【BBED】bbed常用命令
BBED(Oracle Block Browerand EDitor Tool)，用來直接查看和修改數據文件數據的一個工具，是Oracle一款內部工具，可以直接修改Oracle數據文件塊的內容，在一些極端恢復場景下比較有用。該工具不受Oracle支持，所以默認是沒有生成可執行文件的，在使用前需要重新連接。
ls -l $ORACLE_HOME/rdbms/lib/*sbbd*
ls -l $ORACLE_HOME/rdbms/mesg/bbed*
chown oracle:dba /u01/app/oracle/product/11.2.0/dbhome_1/rdbms/lib/*sbbd*
chown oracle:dba /u01/app/oracle/product/11.2.0/dbhome_1/rdbms/mesg/bbed*
--cd $ORACLE_HOME/rdbms/lib
--make -f ins_rdbms.mk $ORACLE_HOME/rdbms/lib/bbed
--make -f $ORACLE_HOME/rdbms/lib/ins_rdbms.mk $ORACLE_HOME/rdbms/lib/bbed
make -f $ORACLE_HOME/rdbms/lib/ins_rdbms.mk BBED=$ORACLE_HOME/bin/bbed $ORACLE_HOME/bin/bbed
vi /home/oracle/file.txt
set line 9999 pagesize 9999
col name format a80
select file#||' '||name||' '||bytes name from v$datafile;
vi /home/oracle/bbed.par
blocksize=8192
listfile=/home/oracle/file.txt
mode=edit
bbed parfile=/home/oracle/bbed.par
bbed PASSWORD=blockedit mode=edit blocksize=8192 listfile=/home/oracle/file.txt
① 若使用bbed的過程中，數據庫有重啟的過程，最好是退出BBED重新進入bbed的環境
② windows下BBED軟件和其他系統下BBED不太一樣，操作的時候塊號比其他系統下多一個
bbed啟動和運行的過程會將運行過的所有命令記錄到當前的目錄下log.bbd文件中，所以bbed要求oracle用戶在當前目錄具有創建文件的權限。
oracle用戶需要對當前操作目錄有權限，否則報錯：
[ZFZHLHRDB2:oracle]:/oracle>bbed
BBED-00303: unable to open file 'log.bbd'
[ZFZHLHRDB2:oracle]:/oracle>touch log.bbd
touch: 0652-046 Cannot create log.bbd.
[ZFZHLHRDB2:oracle]:/oracle>ls -ld
drwxr-xr-x 5 root dba 256 Mar 14 15:41 .
[ZFZHLHRDB2:oracle]:/oracle>cd /home/oracle
[ZFZHLHRDB2:oracle]:/home/oracle>ls -ld
drwxr-xr-x 3 oracle dba 256 Apr 05 17:01 .
[ZFZHLHRDB2:oracle]:/home/oracle>bbed
Password:
BBED-00113: Invalid password. Please rerun utility with the correct password.
[ZFZHLHRDB2:oracle]:/home/oracle>bbed parfile=/home/oracle/bbed.par
Password:
BBED: Release 2.0.0.0.0 - Limited Production on Tue Apr 5 17:28:35 2016
Copyright (c) 1982, 2011, Oracle and/or its affiliates. All rights reserved.
************* !!! For Oracle Internal Use only !!! ***************
BBED> info
File# Name Size(blks)
----- ---- ----------
1 /oracle/app/oracle/oralhr/system01.dbf 96000
2 /oracle/app/oracle/oralhr/sysaux01.dbf 62720
3 /oracle/app/oracle/oralhr/undotbs01.dbf 11520
4 /oracle/app/oracle/oralhr/users01.dbf 640
5 /oracle/app/oracle/oralhr/example01.dbf 40080
先看幫助的說明：
BBED> help all
SET DBA [ dba | file#, block# ]
SET FILENAME 'filename'
SET FILE file#
SET BLOCK [+/-]block#
SET OFFSET [ [+/-]byte offset | symbol |*symbol ]
SET BLOCKSIZE bytes
SET LIST[FILE] 'filename'
SET WIDTH character_count
SET COUNT bytes_to_display
SET IBASE [ HEX | OCT | DEC ]
SET OBASE [ HEX | OCT | DEC ]
SET MODE [ BROWSE | EDIT ]
SET SPOOL [ Y | N ]
SHOW [ <SET parameter> | ALL ]
INFO
MAP[/v] [ DBA | FILENAME | FILE | BLOCK ]
DUMP[/v] [ DBA | FILENAME | FILE | BLOCK |OFFSET | COUNT ]
PRINT[/x|d|u|o|c] [ DBA | FILE | FILENAME |BLOCK | OFFSET | symbol | *symbol ]
EXAMINE[/Nuf] [ DBA | FILE | FILENAME |BLOCK | OFFSET | symbol | *symbol ]
</Nuf>:
N - a number which specifies a repeatcount.
