一.1  BLOG文档结构图

image

 

一.2  前言部分

 

一.2.1  导读和注意事项

各位技术爱好者,看完本文后,你可以掌握如下的技能,也可以学到一些其它你所不知道的知识,~O(∩_∩)O~:

① truncate操作后的恢复方法(重点)

 

  Tips:

       ① 若文章代码格式有错乱,推荐使用QQ、搜狗或360浏览器,也可以下载pdf格式的文档来查看,pdf文档下载地址:http://yunpan.cn/cdEQedhCs2kFz (提取码:ed9b) 

       ② 本篇BLOG中命令的输出部分需要特别关注的地方我都用灰色背景和粉红色字体来表示,比如下边的例子中,thread 1的最大归档日志号为33,thread 2的最大归档日志号为43是需要特别关注的地方;而命令一般使用黄色背景和红色字体标注;对代码或代码输出部分的注释一般采用蓝色字体表示。

 

  List of Archived Logs in backup set 11

  Thrd Seq     Low SCN    Low Time            Next SCN   Next Time

  ---- ------- ---------- ------------------- ---------- ---------

  1    32      1621589    2015-05-29 11:09:52 1625242    2015-05-29 11:15:48

  1    33      1625242    2015-05-29 11:15:48 1625293    2015-05-29 11:15:58

  2    42      1613951    2015-05-29 10:41:18 1625245    2015-05-29 11:15:49

  2    43      1625245    2015-05-29 11:15:49 1625253    2015-05-29 11:15:53

 

 

 

 

[ZFXXDB1:root]:/>lsvg -o

T_XDESK_APP1_vg

rootvg

[ZFXXDB1:root]:/>

00:27:22 SQL> alter tablespace idxtbs read write;

 

 

====》2097152*512/1024/1024/1024=1G 

 

 

 

本文如有错误或不完善的地方请大家多多指正,ITPUB留言或QQ皆可,您的批评指正是我写作的最大动力。

 

 

一.2.2  相关参考文章链接

 

 

一.2.3  本文简介

truncate操作是比较危险的操作,不记录redo,不能通过闪回查询来找回数据,但是只要段所占用的块没有全部被重新占用的情况下,我们还是可以通过一些特殊的办法来找回truncate掉的数据,因为当Truncate命令发起之后,Oracle实际上并没有在删除底层数据块上的数据,而是要等到重用的时候才会把这一部分数据回收,于是这给了我们一个能够恢复数据库的机会。

总体而言,恢复的办法是通过一些大牛写的工具来恢复,分为收费和免费的,我们下边分别说明。实验部分我们只实验fy_recover_data包和gdul工具。

有的实验是很久之前做的,这篇文章发布太晚,因为中间学习了DUL和BBED的相关知识。

一.3  收费软件

 

这里简单列举一下,具体内容请到相关网站了解:

工具名称

下载地址

作者

软件

ODU

http://www.oracleodu.com/cn/

老熊

命令行操作

PRM-DUL

http://www.parnassusdata.com/ 

Maclean Liu

图形界面操作

AUL/mydul

http://www.dbatools.net/mydul/

d.c.b.a/楼方鑫

命令行

 

 

 

一.4  免费软件

一.4.1  fy_recover_data包

作者个人信息:

WWW.HelloDBA.COM                                                   

Created By: Fuyuncat                                               

Created Date: 08/08/2012                                           

Email: Fuyuncat@gmail.com                                          

Copyright (c), 2014, WWW.HelloDBA.COM All rights reserved.         

Latest Version: http://www.HelloDBA.com/download/FY_Recover_Data.zip

该包采用纯plsql语句恢复被truncate掉的表,操作比较简单,下载可以去官网下载,或者小麦苗的云盘共享目录。

 

Fy_Recover_Data是利用Oracle表扫描机制、数据嫁接机制恢复TRUNCATE或者损坏数据的工具包。由纯PLSQL编写,原理图如下:

wps33F2.tmp 

 

 

包内容:

wps3403.tmp

 

一.4.2  gdul工具

GDUL是老耿开发的一款类dul工具,当数据库由于某种原因无法打开时,可以利用GDUL把表数据直接读取出来,工具下载地址参考小麦苗的blog,老耿的信息如下:

*********************************************************************

  GDUL for ORACLE DB.

