Oracle Dump Redo Log File

aishu521發表於2013-03-15

一. dump redo 說明
            關於Dump redo log 的示例，MOS 上的文件：[ID 1031381.6] 有詳細說明。Dump 有兩種方式：
            （1）使用'alter session' 命令dumpredo header。
            （2）使用'alter system dump logfile' 命令dump logfile contents。

可以在mount，nomout和open狀態下使用如上命令，並且使用以上命令需要'alter system' 的許可權。可以使用該命令dump onlie redo log 或者歸檔檔案。當作業系統系統相同的情況下，還可以dump其他db的log 檔案。

有如下dump 方法：
            1.To dump records based in DBA (Data Block Address)
            2.To dump records based on RBA (Redo Block Address)
            3.To dump records based on SCN
            4.To dump records based on time
            5.To dump records based on layer and opcode
            6.Dump the file header information
            7.Dump an entire log file:
二. 具體使用示例
2.1 To dump records based on DBA (Data Block Address)
關於DBA的說明，參考：
            Oracle rdba和 dba 說明
            http://blog.csdn.net/tianlesoftware/article/details/6529346
            根據DBA進行dump，主要是根據file和block 號來進行dump。這個的block 是一個範圍值。
11g命令格式如下：
ALTER SYSTEM DUMP LOGFILE 'filename' DBA MIN fileno . blockno    DBA MAX fileno . blockno;

如果是Oracle 10g，需要省略'.'，即格式如下：

ALTER SYSTEM DUMP LOGFILE 'filename' DBA MIN fileno blockno    DBA MAX fileno blockno;
否則會報： ORA-01963: Must specify a block number 錯誤。
如：
SYS@anqing1(rac1)> select * fromv$version where rownum=1;
BANNER
----------------------------------------------------------------
Oracle Database 10g Enterprise EditionRelease 10.2.0.4.0 - Prod
SQL>select distinct dbms_rowid.rowid_relative_fno(rowid) rel_fno,
dbms_rowid.rowid_block_number(rowid)blockno from ta;
REL_FNO    BLOCKNO
---------- ----------
        1     294608
        1     294609
        1     294612
        1     294614
        1     294621
        1     294635
        1     294643
        1     294654
        1     294656
        1     294657
        6      10385
REL_FNO    BLOCKNO
---------- ----------
        6      10393
        6      10410
         6     10411
        6      10415
        6      10416
        6      10458
        6      10461
        6      10464
        6      10465
        6      10468
        6      10472