u - a letter which specifies a unit size:
b -b1, ub1 (byte)
h -b2, ub2 (half-word)
w -b4, ub4(word)
r -Oracle table/index row
f - a letter which specifies a displayformat:
x -hexadecimal
d -decimal
u -unsigned decimal
o -octal
c -character (native)
n -Oracle number
t -Oracle date
i -Oracle rowid
FIND[/x|d|u|o|c] numeric/character string [TOP | CURR ]
COPY [ DBA | FILE | FILENAME | BLOCK ] TO [DBA | FILE | FILENAME | BLOCK ]
MODIFY[/x|d|u|o|c] numeric/character string
[ DBA | FILE | FILENAME | BLOCK | OFFSET | symbol | *symbol ]
ASSIGN[/x|d|u|o] <targetspec>=<source spec>
<target spec> : [ DBA | FILE |FILENAME | BLOCK | OFFSET | symbol | *symbol ]
<source spec> : [ value | <targetspec options> ]
SUM [ DBA | FILE | FILENAME | BLOCK ] [APPLY ]
PUSH [ DBA | FILE | FILENAME | BLOCK |OFFSET ]
POP [ALL]
REVERT [ DBA | FILE | FILENAME | BLOCK ]
UNDO
HELP [ <bbed command> | ALL ]
VERIFY [ DBA | FILE | FILENAME | BLOCK ]
CORRUPT [ DBA | FILE | FILENAME | BLOCK ]
下面是幾個常用的：
set 設定當前的環境
show 查看當前的環境參數，跟sqlplus的同名命令類似。
dump 列出指定block的內容
find 在指定的block中查找指定的字符串，結果是顯示出字符串，及其偏移量--offset，偏移量就是在block中的字節數
modify 修改指定block的指定偏移量的值，可以在線修改。
copy 把一個block的內容copy到另一個block中
verify 檢查當前環境是否有壞塊
sum 計算block的checksum，modify之後block就被標識為壞塊，current checksum與reqired checksum不一致，sum命令可以計算出新的checksum並應用到當前塊。
undo 回滾當前的修改操作，如果手誤做錯了，undo一下就ok了，回到原來的狀態。
revert 回滾所有之前的修改操作，意思就是 undo all
Set the current data block using the standard Oracle DBA (Data Block Address)format. This is entered as file_id, block.
關於DBA 說明，參考：Oracle rdba和 dba 說明 http://blog.csdn.net/tianlesoftware/article/details/6529346
select rowid,
dbms_rowid.rowid_relative_fno(rowid)rel_fno,
dbms_rowid.rowid_block_number(rowid)blockno,
dbms_rowid.rowid_row_number(rowid) rowno
from scott.emp;
set dba 4,145
如果設置成功，會返回該block的RDBA (Relative Data Block Address)，rdba就是rowid中的rfile#+block#。括號裡面的是DBA值和block 和 file id。 我們驗證一下：
sqlplus執行：
variable dba varchar2(30);
exec :dba :=dbms_utility.make_data_block_address(4,145);
print dba
Sets the current file to the one specified. It must be a valid Oracle data file andit must be enclosed in single quotes. If the file is not in the current path itmust also be fully qualified. If successful, bbed will respond showing the filenow being accessed.
BBED> set filename '/u01/app/oracle/oradata/dave2/users01.dbf'
FILENAME /u01/app/oracle/oradata/dave2/users01.dbf
--必須是一個有效的datafile，並且用單引號括起來
Sets the current file to the number specified. The number specified must be one ofthe file ids supplied in the filelist referenced at startup. If successful,bbed will respond showing the file id now being accessed.
BBED> set file 4
FILE# 4
--注意這裡的number，是我們之前配置的filelist裡的number。它可以和我們db 裡的file id 不一樣。 不過最好是配置一樣的。
Sets the current block. The block is relative to the filename or file already set.The absolute block can be specified, or an offset to the current block can bespecified using the plus (+) or (-) symbols. If successful, bbed will respondshowing the current block.
--注意這裡的block 是一個相對的位置，我們需要先指定一個file，然後在指定block。 即對應file裡的block。可以對當前block的位置進行+和-操作。
BBED> set file 4
FILE# 4
BBED> set block 60882
BBED-00309: out of range block number(60882)
BBED> set file 1
FILE# 1
BBED> set block 60882
BLOCK# 60882
BBED> set bock +10
BBED-00202: invalid parameter (bock)
BBED> set block +10
BLOCK# 60892
BBED> set block -10
BLOCK# 60882
BBED> set file 4 block 520
FILE# 4
BLOCK# 520
BBED> set dba 4,520
DBA 0x01000208 (16777736 4,520)
Sets the current offset. The offset is relative to the block already set. Theabsolute offset can be specified, or an offset to the current offset can bespecified using the plus (+) or minus (-) symbols. If successful, bbed willrespond showing the current offset.