  Version 4.0.0.1, build date: 2016.04.12.

  Copyright (c) 2007, 2016. Andy Geng.  ALL RIGHTS RESERVED.

  Email: dbtool@aliyun.com

  WeChat official account: dbtool

  QQ group: 235019291

*********************************************************************

 

一.4.2.1  gDUL功能特点

v 完整支持多种格式导出,包括expdp,exp,text格式。目前市面上的类dul工具只有gDUL支持expdp格式。

v 支持ASM文件系统,并内置asmcmd命令。

v 支持绝大多数列类型,支持常见的NUMBER,CHAR, VARCHAR2, DATE,LOB, LONG等类型。。其中 SecureFile LOB 支持压缩,尚不支持去重和加密。

v 支持导出常规表、IOT、Cluster 表、分区表、压缩表。

v 支持 truncated 表、删除行恢复。

v 支持常规表空间和 bigfile 表空间。

v 支持主流硬件平台(HP-UX,AIX, Solaris, Linux, Windows),各个平台仅需单一的可执行文件,方便分发。

v 重点是——永久免费使用,无需额外费用,不开源。

 

 

 

一.4.3  dul

DUL 是 Data Unloader 的缩写,是一个荷兰的 Oracle 工程师开发的,他的名字为 Bernard Van Duijnen。 DUL 是一个 C 开发的小程序,编译后整个程序只有一个文件,大小也不过几百 KB,它工作时不需 Oracle RDBMS 以及任何的 Oracle 的程序、组件,它可以直接从一个坏了数据库的数据文件中读取数据,生成 IMP 或 SQL*Loader 可以识别的文件。

DUL 不是一个商用化的产品,Oracle 不卖、不提供也不支持它的使用。DUL 只有在 Oracle 的内部网站才可以下载到,因此也只有 Oracle 的 Supporter 才能下载到有这个工具,如果与 Oracle 的 Supporter 熟悉,没准他私底下会给你一个,这个工具也因此有一些流落到民间,被一些人收入囊中,奉为珍宝。 

不同的平台、不同版本的数据库都有相应的 DUL 软件,9.x 及之前 DUL 是没有 License 限制的,也就是有这个工具可以无限制的使用,不过最新的 DUL 在这方面已经改进了,kamus 说最新 DUL 拿到手只能用一个月。 

关于这一小点稍总结一下,获得 DUL 有以下几种途径: 

wps3404.tmp 如果你是 Oracle 的 Supporter ,可以在内部网站下载,地址为: http://www.nl.oracle.com/support/dul/ 

wps3405.tmp 如果你有 Oracle 的 Supporter 的朋友可以向他们要一个,itpub 也几位斑竹都到 Oracle 了,如coolyl,kamus,lunar。 

wps3406.tmp 一些 dul 流落到民间,可以向有这软件的朋友要一个,不过他们不一定有你需要的那个。 

 

所以关于DUL我们不做过多的解释。

 

一.4.4  bbed来恢复

这个比较复杂,若对oracle不熟悉或者bbed不熟悉都不推荐使用这个,具体案例参考:http://blog.itpub.net/26736162/viewspace-2080727/

 

 

第二章 实验部分

二.1  实验环境介绍

项目

db

db 类型

单实例

db version

11.2.0.4.0

db 存储

FS

主机IP地址/hosts配置

192.168.59.129

OS版本及kernel版本

AIX 7.1 64位

归档模式

Archive Mode

ORACLE_SID

oralhr

 

 

二.2  实验目标

将truncate掉的表数据成功找回。

 

二.3  实验过程

 

二.3.1  fy_recover_data包恢复truncate的表

[ZFXDESKDB1:oracle]:/oracle>ORACLE_SID=oraESKDB1

[ZFXDESKDB1:oracle]:/oracle>sqlplus / as sysdba

 

SQL*Plus: Release 11.2.0.4.0 Production on Mon Mar 21 15:51:55 2016

 

Copyright (c) 1982, 2013, Oracle.  All rights reserved.

