首页 > 代码库 > latch: cache buffers chains故障处理总结

latch: cache buffers chains故障处理总结

故障分析思路 

 

查看等待事件,判断故障起因

 

SQL>select * from (select sid,event,p1,p2,p3,p1text,WAIT_TIME,SECONDS_IN_WAIT from v$session_wait where wait_class# <> 6 order by wait_time desc) where rownum <=10;

 

确认为latch: cache buffers chains引起的故障后,查看latch的命中率

 

SQL>SELECT name, gets, misses, sleeps,        immediate_gets, immediate_misses      FROM v$latch    WHERE name = ‘cache buffers chains‘;

 

各列名称意义如下

 

NAME:latch名称 IMMEDIATE_GETS:以Immediate模式latch请求数 IMMEDIATE_MISSES:请求失败数 GETS:以Willing to wait请求模式latch的请求数 MISSES:初次尝试请求不成功次数 SPIN_GETS:第一次尝试失败,但在以后的轮次中成功 SLEEP[x]:成功获取前sleeping次数 WAIT_TIME:花费在等待latch的时间

 

这里需要注意MISSES/GETS如果在达10%左右,则说明有比较严重的latch争用,也可以通过查询v$latch_children视图查看其他latch信息 ,语句如下

 

SQL> SELECT *     FROM (SELECT   addr, child#, gets, misses, sleeps, immediate_gets igets,                   immediate_misses imiss, spin_gets sgets               FROM v$latch_children              WHERE NAME = ‘cache buffers chains‘           ORDER BY sleeps DESC)    WHERE ROWNUM < 11;

关于latch的统计信息,主要关注以下几部分

misses/gets的比率是多少

获自spinning的misses的百分比是多少

latch请求了多少次

latch休眠了多少次

 

查看热点对象和访问信息,TCH列表示对象被访问的次数

 

SQL> SELECT *     FROM (  SELECT addr,                    ts#,                    file#,                    dbarfil,                    dbablk,                    tch               FROM x$bh           ORDER BY tch DESC)    WHERE ROWNUM < 11; 

 

通过对象的文件号和块号查看具体对象信息

 

SQL>select owner, segment_name, partition_name, tablespace_name  from dba_extents where relative_fno = &v_dba_rfile and &v_dba_block between block_id and block_id + blocks - 1;

也可以通过如下sql查找热点块,主要

SELECT *

  FROM (SELECT O.OWNER, O.OBJECT_NAME, O.OBJECT_TYPE, SUM(TCH) TOUCHTIME

          FROM X$BH B, DBA_OBJECTS O

         WHERE B.OBJ = O.DATA_OBJECT_ID

           AND B.TS# > 0

         GROUP BY O.OWNER, O.OBJECT_NAME, O.OBJECT_TYPE

         ORDER BY SUM(TCH) DESC)

 WHERE ROWNUM <= 10;

 

查看引起latch: cache buffers chains的sql

 

SQL> select * from (select      count(*),      sql_id,      nvl(o.object_name,ash.current_obj#) objn,     substr(o.object_type,0,10) otype,      3    4    5    6        CURRENT_FILE# fn,          CURRENT_BLOCK# blockn    from  v$active_session_history ash        , all_objects o    where event like ‘latch: cache buffers chains‘      and o.object_id (+)= ash.CURRENT_OBJ#    group by sql_id, current_obj#, current_file#,                   current_block#, o.object_name,o.object_type    order by  count(*) desc )where rownum <=10;

 

根据上面得到的sql_id信息查看sql全文

 

SQL>select sql_fulltext from v$sqlarea where sql_id=‘&sqlid‘;

 

查看SQL的执行计划 SQL>SELECT * FROM table(DBMS_XPLAN.DISPLAY_CURSOR((‘&sql_id‘,0));

 

在认为sql执行计划不准确的情况也可以通过sql_id查看sql的address和hash_value查看sql的实际执行计划

 

SQL>SELECT  address, hash_value FROM v$sql       WHERE sql_id=‘&sql_id‘;  SQL>SELECT operation, options, object_name, cost FROM v$sql_plan       WHERE address = ‘&addr‘ AND hash_value = http://www.mamicode.com/‘hash_v‘;

 

当某个会话长时间持有latch时,可以通过联合v$latchholder和v$session视图查看sql信息

 

SQL>SELECT s.sql_hash_value,s.sql_id,s.address, l.name   FROM V$SESSION s, V$LATCHHOLDER l WHERE s.sid = l.sid;

 

故障处理思路

 

1、根据sql执行计划判断该执行计划是否正确,sql执行过长往往意味着过长时间的持有latch。

 

2、优化nested loop join,如果有可能使用hash join代替nested loop join。也可以利用对热块索引进行hash分区,或者使用hash簇的方式减缓热块现象。

 

3、调整表的pctfree值,将数据尽可能的分布到多个块中

 

4、调整应用

 

5、等。呵呵当出现latch争用时,故障时刻确实没有较好的方式解决,找到病因才是关键。

latch: cache buffers chains故障处理总结