--偏移量是相對某個block裡的偏移量，可以用+和-進行操作
BBED> set offset 20
OFFSET 20
BBED> set offset +2
OFFSET 22
BBED> set offset -2
OFFSET 20
Sets the blocksize of the current file. The blocksize must match the file selectedor an error will be reported. If successful, bbed will respond showing thecurrent blocksize.
設置當前datafile 的blocksize 大小，該大小必須和datafile 的實際block 匹配，否則會報錯。
BBED> set blocksize 8192
BLOCKSIZE 8192
Sets the listfile to the specified file. This option can be used if the listfile wasnot specified on the command line. The listfile must be enclosed in singlequotes. If successful, bbed will respond showing the current listfile.
--在前面講過，可以通過parameter file 來指定bbed的屬性，當然也可以通過set 來指定這些信息。對於listfile的文件，必須用單引號括起來。
BBED> set listfile '/u01/filelist.txt'
LISTFILE /u01/filelist.txt
Sets the current screen width. If not specified bbed will assume an 80-characterdisplay.
設定當前屏幕的寬度，默認是80.
BBED> set width 200
WIDTH 200
Sets the number of bytes of the data block to display from the dump command. Thedefault is 512.
Tosee an entire 8Kb block therefore you would need to dump the block eight timesat offsets 0, 512, 1024, 1536, 2048, 2560, 3092 and 3604.
By setting the count higher bbed will dump more of the block each time. Byreducing it a smaller dump can be achieved.
設置dump 命令顯示bytes的數量。默認是512 bytes。
BBED> set count 512
COUNT 512
Sets the internal number base. The default is decimal. However it can also be set tohexadecimal or octal. This allows the set file, set block and set offsetcommands to use an alternate base to decimal. If successful, bbed will respondshowing the current base:
--設置內部的數字格式，默認是十進制。 也可以設置為十六進制或者八進制。設置完數字格式之後，可是使用該格式來設置blcok，offset等。
BBED> set ibase hex
IBASE Hex
BBED> set block +D
BLOCK# 14
BBED> set ibase decimal
IBASE Dec
Sets the bbed mode. The options are browse or edit. In browse mode no changes can bemade. This is the suggested mode for first-time users, or if you are intendingto use the tool only to inspect data blocks.
--設置bbed 的模式，該默認有2種：browse 和 edit。 browse 模式不允許進行修改。 如果要修改，就選擇edit模式。 這個在我們的之前的配置文件裡，我們選擇了edit。
BBED> set mode browse
MODE Browse
BBED> set mode edit
MODE Edit
顯示當前的配置選項。
show
BBED> show
FILE# 1
BLOCK# 14
OFFSET 0
DBA 0x0040000e(4194318 1,14)
--注意這裡的block 變成了14. 是我們剛才設置的。
FILENAME /u01/app/oracle/oradata/dave2/system01.dbf
BIFILE bifile.bbd
LISTFILE /u01/filelist.txt
BLOCKSIZE 8192
MODE Edit
EDIT Unrecoverable
IBASE Dec
OBASE Dec
WIDTH 200
COUNT 512
LOGFILE log.bbd
SPOOL No
顯示當前可以進行browse 或者edit 的file。即我們filelist 裡指定的datafile信息。
Info
The map command shows a map of the current block. It can be combined with the /voption to produce a more verbose output. The map shows the offsets throughout the block where certain information can be found such as the block header, the data block header or the row directory.
If the set commands have not been used to set a current block, or it the user simply wishes to examine another block while keeping the current block their focus,the file name, file id, block or DBA can be specified with the command.
Map會通過偏移量來顯示block裡的詳細信息，如block header，data block header 和row directory。 使用/v 選項，可以查看更詳細的信息。
在不指定block的情況下，會顯示當前block的信息，如果想顯示其他block的信息，可以使用file name，file id，block 和DBA 來指定要顯示的block。
--通過dba 來指定某個block
map /v dba 4,145
Map 顯示的具體信息解釋如下：
struct kcbh, 20 bytes
Block Header Structure
ub1 type_kcbh
Block type (see Header Block Types below)
ub1 frmt_kcbh
Block format 1=Oracle 7, 2=Oracle 8+
ub1 spare1_kcbh
Not used
ub1 spare2_kcbh
Not used
ub4 rdba_kcbh
RDBA -Relative Data Block Address
ub4 bas_kcbh
SCN Base
ub2 wrp_kcbh
SCN Wrap
ub1 seq_kcbh
Sequence number, incremented for every change made to the block at the same SCN
ub1 flg_kcbh
Flag:
0x01 New Block
0x02 Delayed Logging Change advanced SCN/seq 0x04 Check value saved - block XOR‘s to zero
0x08 Temporary block
ub2 chkval_kcbh
Optional block checksum (if DB_BLOCK_CHECKSUM=TRUE)
ub2 spare3_kcbh
Not used
struct ktbbh, 72 bytes
Transaction Fixed Header Structure
ub1 ktbbhtyp
Block type (1=DATA, 2=INDEX)
union ktbbhsid, 4 bytes
Segment/Object ID
struct ktbbhcsc, 8 bytes
SCN at last block cleanout
b2 ktbbhict
Number of ITL slots
ub1 ktbbhflg
0=on the freelist
ub1 ktbbhfsl
ITL TX freelist slot
ub4 ktbbhfnx