 

 

Connected to:

Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production

With the Partitioning, Real Application Clusters, Automatic Storage Management, OLAP,

Data Mining and Real Application Testing options

 

SYS@oraESKDB1> set time on;

15:52:10 SYS@oraESKDB1> set timing on;

15:52:10 SYS@oraESKDB1> set serveroutput on;

15:52:10 SYS@oraESKDB1> create table scott.TB_0321    as SELECT * FROM dba_objects;

 

Table created.

 

Elapsed: 00:00:00.59

15:52:18 SYS@oraESKDB1> SELECT COUNT(1) FROM   scott.TB_0321;

 

  COUNT(1)

----------

     86651

 

Elapsed: 00:00:00.19

15:52:24 SYS@oraESKDB1> INSERT INTO scott.TB_0321 SELECT * FROM scott.TB_0321;

 

 

86651 rows created.

 

Elapsed: 00:00:00.26

15:52:30 SYS@oraESKDB1> COMMIT;

Commit complete.

 

Elapsed: 00:00:00.01

15:52:30 SYS@oraESKDB1> INSERT INTO scott.TB_0321 SELECT * FROM scott.TB_0321;

COMMIT;

 

173302 rows created.

 

Elapsed: 00:00:00.43

15:53:02 SYS@oraESKDB1> SELECT COUNT(1) FROM   scott.TB_0321;

 

  COUNT(1)

----------

  346604

 

Elapsed: 00:00:00.27

16:15:18 SYS@oraESKDB1> SELECT d.BYTES/1024/1024 FROM dba_segments d WHERE d.segment_name ='TB_0321';

 

D.BYTES/1024/1024

-----------------

               40

 

Elapsed: 00:00:00.44

16:15:25 SYS@oraESKDB1> truncate table scott.TB_0321;

 

Table truncated.

 

Elapsed: 00:00:00.20

16:15:46 SYS@oraESKDB1> SELECT COUNT(1) FROM   scott.TB_0321;

 

  COUNT(1)

----------

         0

 

Elapsed: 00:00:00.01

 

====》数据已经被truncate掉了,下边我们来恢复

 

 

16:15:52 SYS@oraESKDB1> @/oracle/FY_Recover_Data.pck

 

Package created.

 

Elapsed: 00:00:00.06

 

Package body created.

 

Elapsed: 00:00:00.03

16:15:59 SYS@oraESKDB1> exec fy_recover_data.recover_truncated_table('scott','TB_0321');

16:16:06: Use existing Directory Name: FY_DATA_DIR

16:16:07: Recover Table: SCOTT.TB_0321$

16:16:09: Restore Table: SCOTT.TB_0321$$

16:16:24: Copy file of Recover Tablespace: FY_REC_DATA_COPY.DAT1

16:16:24: begin to recover table SCOTT.TB_0321

16:16:24: Use existing Directory Name: TMP_HF_DIR

16:17:09: Recovering data in datafile +DATA/oraeskdb/datafile/users.351.902678817

16:17:09: Use existing Directory Name: TMP_HF_DIR

16:39:16: 4984 truncated data blocks found.

16:39:16: 346604 records recovered in backup table SCOTT.TB_0321$$

16:39:17: Total: 4984 truncated data blocks found.

16:39:17: Total: 346604 records recovered in backup table SCOTT.TB_0321$$

16:39:17: Recovery completed.

16:39:17: Data has been recovered to SCOTT.TB_0321$$

 

PL/SQL procedure successfully completed.

