (1,'scott');

1 row created.

SQL> insert into tb1 values (2,'tom');

1 row created.

SQL> commit;

Commit complete.

會話1修改資料：

SQL> select userenv('sid') from dual;

USERENV(
 
'SID')
--------------
             1
             
SQL> begin
for i in 1..100000 loop
update tb1 set name='rose' where id=2;
commit;
end loop;
end;
/

會話2修改資料：

SQL> select userenv('sid') from dual;

USERENV('SID')
--------------
            27

SQL> begin
for i in 1..100000 loop
update tb1 set name='john'
 
 where id=1;
commit;
end loop;
end;
/

在會話3檢視等待情況：

--定位到造成等待的SQL
SQL> SELECT g.inst_id,g.sid,g.serial#,g.event,g.username,g.sql_hash_value,s.sql_fulltext
FROM gv$session g,v$sql s
WHERE g.sql_hash_value = s.hash_value and username='HR' and event='buffer busy waits';
   INST_ID        SID    SERIAL# EVENT                          USERNAME   SQL_HASH_VALUE SQL_FULLTEXT
---------- ---------- ---------- ------------------------------ ---------- -------------- --------------------------------------------------
         1          1          7 buffer busy waits              HR              401484711 UPDATE TB1 SET NAME='rose' WHERE ID=2
         1         27         49 buffer busy waits              HR             2040921087 UPDATE TB1 SET NAME='john' WHERE ID=1

--定位到熱點快
SQL> select event,sid,p1,p2,p3 from v$session_wait_history where event='buffer busy waits'
EVENT                       SID         P1         P2         P3
-------------------- ---------- ---------- ---------- ----------
buffer busy waits             1          4       5471          1
buffer busy waits             1          4       5471          1
buffer busy waits             1          4       5471          1
buffer busy waits             1          4       5471          1
buffer busy waits             1          4       5471          1
buffer busy waits            27          4       5471          1
buffer busy waits            27          4       5471          1
buffer busy waits            27          4       5471          1
buffer busy waits            27          4       5471          1
buffer busy waits            27          4       5471          1

10 rows selected.

--P1=file#
--P2=block#

--定位到塊所屬的段
SQL> SELECT OWNER, SEGMENT_NAME, SEGMENT_TYPE, TABLESPACE_NAME, A.PARTITION_NAME FROM DBA_EXTENTS A WHERE FILE_ID = 4 AND 5471 BETWEEN BLOCK_ID AND BLOCK_ID + BLOCKS - 1;

OWNER      SEGMENT_NAME         SEGMENT_TYPE       TABLESPACE_NAME                PARTITION_NAME
---------- -------------------- ------------------ ------------------------------ ------------------------------
HR         TB1                  TABLE              USERS

解決方法

As buffer busy waits are due to contention for particular blocks then you cannot take any action until you know which blocks are being competed for and why. Eliminating the cause of the contention is the best option. Note that “buffer busy waits” for data blocks are often due to several processes repeatedly reading the same blocks (eg: if lots of people scan the same index) - the first session processes the blocks that are in the buffer cache quickly but then a block has to be read from disk - the other sessions (scanning the same index) quickly ‘catch up’ and want the block which is currently being read from disk - they wait for the buffer as someone is already reading the block in.

The following hints may be useful for particular types of contention - these are things that MAY reduce contention for particular situations:

參考： WAITEVENT: “buffer busy waits” Reference Note (Doc ID 34405.1)

歡迎關注我的公眾號，一起學習。