DBA of next block on the freelist
struct ktbbhitl[2], 48 bytes
ITL list index
struct kdbh, 14 bytes
Data Header Structure
ub1 kdbhflag
N=pctfree hit(clusters); F=do not put on freelist; K=flushable cluster keys
b1 kdbhntab
Number of tables (>1 in clusters)
b2 kdbhnrow
Number of rows
sb2 kdbhfrre
First free row entry index; -1 = you have to add one
sb2 kdbhfsbo
Freespace begin offset
sb2 kdbhfseo
Freespace end offset
b2 kdbhavsp
Available space in the block
b2 kdbhtosp
Total available space when all TXs commit
struct kdbt[1], 4 bytes
Table Directory Entry Structure
b2 kdbtoffs
b2 kdbtnrow
sb2 kdbr[1]
Row Directory
ub1 freespace[8030]
Free Space
ub1 rowdata[38]
Row Data
ub4 tailchk
(See Tailchecks below)
Different block types are designated by the first byte of the block. The following tableshows how to decode the block type:
不同的block 可以第一個byte的值是不一樣的。 具體值對應block 類型如下。
Header Block Types
ID
Type
01
Undo segment header
02
Undo data block
03
Save undo header
04
Save undo data block
05
Data segment header (temp, index, data and so on)
06
KTB managed data block (with ITL)
07
Temp table data block (no ITL)
08
Sort Key
09
Sort Run
10
Segment free list block
11
Data file header
可以通過dump block來查看對應的具體的值。 下文講dump時會有相關的示例。
oracleblocks 的最後4個bytes 是tail check。 下面看一下oracle 9i block的tail check 組成。
Tailchecks
Thetail of an Oracle 8+ block is a concatenation of the lower order two bytes ofthe SCN base, the block type and the SCN sequence number.
Oracleblock tail 由4個bytes組成，但實際上只用了低2個bytes來存放。 2個bytes的tail 由scn base，block type 和 scn sequence 組成。
E.g, if the SCN base number is 0x00029728,the block type is 06 and the SCN sequence number is 0x02, the tail check wouldbe 0x97280602:
SCN base Type SCN seq
9728 06 02
Althoughthis tail check value is generated from three components, Oracle treats thefinal value as a single unsigned integer stored as a word (4-bytes). Onlittle-endian architecture machines, which include Intel, the value will bestores as low-order byte first.
雖然tail check 由3個部分組成，但是oracle 把這3部分作為一個整體來存儲，並且占用4個bytes。
對於little-endian（低端）架構的機器，包括Intel, 他們會先存放low-order byte，即低位字節。
Thereforeif the tail check is examined in the block using a standard block editor, orthe dump command which will be explained in the next section, the byte ordermay look different. A tail check of 0x97280602 stored on an Intel machine wouldbe written to disk as "02062897".
可以通過標准block editor 或者dump 來查看tail check。 對於不同的機器，他們存儲的順序是不一樣的。比如tail check 0x97280602 在Intel Machine 就被存儲為02062897，因為它會先保存low-order bytes。
The dump command dumps the content of the block to the screen. It can be combinedwith the /v option to produce a more verbose output. TheDBA, Filename, File, Block and/or Offset to dump can be specified with thecommand. If these are not specified the current file, block and offsetas established with the set command will be dumped. The size of the dump islimited by the set count option and defaults to 512 bytes or alternatively thesize of the dump can be specified with the command.
dump命令可以將block 的內容顯示到屏幕。 每次顯示的bytes由count 控制，默認是512 bytes。 使用 /v 選項，可以顯示更多詳細信息。
dump /v dba 4,145 offset 0 count 128
The examine command is used to display data from the block in raw or formattedoutput. The DBA, Filename, File, Block and/or Offset to examine can bespecified with the command. If these are not specified the current file, blockand offset as established with the set command will be examined. If the examinecommand is issued with just the block and offset to examine, bbed will displaythe data structure at that offset.
--examine命令也是用來顯示datablock的內容的。
Unlikethe print command it cannot interpret data structures, but it can be used todisplay row information. Combined with knowledge of the data type of the row,it can be used to retrieve complete rows from the block:
--print命令不能對data structures 進行一個解釋說明。
The examine command will interpret the data in the block according to the followingswitches:
Switch
Display Format
/b
b1, ub1 (byte)
/h
b2, ub2 (half-word)
/w
b4, ub4 (word)
/l
b8, ub8 (long) (was b4/ub4 in Oracle7).
/r
Oracle table/index row
Theexamine command allows switches from the print command to be combined withthese specific switches to interpret data.