 

Elapsed: 00:23:11.59

 

16:39:17 SYS@oraESKDB1> SELECT COUNT(1) FROM   scott.TB_0321$$;

 

  COUNT(1)

----------

   346604

 

Elapsed: 00:00:01.55

16:40:51 SYS@oraESKDB1>

16:40:51 SYS@oraESKDB1> alter table scott.TB_0321 nologging;

 

Table altered.

 

Elapsed: 00:00:00.03

16:41:43 SYS@oraESKDB1> insert /*+append*/ into scott.TB_0321 select * from scott.TB_0321$$;

 

346604 rows created.

 

Elapsed: 00:00:00.86

16:41:52 SYS@oraESKDB1> commit;

 

Commit complete.

 

Elapsed: 00:00:00.01

16:41:55 SYS@oraESKDB1> alter table scott.TB_0321 logging;

 

Table altered.

 

Elapsed: 00:00:00.02

16:42:06 SYS@oraESKDB1>

16:42:06 SYS@oraESKDB1> drop tablespace   FY_REC_DATA  including contents and datafiles;

 

Tablespace dropped.

 

Elapsed: 00:00:08.00

16:42:35 SYS@oraESKDB1> drop tablespace   FY_RST_DATA  including contents and datafiles;

 

Tablespace dropped.

 

Elapsed: 00:00:07.59

16:42:44 SYS@oraESKDB1>

 

 

数据成功恢复。

 

 

二.3.2  gdul恢复truncate的表

set time on;

set timing on;

set serveroutput on;

drop table scott.TB_0322_05;

create table scott.TB_0322_05    as SELECT * FROM dba_objects;

 

SELECT COUNT(1) FROM   scott.TB_0322_05;

INSERT INTO scott.TB_0322_05 SELECT * FROM scott.TB_0322_05;

COMMIT;

INSERT INTO scott.TB_0322_05 SELECT * FROM scott.TB_0322_05;

COMMIT;

INSERT INTO scott.TB_0322_05 SELECT * FROM scott.TB_0322_05;

COMMIT;

INSERT INTO scott.TB_0322_05 SELECT * FROM scott.TB_0322_05;

COMMIT;

SELECT COUNT(1) FROM   scott.TB_0322_05;

 

SELECT d.BYTES/1024/1024 FROM dba_segments d WHERE d.segment_name ='TB_0322_05';

 

 

truncate table scott.TB_0322_05;

 

alter system checkpoint;

 

col ownere format a10

col DIRECTORY_NAME format a30

col DIRECTORY_PATH format a50

select OWNER,DIRECTORY_NAME,DIRECTORY_PATH from  dba_directories;

 

 

bootstrap

desc scott.TB_0322_05

unload table  scott.TB_0322_05

scan tablespace 4

untrunc table  scott.TB_0322_05

 

cp SCOTT_TB_0322_05.dmp /oracle/app/oracle/admin/oralhr/dpdump/

impdp  scott/tiger directory=DATA_PUMP_DIR dumpfile=SCOTT_TB_0322_05.dmp LOGFILE=SCOTT_TB_0322_05.log TABLES=TB_0322_05

 

15:41:04 SQL> set time on;

15:59:49 SQL> set timing on;

15:59:49 SQL> set serveroutput on;

15:59:49 SQL> drop table scott.TB_0322_05;

create table scott.TB_0322_05    as SELECT * FROM dba_objects;

 

SELECT COUNT(1) FROM   scott.TB_0322_05;

INSERT INTO scott.TB_0322_05 SELECT * FROM scott.TB_0322_05;

 

Table dropped.

