记一次ORACLE的UNDO表空间爆满分析过程

这篇文章是记录一次ORACLE数据库UNDO表空间爆满的分析过程，主要整理、梳理了同事分析的思路。具体过程如下所示：

早上收到一数据库服务器的UNDO表空间的告警邮件，最早一封是7：55发出的（监控作业是15分钟一次），从告警邮件分析，好像是UNDO表空间突然一下子被耗尽了。

DB	Tablespace	Allocated	Free	Used	% Free	% Used
192.168.xxx.xxx:1521	UNDOTBS1	16384	190.25	16193.75	1.16	99

使用一些SQL分析了undo表空间使用情况，以及undo segment状态等等，非常想定位到是哪个或那些SQL耗尽了UNDO表空间，但是没有一个SQL能实现我的想法，抑或是我不了解。

SELECT  UPPER(F.TABLESPACE_NAME)                           AS "TABLESPACE_NAME",

       ROUND(D.MAX_BYTES,2)                               AS "TBS_TOTAL_SIZE" ,

       ROUND(D.AVAILB_BYTES ,2)                           AS "TABLESPACE_SIZE",

       ROUND(D.MAX_BYTES - D.AVAILB_BYTES +USED_BYTES,2)  AS "TBS_AVABLE_SIZE",

       ROUND((D.AVAILB_BYTES - F.USED_BYTES),2)           AS "TBS_USED_SIZE",

       TO_CHAR(ROUND((D.AVAILB_BYTES - F.USED_BYTES) / D.AVAILB_BYTES * 100,

                    2),

              '999.99')                                  AS "USED_RATE(%)",

      ROUND(F.USED_BYTES, 6)                             AS "FREE_SIZE(G)"

FROM (SELECT TABLESPACE_NAME,

              ROUND(SUM(BYTES) / (1024 * 1024 * 1024), 6) USED_BYTES,

              ROUND(MAX(BYTES) / (1024 * 1024 * 1024), 6) MAX_BYTES

         FROM SYS.DBA_FREE_SPACE

        GROUP BY TABLESPACE_NAME) F,

      (SELECT DD.TABLESPACE_NAME,

              ROUND(SUM(DD.BYTES) / (1024 * 1024 * 1024), 6)  AVAILB_BYTES,

              ROUND(SUM(DECODE(DD.MAXBYTES, 0, DD.BYTES, DD.MAXBYTES))/(1024*1024*1024),6)   MAX_BYTES

         FROM SYS.DBA_DATA_FILES DD

        GROUP BY DD.TABLESPACE_NAME) D

HERE D.TABLESPACE_NAME = F.TABLESPACE_NAME

 AND D.TABLESPACE_NAME=&UNDO_TABLESPACE_NAME

RDER BY 5 DESC;

select usn,xacts,rssize/1024/1024/1024,hwmsize/1024/1024/1024,shrinks

from v$rollstat order by rssize;

既然直接入手，无法定位，那就曲线分析，首先检查、分析了一下redo log，发现在7点这段时间，日志切换了83次之多，横向、纵向对比，明显异常，如下截图所示：

SELECT 

TO_CHAR(FIRST_TIME,'YYYY-MM-DD') DAY,

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'00',1,0)),'99') "00",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'01',1,0)),'99') "01",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'02',1,0)),'99') "02",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'03',1,0)),'99') "03",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'04',1,0)),'99') "04",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'05',1,0)),'99') "05",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'06',1,0)),'99') "06",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'07',1,0)),'99') "07",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'08',1,0)),'99') "0",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'09',1,0)),'99') "09",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'10',1,0)),'99') "10",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'11',1,0)),'99') "11",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'12',1,0)),'99') "12",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'13',1,0)),'99') "13",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'14',1,0)),'99') "14",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'15',1,0)),'99') "15",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'16',1,0)),'99') "16",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'17',1,0)),'99') "17",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'18',1,0)),'99') "18",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'19',1,0)),'99') "19",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'20',1,0)),'99') "20",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'21',1,0)),'99') "21",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'22',1,0)),'99') "22",