--examine可以根據switch的方式和print 命令進行一個結合來對data 進行解釋說明。
Forexample if we wanted to interpret data as an Oracle table row with the firstcolumn character and the second and third columns numeric, we would execute thecommand as follows:
BBED>x /rcnn
Thefollowing example shows the print and examine commands being used to step throughthe first and second rows of a block, with the data interpreted as a row in theformat: character, number, number:
x /rcnn
select * from scott.emp;
The find command is used to locate data within a block. The command allows hex,string or numeric data to be searched for. The pattern can be searched for fromthe top of the block (offset 0) using the TOP directive, or from the currentposition using the CURR directive.
find命令可以用來搜索關鍵字。 可以從offset 0 搜索到top 或者從當前的offset 搜索到top。
Switchesare used to determine the data type of the pattern to search for. These areshown below:
Switch
Datatype
/x
Hexadecimal
/d
Decimal
/u
unsigned decimal
/o
Octal
/c
character (native)
Note: Number and Dates are not supported bythe find command.
find 命令支持的switch 類型如上表，注意，find 不支持number和Date 。
select rowid,
dbms_rowid.rowid_relative_fno(rowid)rel_fno,
dbms_rowid.rowid_block_number(rowid)blockno,
dbms_rowid.rowid_row_number(rowid) rowno
from scott.emp;
set file 4
set block 145
set offset 0
find /c SCOTT top
d /v dba 4,145 offset 7864 count 128
SCOTT
The copy command is used to copy blocks from one location to another. As with other commands, the file or filename and offset can be specified, or the DBA can be specified instead.
可以把一個塊的內容拷貝到另一個塊中。
命令格式如下：
BBED> copy dba 1,115362 to dba 1,115363
copy 是個危險的命令，慎用。
BBED> help modify
MODIFY[/x|d|u|o|c] numeric/character string
[ DBA | FILE | FILENAME | BLOCK | OFFSET | symbol | *symbol ]
BBED>
The modify command is used to change data inside a block. The DBA, Filename, File,Block and/or Offset to modify can be specified with the command. If these are not specified the current file, block and offset as established with the set command will be modified. Alternatively a symbol or symbol pointer can bespecified for modification.
The pattern of bytes used to overwrite the original can be specified inhexadecimal, decimal, unsigned decimal, octal or character data using the sameswitches as the find command.
可以修改塊中的內容，該步驟不演示，下面實驗有。
在file 1,block 115362 有我們的Dave，我們這裡把Dave 改成dmm。
modify /c HAHAH dba 4,31 offset 7864 該步驟不演示，下面實驗有。
d /v dba 4,31 offset 7864 count 128
select * from scott.emp;
現在無改變。
sum apply
重啟庫才能生效。
shutdown immediate;
startup
select * from scott.emp;
The assign command does symbolic assignment, with type and range checking. Either target or source can be omitted for the current offset.
For example, the following command assigns structure at current offset to file 4,block 2 ”s first ITL entry
BBED> assign dba 4, 145 ktbbhitl[0]
指定結構 ，不演示。
Thesum command is used to check and set the block checksum. The DBA, Filename,File, Block and/or Offset to check can be specified with the command. If theseare not specified the current file, block and offset as established with theset command will be checked.
Theapply directive can be used to update the checksum.
我們可以使用bbed 對block 進行修改。 要使這些修改生效，就要使用sum命令。
該步驟不演示，下面實驗有。
sum dba 4,31 apply 該步驟不演示，下面實驗有。
重啟庫或清除緩沖區才能生效。
Thepush and pop commands are used to push a file, block and offset location onto amemory backed stack and then pop them back. This allows a current locationbeing edited to be temporarily saved while another location is examined ormodified.
Note that the stack only stores the locationœ it does notsave the contents.
--push命令將對象放到內存的stack，pop 將對象從內存寫回磁盤。
Thefollowing example shows file 7, block 16, offset 8163 being examined. Thelocation is saved with the push command. We then move to file 6, block 1 beforereturning to DBA 7,16 with the pop command.
BBED> push dba 7,16
DBA 0x01c00010 (29360144 7,16)
OFFSET 8163
BBED> set dba 6,1
DBA 0x01800001 (25165825 6,1)
BBED> pop
DBA 0x01c00010 (29360144 7,16)
OFFSET 8163
The command pop all can be used to remove all push‘d entries from the stack. Thecommand show all can be used to show all saved locations.