 

Elapsed: 00:00:00.07

15:59:49 SQL> COMMIT;

INSERT INTO scott.TB_0322_05 SELECT * FROM scott.TB_0322_05;

COMMIT;

INSERT INTO scott.TB_0322_05 SELECT * FROM scott.TB_0322_05;

COMMIT;

INSERT INTO scott.TB_0322_05 SELECT * FROM scott.TB_0322_05;

COMMIT;

SELECT COUNT(1) FROM   scott.TB_0322_05;

 

SELECT d.BYTES/1024/1024 FROM dba_segments d WHERE d.segment_name ='TB_0322_05';

 

 

truncate table scott.TB_0322_05;

 

alter system checkpoint;

 

Table created.

 

Elapsed: 00:00:00.97

15:59:50 SQL> 15:59:50 SQL>

  COUNT(1)

----------

     75707

 

Elapsed: 00:00:00.86

15:59:51 SQL>

75707 rows created.

 

Elapsed: 00:00:00.23

15:59:52 SQL>

Commit complete.

 

Elapsed: 00:00:00.17

15:59:52 SQL>

151414 rows created.

 

Elapsed: 00:00:00.50

15:59:52 SQL>

Commit complete.

 

Elapsed: 00:00:00.23

15:59:52 SQL>

302828 rows created.

 

Elapsed: 00:00:01.63

15:59:54 SQL>

Commit complete.

 

Elapsed: 00:00:00.22

15:59:54 SQL>

605656 rows created.

 

Elapsed: 00:00:06.19

16:00:00 SQL>

Commit complete.

 

Elapsed: 00:00:00.02

16:00:01 SQL>

  COUNT(1)

----------

   1211312

 

Elapsed: 00:00:00.07

16:00:01 SQL> 16:00:01 SQL>

D.BYTES/1024/1024

-----------------

              136

 

Elapsed: 00:00:00.17

16:00:01 SQL> 16:00:01 SQL> 16:00:01 SQL>

Table truncated.

 

Elapsed: 00:00:01.26

16:00:02 SQL> 16:00:02 SQL>

System altered.

 

Elapsed: 00:00:00.15

16:00:02 SQL>

16:00:02 SQL> SELECT COUNT(1) FROM   scott.TB_0322_05;

 

  COUNT(1)

----------

         0

 

Elapsed: 00:00:00.00

16:02:35 SQL>

 

[oracle@ZFFR4CB1101:/home/oracle/gdul]$ ./gdul

 

*********************************************************************

  GDUL for ORACLE DB.

  Version 3.5.0.1, build date: 2016.03.07.

  Copyright (c) 2007, 2016. Andy Geng.  ALL RIGHTS RESERVED.

  Email: gengyonghui@aliyun.com

  QQ group: 235019291, WeChat Official Account: dbtool

*********************************************************************

 

GDUL> bootstrap

Bootstrap finish.

GDUL> desc scott.TB_0322_05

 

object_id: 78302, dataobj#: 78303, cluster tab#: 0

segment header: (ts#: 4, rfile#: 4, block#: 682))

 

Seg Column#  Column#    Name                 Null?           Type     

------------ ---------- -------------------- --------------- --------------

1            1          OWNER                                VARCHAR2(30)

2            2          OBJECT_NAME                          VARCHAR2(128)

3            3          SUBOBJECT_NAME                       VARCHAR2(30)

4            4          OBJECT_ID                            NUMBER   

5            5          DATA_OBJECT_ID                       NUMBER   

6            6          OBJECT_TYPE                          VARCHAR2(19)

7            7          CREATED                              DATE     

8            8          LAST_DDL_TIME                        DATE     

9            9          TIMESTAMP                            VARCHAR2(19)

10           10         STATUS                               VARCHAR2(7)

11           11         TEMPORARY                            VARCHAR2(1)

12           12         GENERATED                            VARCHAR2(1)

13           13         SECONDARY                            VARCHAR2(1)

14           14         NAMESPACE                            NUMBER   

15           15         EDITION_NAME                         VARCHAR2(30)

 

GDUL> unload table  scott.TB_0322_05

2016-03-22 16:01:54...unloaded "SCOTT"."TB_0322_05"   0 rows

GDUL> scan tablespace 4

start scan tablespace 4...

scan tablespace completed.

