自治事件引起死锁

今天遇到一个比较特殊的死锁现象,记录下来

DEADLOCK DETECTED ( ORA-00060 )
[Transaction Deadlock]
The following deadlock is not an ORACLE error. It is a
deadlock due to user error in the design of an application
or from issuing incorrect ad-hoc SQL. The following
information may aid in determining the deadlock:
Deadlock graph:
                       ---------Blocker(s)--------  ---------Waiter(s)---------
Resource Name          process session holds waits  process session holds waits
TX-00090022-000002ba        15     145     X             15     145           X
session 145: DID 0001-000F-00000019     session 145: DID 0001-000F-00000019
Rows waited on:
Session 145: obj - rowid = 0000E0A3 - AAAOCjAAFAAAAA8AAA
  (dictionary objn - 57507, file - 5, block - 60, slot - 0)
Information on the OTHER waiting sessions:
End of information on OTHER waiting sessions.
Current SQL statement for this session:
UPDATE T SET Y = Y WHERE X = :B1 
----- PL/SQL Call Stack -----
  object      line  object
  handle    number  name
0x67987910         4  CHF.T
===================================================

问题原因:自治事件导致(重现)

drop table t;                                                                                                        
create table t ( x int, y int );
                                                                                                                 
create or replace trigger t before update on t
for each row
declare
    pragma autonomous_transaction;
begin
    update t set y = y where x = :new.x;
    commit;
end;
/
                                                                                                              
insert into t values ( 1, 1 );
commit;
update t set y = y where x = 1;
发表在 Oracle | 一条评论

DUMPING ORACLE BLOCKS

我见识过的,讲的最好的关于dump相关的文章,这个里面关于dump block,dump index,dump controlfile说的比较详细,看了收获不小,现在贡献出来和大家分享
如何阅读oracle数据块的dump文件
//www.xifenfei.com/wp-content/uploads/2011/09/read_oracle_dump.pdf

发表在 Oracle | 一条评论

假坏块引起恐慌

alert文件出现如下日志
Fri Sep 16 02:58:25 2011
Hex dump of (file 19, block 1444767) in trace file /opt/oracle/admin/ora9i/udump/ora9i_ora_24702.trc
Corrupt block relative dba: 0x04d60b9f (file 19, block 1444767)
Fractured block found during backing up datafile
Data in bad block:
type: 6 format: 2 rdba: 0x04d60b9f
last change scn: 0x0abf.56961827 seq: 0×1 flg: 0×04
spare1: 0×0 spare2: 0×0 spare3: 0×0
consistency value in tail: 0x7c340601
check value in block header: 0xba17
computed block checksum: 0×6413
Reread of blocknum=1444767, file=/opt/oracle/oradata/ora9i/TS_INDX_Base.005.dbf. found valid data

因为这个是我们比较重要的业务服务器,如果出现坏块,不能及时处理,后果将不堪设想,因此我马上检查

先查看是什么对象:(结果是一个分区表的index,担心减少一半,index大不了重建,无大事)

SELECT OWNER, SEGMENT_NAME, SEGMENT_TYPE, TABLESPACE_NAME, A.PARTITION_NAME
  FROM DBA_EXTENTS A
 WHERE FILE_ID = &FILE_ID
   AND &BLOCK_ID BETWEEN BLOCK_ID AND BLOCK_ID + BLOCKS - 1;

然后检查是否真的坏块
方法一:dbv

[oracle@DB1 bdump]$ dbv file=/opt/oracle/oradata/ora9i/TS_INDX_Base.005.dbf blocksize=8192

DBVERIFY: Release 10.2.0.4.0 - Production on Fri Sep 16 09:15:11 2011

Copyright (c) 1982, 2007, Oracle.  All rights reserved.

DBVERIFY - Verification starting : FILE = /opt/oracle/oradata/ora9i/TS_INDX_Base.005.dbf


DBVERIFY - Verification complete

Total Pages Examined         : 2207744
Total Pages Processed (Data) : 0
Total Pages Failing   (Data) : 0
Total Pages Processed (Index): 2201581
Total Pages Failing   (Index): 0
Total Pages Processed (Other): 3271
Total Pages Processed (Seg)  : 0
Total Pages Failing   (Seg)  : 0
Total Pages Empty            : 2892
Total Pages Marked Corrupt   : 0
Total Pages Influx           : 0
Highest block SCN            : 1454468581 (2751.1454468581)

结论无坏块

方法二:bbed

[oracle@DB1 lib]$ bbed
Password: 

BBED: Release 2.0.0.0.0 - Limited Production on Fri Sep 16 10:01:26 2011

Copyright (c) 1982, 2007, Oracle.  All rights reserved.

************* !!! For Oracle Internal Use only !!! ***************

BBED> set filename '/opt/oracle/oradata/ora9i/TS_INDX_Base.005.dbf'
        FILENAME        /opt/oracle/oradata/ora9i/TS_INDX_Base.005.dbf

BBED> set block 1444767
        BLOCK#          1444767

BBED> VERIFY 
DBVERIFY - Verification starting
FILE = /opt/oracle/oradata/ora9i/TS_INDX_Base.005.dbf
BLOCK = 1444767


DBVERIFY - Verification complete

Total Blocks Examined         : 1
Total Blocks Processed (Data) : 0
Total Blocks Failing   (Data) : 0
Total Blocks Processed (Index): 1
Total Blocks Failing   (Index): 0
Total Blocks Empty            : 0
Total Blocks Marked Corrupt   : 0
Total Blocks Influx           : 0

再次证明无坏块

方法三:rman
RMAN> backup validate datafile 19;
结果查询V$BACKUP_CORRUPTION也无坏块

方法四:使用index查询
结果也正常,没报错,alert中无错误记录

发现alert报错,但是实际没有错误,查询相关资料发现
从日志中可以看到,提示Corrupt的block对应的dba为0x04d60b9f (file 19, block 1444767),data block的类型为6(6为trans data,所有的data和index blocks都是该类型)。Oracle发现block有可能corrupt后,进行了reread,结果为found valid data,说明数据块未损坏。Fractured block found,表示rman发现这个数据块正在被使用,这时rman会进行重新读取,如果再次失败,才认为是坏块。如果第二次尝试读取时成功,则表示数据完好,不会产生影响。此类信息在IO负载较高的情况下进行rman备份时比较容易出现。
查询系统发现,正好该时间段是系统io负载最高的时候

发表在 Oracle备份恢复 | 评论关闭