Therevert command is used to restore a file, filename, block or DBA to it‘soriginal state when bbed was started.
revert是恢復自bbed 啟動以來的所有修改。
revert dba 4,31該步驟不演示，下面實驗有。
sum dba 4,31 apply
重啟庫才能看到變化。
shutdown immediate
startup
或者：
alter system flush buffer_cache;
select * from scott.emp;
還是沒有改變，flush 一下buffer cache之後就更改回來了。我們剛才在之前的測試時，是重啟了DB。 看來也是啟了flush buffer cache的作用。
Theundo command rolls back the last modify or assign command. If the undo commandis issued again the modification is re-done.
undo命令是回滾最後一次的操作。
該步驟不演示，下面實驗有。
modify /c HAHAH dba 4,31 offset 7864該步驟不演示，下面實驗有。
d /v dba 4,31 offset 7864 count 128
sum apply
一定要提交。
如果不提交之前可以
undo 該步驟不演示，下面實驗有。
alter system flush buffer_cache;
select * from scott.emp;
Theverify command is used to verify the integrity of the block. It performs asimilar function to the dbverify utility.
verify命令用來驗證block的完整性。
verify dba 4,145
查看有沒有壞塊
The corrupt command is used to mark blocks as media corrupt.
corrupt命令將一個block 標記為corrupt，這樣db 在操作時就會跳過該block，從而避免錯誤。
verify dba 4,31 該步驟不演示，下面實驗有。
corrupt dba 4,31該步驟不演示，下面實驗有。
verify dba 4,31
注意： undo 命令不能undo 一個corruption，但是revert 命令卻可以。
revert dba 4,31
sum apply
verify dba 4,31
本例綜合運用上面的各種命令，修改塊的內容，並撤銷修改。
在sqlplus中操作
select * from scott.emp;
現在無改變。
select rowid,
dbms_rowid.rowid_relative_fno(rowid)rel_fno,
dbms_rowid.rowid_block_number(rowid)blockno,
dbms_rowid.rowid_row_number(rowid) rowno
from scott.emp;
在bbed中操作
set dba 4,145 offset 0
find /c SCOTT top
dump /v dba 4,145 offset 7864 count 128
注意這裡面的Offsets:7864 to 7991， 它指的是這一行的一個地址。其中
S 的offset 是7864
C 的offset 是7865
O 的offset 是7866
T 的offset 是7867
空格也算offset。
modify /c HAHAH dba 4,145 offset 7864
sum apply
或者
sum dba 4,145 apply
dump /v dba 4,145 offset 7864 count 128
在sqlplus中操作
alter system flush buffer_cache;
select * from scott.emp;
在bbed中操作
revert dba 4,145
sum dba 4,145 apply
在sqlplus中操作
alter system flush buffer_cache;
select * from scott.emp;
When rows are deleted in Oracle the data is not actually removed. The row is simply marked as deleted and the free space counters and pointers adjusted accordingly. The status of a row is stored in the Row Header which occupies the first few bytes of each row.
當row 被delete 的時候，實際上data 並沒有被remove，只是將該row 標記為delete，然後其對應的空間被統計為free space。 row 的status 存在每個row的row header裡。
The Row Header consists of the Row Flag, Lock Byte (ITL entry) and Column Count.The first of these - the Row Flag - is a single byte that holds a bitmask thatshows the status of the row. The bitmask is decoded as follows:
RowHeader 包含Row Flag，Lock Byte（ITL）和column Count。其中Row Flag占用1個byte，並且以bitmask 來保存。bitmask 的解釋如下：
Cluster Key
Cluster Table Member
Head of row piece
Deleted
First data piece
Last data piece
1st Column continues from previous piece
Last column continues in next piece
128
64
32
16
8
4
2
1
在sqlplus中操作
select * from scott.emp;
select rowid,
dbms_rowid.rowid_relative_fno(rowid)rel_fno,
dbms_rowid.rowid_block_number(rowid)blockno,
dbms_rowid.rowid_row_number(rowid) rowno
from scott.emp;
col segment_name for a10
select extent_id,segment_name,bytes/1024 k,file_id,block_id from dba_extents where owner='SCOTT';
alter system checkpoint;
alter system dump datafile 4 block 145;
查詢dump到哪個trace文件中：
oradebug setmypid
oradebug tracefile_name
more /u01/app/oracle/oradata/PROD/dump/diag/rdbms/prod/PROD/trace/PROD_ora_6241.trc
可以看到這個表有14條記錄。
注意這裡的fb: --H-FL--。 其有8個選項，每個值分別與bitmask 對應。
Therefore,columns that fit within a single block, are not chained, migrated or part of aclustered table and are not deleted will have the following attributes:
（1）Head of Row Piece
（2）First Data Piece
（3）Last Data Piece
如果一個row 沒有被刪除，那麼它就具有上面的3個屬性，即Flag 表示為：--H-FL--. 這裡的字母分別代表屬性的首字母。其對應的值：32 + 8 + 4 =44 or 0x2c.
如果一個row 被delete了，那麼row flag 就會更新，bitmask 裡的deleted 被設置為16. 此時row flag 為： 32 + 16 + 8 + 4 = 60 or 0x3c.
delete from scott.emp where empno=7900;
commit;
select * from scott.emp;
alter system checkpoint;
alter system dump datafile 4 block 145;
查詢dump到哪個trace文件中：
oradebug setmypid
oradebug tracefile_name
more /u01/app/oracle/oradata/PROD/dump/diag/rdbms/prod/PROD/trace/PROD_ora_6241.trc
注意上面的標簽被刪除的數據是HDFL，一般是H-FL。
現在我們用bbed 將刪除的內容找回來。
在bbed中操作
set dba 4,145 offset 0
find /c JAMES top
dump /v dba 4,145 offset 7706 count 128
d /v dba 4,145 offset 7705 count 128
比剛剛多了兩個字符。
8個字符才是一個完整的信息，所以要改變4個偏移量才能展示完整的信息。
7706-8=7698，再加1=7699.