GDUL> untrunc table  scott.TB_0322_05

2016-03-22 16:04:29...untruncating table TB_0322_05 1211312 rows unloaded.

GDUL>

 

16:02:35 SQL> select * from dba_directories;

 

OWNER                          DIRECTORY_NAME                 DIRECTORY_PATH

------------------------------ ------------------------------ -----------------------------------------------------------------------

SYS                            SUBDIR                         /u01/app/oracle/product/11.2.0/dbhome_1/demo/schema/order_entry//2002/Sep

SYS                            SS_OE_XMLDIR                   /u01/app/oracle/product/11.2.0/dbhome_1/demo/schema/order_entry/

SYS                            LOG_FILE_DIR                   /u01/app/oracle/product/11.2.0/dbhome_1/demo/schema/log/

SYS                            MEDIA_DIR                      /u01/app/oracle/product/11.2.0/dbhome_1/demo/schema/product_media/

SYS                            XMLDIR                         /u01/app/oracle/product/11.2.0/dbhome_1/rdbms/xml

SYS                            DATA_FILE_DIR                  /u01/app/oracle/product/11.2.0/dbhome_1/demo/schema/sales_history/

SYS                            DATA_PUMP_DIR                  /u01/app/oracle/product/11.2.0/dbhome_1/rdbms/log/

SYS                            ORACLE_OCM_CONFIG_DIR          /u01/app/oracle/product/11.2.0/dbhome_1/ccr/state

 

8 rows selected.

 

Elapsed: 00:00:00.00

16:05:29 SQL>

 

[oracle@ZFFR4CB1101:/home/oracle/gdul/dump]$ impdp  scott/tiger directory=DATA_PUMP_DIR dumpfile=SCOTT_TB_0322_05.dmp LOGFILE=SCOTT_TB_0322_05.log TABLES=TB_0322_05

 

Import: Release 11.2.0.3.0 - Production on Tue Mar 22 16:16:48 2016

 

Copyright (c) 1982, 2011, Oracle and/or its affiliates.  All rights reserved.

 

Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production

With the Partitioning, Real Application Clusters, Automatic Storage Management, OLAP,

Data Mining and Real Application Testing options

Master table "SCOTT"."SYS_IMPORT_TABLE_01" successfully loaded/unloaded

Starting "SCOTT"."SYS_IMPORT_TABLE_01":  scott/******** directory=DATA_PUMP_DIR dumpfile=SCOTT_TB_0322_05.dmp LOGFILE=SCOTT_TB_0322_05.log TABLES=TB_0322_05

Processing object type TABLE_EXPORT/TABLE/TABLE_DATA

. . imported "SCOTT"."TB_0322_05"                        117.1 MB 1211312 rows

Job "SCOTT"."SYS_IMPORT_TABLE_01" successfully completed at 16:16:59

 

[oracle@ZFFR4CB1101:/home/oracle/gdul/dump]$

[oracle@ZFFR4CB2101:/home/oracle]$ sqlplus / as sysdba

 

SQL*Plus: Release 11.2.0.3.0 Production on Tue Mar 22 16:17:39 2016

 

Copyright (c) 1982, 2011, Oracle.  All rights reserved.

 

 

Connected to:

Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production

With the Partitioning, Real Application Clusters, Automatic Storage Management, OLAP,

Data Mining and Real Application Testing options

 

SQL> SELECT COUNT(1) FROM   scott.TB_0322_05;

 

  COUNT(1)

----------

   1211312

 

SQL>

 

数据成功恢复。

 

 

 

二.4  实验总结

 

总体而言用fy_recover_data包或GDUL工具都是非常好的,fy_recover_data可以恢复truncate的数据,但不能恢复drop的数据,而GDUL工具就比较全面了,具体可以参考前边的简介或下载文档来看,小麦苗的共享云盘里也有比较全的文档。