OS:AIX52
DB:ORACLE9205 两节点的RAC
在TOP5中出现了global cache cr request事件,我自已查了一下segment_statistic这个视图,如下:
SQL> select wh.kcbwhdes "module",
2 sw.why0 "calls",
3 sw.why2 "waits",
4 sw.other_wait "caused waits"
5 from x$kcbwh wh, x$kcbsw sw
6 where wh.indx = sw.indx
7 and sw.other_wait > 0
8 order by sw.other_wait desc;
module calls waits caused waits
-------------------------------- ---------- ---------- ------------
kdswh01: kdstgr 2799526702 0 8277343
kdswh05: kdsgrp 727843717 0 1327815
kdiwh09: kdiixs 2310359964 0 314096
kdiwh15: kdifxs 514481080 0 311693
kdswh02: kdsgrp 820853004 0 35859
kdiwh07: kdifbk 734658715 0 18183
kdiwh100: kdircys 17058136 0 8165
kdiwh22: kdifind 399378239 0 6326
ktuwh05: ktugct 171461885 0 4166
ktuwh27: kturbk 4234766 0 1723
ktuwh01: ktugus 3479334 0 733
kdswh07: kdstlps 5384421 0 708
ktspbwh1: ktspfsrch 170125216 0 631
ktewh80: kteopdelete 3875 0 292
以下是相关解释:
Kdstgr Performs full table scan get row. Rows are accessed by a full table scan. Check the number of FULL table scans
Kdsgrp Performs get row piece. Typically, row pieces are affected only in case of chained and migrated rows. Row chaining has to be analyzed and fixed.
Kdiixs Performs index range scan.
从以上的现象上看,好象主要是由全表扫描和行链接和迁移造成的,但是我很少看到有全表扫描的语句,最常见的是一个Xa.commit.new什么的好象在做全表扫描.
这个不能确定,另外解决行链接和迁移需要做表移动吧,这样在多大程度上有效呢?
因为是生产数据库,先确认一下.谢谢!
来自 “ ITPUB博客 ” ,链接:https://blog.itpub.net/9375/viewspace-555101/
,如需转载,请注明出处,否则将追究法律责任。