尋找原則：和row directory核對，尋找前面最接近的值。
d /v dba 4,145 offset 7699
這裡已經出現了我們3c（deleted）標志，但是注意這裡的位置的根據我們的查找的字符串來分的，實際在block裡的分割方式不一樣按照我們的offset 來進行。 我們可以通過row directory 來進行一個確認。
驗證一下
p kdbr
p *kdbr[9]
p *kdbr[10]
p *kdbr[11]
p *kdbr[12]
確定7699就是3c的開頭。
modify /x 2c offset 7699
sum apply
在sqlplus中操作
alter system flush buffer_cache;
select * from scott.emp;
之前delete 的數據已經恢復出來。
但是：
系統認為已經刪掉了。
select count(*) from scott.emp;
alter table scott.emp move;
將表scott.emp移動到我們的單獨的datafile裡(目的是查找清晰，並恢復上面的破壞)：
col tablespace_name for a15
select tablespace_name,contents,status from dba_tablespaces;
col name for a50
select file#,name from v$datafile;
create tablespace lxtbs datafile '/u01/app/oracle/oradata/PROD/disk3/lxtbs01.dbf' size 50m;
alter table scott.emp move tablespace lxtbs;
alter index scott.pk_emp rebuild tablespace lxtbs;
col table_name for a10
select table_name,tablespace_name,blocks from dba_tables where owner='SCOTT';
select * from scott.emp;
select file#||' '||name||' '||bytes from v$datafile;
shutdown immediate;
關庫後再操作：
cp /u01/app/oracle/oradata/PROD/disk3/lxtbs01.dbf /home/oracle
cd /home/oracle
ls
將copy 的bak datafile 添加到bbed 的parfile裡面
vi file.txt
添加：
6 /u01/app/oracle/oradata/PROD/disk3/lxtbs01.dbf 52428800
7 /home/oracle/lxtbs01.dbf 52428800
最後的7號文件是我們添加的。
startup
select rowid,
dbms_rowid.rowid_relative_fno(rowid)rel_fno,
dbms_rowid.rowid_block_number(rowid)blockno,
dbms_rowid.rowid_row_number(rowid) rowno
from scott.emp;
col owner for a10
col segment_name for a10
select owner,segment_name,header_file,header_block,blocks from dba_segments where owner='SCOTT';
從這個查詢結果，我們可以看到，對象保存在datafile 6裡，從130 的block 開始存儲，占用8個blocks。
這裡要注意的一點是：dba_segments 視圖裡的block 是從0開始的統計的，而bbed 裡是從1. 所以我們在bbed中指定block時，需要加1（130會報錯），也就是131
quit
bbed parfile=/home/oracle/bbed.par
blockedit
下面能查出東西來就行。
set dba 6,130 offset 0
p ktbbh
set dba 6,131 offset 0
p ktbbh
做一個誤刪除操作:
delete from scott.emp;
commit;
select * from scott.emp;
set width 70
info
其中7是我們的舊的數據文件。
copy dba 7,131 to dba 6,131
copy剩下的7個塊：
copy dba 7,132 to dba 6,132
copy dba 7,133 to dba 6,133
copy dba 7,134 to dba 6,134
copy dba 7,135 to dba 6,135
copy dba 7,136 to dba 6,136
copy dba 7,137 to dba 6,137
copy dba 7,138 to dba 6,138
sum apply
alter system flush buffer_cache;
select * from scott.emp;
如果出不來就再做一遍copy,數據就回來了，或者重啟db數據也能回來。
但是：
select count(*) from scott.emp;
需要再move一次。
alter table scott.emp move tablespace users;
alter index scott.pk_emp rebuild tablespace users;
alter table scott.emp move tablespace lxtbs;
alter index scott.pk_emp rebuild tablespace lxtbs;
select count(*) from scott.emp;
有些人喜歡用bbed干一些弄簡為繁的事情，比如控制文件丟失，或者介質故障開不了庫，手工修改scn號開庫，或壞塊修復等。如果不是高手，最好不要用bbed進行高風險的工作，其實Oracle Rman提供了修復壞塊的工具block recover。但前提條件是你得有一個可用的備份存在。
最好重新做一次備份：
rman target /
delete noprompt backup;
delete noprompt copy;
list backup;
list copy;
run{
shutdown immediate;
startup mount;
allocate channel c1 type disk;
allocate channel c2 type disk;
backup database format '/home/oracle/rman_bak/cold_bak/%d_%s_%p.bak';
alter database open;
}
run{
sql 'alter system switch logfile';
allocate channel c1 type disk;
allocate channel c2 type disk;
backup database format '/home/oracle/rman_bak/hot_bak/%d_%s_%p_hot.bak' filesperset 3;
sql 'alter system switch logfile';
}
/************************重新登錄*************************/
sqlplus重新登錄：
quit
sqlplus '/as sysdba'
bbed重新登錄
quit
bbed parfile=/home/oracle/bbed.par
blockedit
/************************重新登錄*************************/
先用bbed搞壞數據塊。
select rowid,
dbms_rowid.rowid_relative_fno(rowid)rel_fno,
dbms_rowid.rowid_block_number(rowid)blockno,
dbms_rowid.rowid_row_number(rowid) rowno
from scott.dept;
col segment_name for a10
select owner,segment_name,header_file,header_block,blocks from dba_segments where owner='SCOTT';
加一變成129.