TO_CHAR(SUM(DECODE(TO_CHAR(FIRST_TIME,'HH24'),'23',1,0)),'99') "23"

FROM

V$LOG_HISTORY

GROUP BY 

TO_CHAR(FIRST_TIME,'YYYY-MM-DD') 

ORDER BY 1 DESC;

生成了实例在7:00~8:00时间段的AWR报告，从下面指标我们可以看出，数据库实例在这段时间呢，其实是非常空闲的，因为DB Time为9.74（mins)

另外，从Time Model Statistics部分来看，主要时间花在background elapsed time，而不是DB Time,我们可以判断时间主要耗费在后台进程，而不是前台进程。另外sql execute elapsed time耗用了DB Time的70.36的时间。

然后我们来看SQL order by Gets部分信息, 第一个SQL是删除WRH$_SQL_PLAN的记录，当然也有删除wrh$_sqltext、WRH$_SEG_STAT_OBJ表记录的SQL，如下所示

DELETE

FROM WRH$_SQL_PLAN tab

WHERE (:beg_snap <= tab.snap_id

AND tab.snap_id  <= :end_snap

AND dbid          = :dbid)

AND NOT EXISTS

  (SELECT 1

  FROM WRM$_BASELINE b

  WHERE (tab.dbid   = b.dbid)

  AND (tab.snap_id >= b.start_snap_id)

  AND (tab.snap_id <= b.end_snap_id)

  )

DELETE

FROM wrh$_sqltext tab

WHERE (tab.dbid   = :dbid

AND :beg_snap    <= tab.snap_id

AND tab.snap_id  <= :end_snap

AND tab.ref_count = 0)

AND NOT EXISTS

  (SELECT 1

  FROM WRM$_BASELINE b

  WHERE (b.dbid    = :dbid2

  AND tab.snap_id >= b.start_snap_id

  AND tab.snap_id <= b.end_snap_id)

  );

DELETE

FROM WRH$_SEG_STAT_OBJ tab

WHERE (:beg_snap <= tab.snap_id

AND tab.snap_id  <= :end_snap

AND dbid          = :dbid)

AND NOT EXISTS

  (SELECT 1

  FROM WRM$_BASELINE b

  WHERE (tab.dbid   = b.dbid)

  AND (tab.snap_id >= b.start_snap_id)

  AND (tab.snap_id <= b.end_snap_id)

  );

查看SQL ordered by Reads部分信息，发现主要也是删除系统表WRH$_SQL_PLAN记录 （这个表是非常大的）

DELETE

FROM WRH$_SQL_PLAN tab

WHERE (:beg_snap <= tab.snap_id

AND tab.snap_id  <= :end_snap

AND dbid          = :dbid)

AND NOT EXISTS

  (SELECT 1

  FROM WRM$_BASELINE b

  WHERE (tab.dbid   = b.dbid)

  AND (tab.snap_id >= b.start_snap_id)

  AND (tab.snap_id <= b.end_snap_id)

  )

然后我们查看AWR报告的Tablespace IO Stats部分，IO主要集中在SYSAUX,UNDOTBS1这两个表空间，然后你会发现那个表WRH$_SQL_PLAN就是在SYSAUX下

所以，上面种种证据显示，让我们几乎可以断定主要是下面这个SQL导致了UNDO表空间使用的暴增。当然分析过程中，还有一些旁听佐证。在此感觉没有必要一一列举了。

DELETE

FROM WRH$_SQL_PLAN tab

WHERE (:beg_snap <= tab.snap_id

AND tab.snap_id  <= :end_snap

AND dbid          = :dbid)

AND NOT EXISTS

  (SELECT 1

  FROM WRM$_BASELINE b

  WHERE (tab.dbid   = b.dbid)

  AND (tab.snap_id >= b.start_snap_id)

  AND (tab.snap_id <= b.end_snap_id)

  )