set dba 4,128 offset 0
p ktbbh
set dba 4,129 offset 0
p ktbbh
map
d /v offset 0 count 128
modify /x 12345678 offset 0 單步執行，或copy命令制作壞塊
sum apply
d /v offset 0 count 128
alter system flush buffer_cache;
select * from scott.dept;
--當Oracle 認為一個block 是corrupt時，會將該block的sequence number 標記為0xff. 該值可以通過seq_kcbh 屬性查看。
set dba 4,129
p kcbh
verify dba 4,129
不建議使用bbed來操作，rman有工具可以很好的處理，bbed做的話就是revert回去就好了。
BBED> revert
All changes made in this session will be rolled back. Proceed? (Y/N) Y
Reverted file '/oracle/app/oracle/oralhr/users01.dbf', block 520
Reverted file '/oracle/app/oracle/oralhr/users01.dbf', block 523
Warning: contents of previous BIFILE will be lost. Proceed? (Y/N) Y
下邊來嘗試rman恢復：
rman target /
不要執行：只適用於一兩個壞塊 blockrecover datafile 4 block 129,130;
如果有多個壞塊，最好先校驗：
backup validate datafile 4;
select * from v$database_block_corruption;
修復全部壞塊。
blockrecover corruption list;
alter system flush buffer_cache;
select * from scott.dept;
用rman的備份，修復了這個塊。
select * from v$database_block_corruption;
set dba 4,129
d /v offset 0 count 128
再次搞壞數據塊：
set dba 4,129
d /v offset 0 count 128
modify /x 12345678 offset 0
sum apply
alter system flush buffer_cache;
select * from scott.dept;
rman內操作：
backup datafile 4;
RMAN> backup datafile 4;
Starting backup at 2016-04-06 14:25:42
using channel ORA_DISK_1
channel ORA_DISK_1: starting full datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
input datafile file number=00004 name=/oracle/app/oracle/oralhr/users01.dbf
channel ORA_DISK_1: starting piece 1 at 2016-04-06 14:25:42
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03009: failure of backup command on ORA_DISK_1 channel at 04/06/2016 14:25:45
ORA-19566: exceeded limit of 0 corrupt blocks for file /oracle/app/oracle/oralhr/users01.dbf
col tablespace_name for a10
col owner for a10
col owner for a10
col PARTITION_NAME for a10
SELECT tablespace_name,segment_type,owner,segment_name,partition_name
FROM dba_extents
WHERE file_id = &file_id AND &block_id BETWEEN block_id AND block_id+blocks-1;
輸入：
4
129
run{
set maxcorrupt for datafile 4 to 2;
backup datafile 4 tag='2corruptblock';
}
這樣才能備份，但是推薦先修復壞塊：
backup validate datafile 4;
blockrecover corruption list;
backup datafile 4;
About Me
..........................................................................................................................................................................................................
● 本文作者：小麥苗，只專注於數據庫的技術，更注重技術的運用
● 本文在ITpub（http://blog.itpub.net/26736162）、博客園(http://www.cnblogs.com/lhrbest)和個人微信公眾號（xiaomaimiaolhr）上有同步更新
● QQ群：230161599 微信群：私聊
● 本文itpub地址：http://blog.itpub.net/26736162/viewspace-2123465/ 本文博客園地址：http://www.cnblogs.com/lhrbest/p/5771282.html
● 本文pdf版：http://yunpan.cn/cdEQedhCs2kFz （提取碼：ed9b）
● 小麥苗分享的其它資料：http://blog.itpub.net/26736162/viewspace-1624453/
● 聯系我請加QQ好友(642808185)，注明添加緣由
● 於 2016-08-01 15:00~ 2016-08-01 19:00 在中行完成
● 【版權所有，文章允許轉載，但須以鏈接方式注明源地址，否則追究法律責任】
..........................................................................................................................................................................................................
長按下圖識別二維碼或微信客戶端掃描下邊的二維碼來關注小麥苗的微信公眾號：xiaomaimiaolhr,學習最實用的數據庫技術。