確定log file 位置：
SYS@anqing1(rac1)> selecta.group#,a.status,b.member from v$log a,v$logfile b where a.group#=b.group#;
   GROUP# STATUS           MEMBER
---------- --------------------------------------------------------------------
        3 INACTIVE         +DATA/anqing/onlinelog/redo03.log
        2 INACTIVE        +DATA/anqing/onlinelog/redo02.log
        1 CURRENT         +DATA/anqing/onlinelog/group_1.277.751552735
        1 CURRENT         +FRA/anqing/onlinelog/group_1.426.751552739
        4 CURRENT         +DATA/anqing/onlinelog/group_4.282.751560131
        5 INACTIVE        +DATA/anqing/onlinelog/group_5.283.751560139
6 rows selected.
SYS@anqing1(rac1)> alter system dump logfile'+data/anqing/onlinelog/redo02.log' dbamin 6 10458    dba max 6 10472;
System altered.
SYS@anqing1(rac1)> oradebug setmypid
Statemnt processed.
SYS@anqing1(rac1)> oradebug tracefile_name
/u01/app/oracle/admin/anqing/udump/anqing1_ora_30373.trc
SYS@anqing1(rac1)>
[oracle@rac1 ~]$cat /u01/app/oracle/admin/anqing/udump/anqing1_ora_30373.trc
/u01/app/oracle/admin/anqing/udump/anqing1_ora_30373.trc
Oracle Database 10g Enterprise EditionRelease 10.2.0.4.0 - Production
With the Partitioning, Real ApplicationClusters, OLAP, Data Mining
and Real Application Testing options
ORACLE_HOME =/u01/app/oracle/product/10.2.0/db_1
System name:    Linux
Node name:      rac1
Release:        2.6.18-194.el5
Version:        #1 SMP Tue Mar 16 21:52:43 EDT 2010
Machine:        i686
Instance name: anqing1
Redo thread mounted by this instance: 1
Oracle process number: 28
Unix process pid: 30373, image: oracle@rac1(TNS V1-V3)
*** 2011-08-08 22:10:37.053
*** ACTION NAME:() 2011-08-08 22:10:37.052
*** MODULE NAME:(sqlplus@rac1 (TNS V1-V3))2011-08-08 22:10:37.052
*** SERVICE NAME:(SYS$USERS) 2011-08-0822:10:37.052
*** SESSION ID:(127.27077) 2011-08-0822:10:37.052
DUMP OF REDO FROM FILE'+data/anqing/onlinelog/redo02.log'
Opcodes *.*
DBAs: (file # 6, block # 10458) thru (file #6, block # 10472)
RBAs: 0x000000.00000000.0000 thru0xffffffff.ffffffff.ffff
SCNs: scn: 0x0000.00000000 thru scn:0xffff.ffffffff
Times: creation thru eternity
FILEHEADER:
       Compatibility Vsn = 169870336=0xa200400
       Db ID=715014091=0x2a9e3fcb, Db Name='ANQING'
       Activation ID=682502900=0x28ae2af4
       Control Seq=4261=0x10a5, File size=102400=0x19000
       File Number=2, Blksiz=512, File Type=2 LOG
descrip:"Thread0001, Seq# 0000000119, SCN 0x0000006ab1bc-0x0000006e0c84"
thread: 1 nab: 0x18ffd seq: 0x00000077 hws:0x2 eot: 0 dis: 0
resetlogs count: 0x2ccbc8df scn:0x0000.000704fa (460026)
resetlogs terminal rcv count: 0x0 scn:0x0000.00000000
prevresetlogs count: 0x2ebbfe8f scn: 0x0000.0006ce7b (446075)
prevresetlogs terminal rcv count: 0x0 scn: 0x0000.00000000
Low scn: 0x0000.006ab1bc (6992316) 07/30/2011 05:39:17
Nextscn: 0x0000.006e0c84 (7212164) 08/03/2011 14:14:34
Enabled scn: 0x0000.000704fa (460026)05/19/2011 12:38:55
Thread closed scn: 0x0000.006ab1bc (6992316)07/30/2011 05:39:17
Diskcksum: 0xd530 Calc cksum: 0xd530
Terminal recovery stop scn: 0x0000.00000000
Terminal recovery 01/01/1988 00:00:00
Mostrecent redo scn: 0x0000.00000000
Largest LWN: 1874 blocks
End-of-redo stream : No
Unprotected mode
Miscellaneous flags: 0x0
Thread internal enable indicator: thr: 0, seq:0 scn: 0x0000.00000000
REDO RECORD - Thread:1 RBA:0x000077.00000cd6.017c LEN: 0x0048 VLD: 0x01
SCN: 0x0000.006ac8a9 SUBSCN: 1 07/30/2011 08:28:08
CHANGE #1 MEDIA RECOVERY MARKERSCN:0x0000.00000000 SEQ: 0 OP:18.3
Reuse redo entry
Range reuse: tsn=1 base=8542185 nblks=8
REDO RECORD - Thread:1 RBA:0x000077.00006e99.0078 LEN: 0x0044 VLD: 0x01
SCN: 0x0000.006b7ae1 SUBSCN: 1 07/31/2011 06:18:10
CHANGE #1 MEDIA RECOVERY MARKERSCN:0x0000.00000000 SEQ: 0 OP:17.4
Datafile resize marker - file: 2 old size:153600 new size: 156160
REDO RECORD - Thread:1 RBA:0x000077.00007884.017c LEN: 0x0048 VLD: 0x01
SCN: 0x0000.006b8d16 SUBSCN: 1 07/31/2011 08:33:09
CHANGE #1 MEDIA RECOVERY MARKERSCN:0x0000.00000000 SEQ: 0 OP:18.3
Reuse redo entry
Range reuse: tsn=1 base=8542217 nblks=1024
*** 2011-08-08 22:10:54.783
END OF REDO DUMP
----- Redo read statistics for thread 1-----
Read rate (ASYNC): 51197Kb in 17.78s =>2.81 Mb/sec
Total physical reads: 51197Kb
Longest record: 8Kb, moves: 0/85186 (0%)
Change moves: 30223/143291 (21%), moved:5Mb
Longest LWN: 937Kb, moves: 6/66426 (0%),moved: 0Mb
Last redo scn: 0x0000.006e087b (7211131)
----------------------------------------------
[oracle@rac1 ~]$
2.2. To dumprecords based on RBA (Redo Block Address)
            在第一節裡是根據日誌的block number進行dump的，也可以根據日誌的sequence#進行dump。
格式如下：
ALTER SYSTEM DUMP LOGFILE 'filename' RBA MIN seqno .blockno RBA MAX seqno . blockno;
SYS@anqing1(rac1)> alter system dump logfile'+data/anqing/onlinelog/redo02.log' rbamin 121 10458 rba max 125 10472;
System altered.
其他資訊的檢視參考第一節。
2.3. To dumprecords based on SCN        可以直接根據SCN 的值來dump redo log。

格式如下：
ALTER SYSTEM DUMP LOGFILE 'filename' SCN MIN minscn SCNMAX maxscn;
SYS@anqing1(rac1)> select current_scn from v$database;
CURRENT_SCN
-----------
7479813

SYS@anqing1(rac1)> alter system dump logfile'+data/anqing/onlinelog/redo02.log' scn min 7479801 scn max 7479813;
System altered.
也可以使用該命令來檢查dumpfile，如：
SQL> ALTER SYSTEM DUMP LOGFILE 'filename' SCN MIN 1SCN MAX 1;
如果以上語句成功執行，則archivelog 沒有問題。
2.4. To dumprecords based on time.
根據時間來dump。
格式如下：
ALTER SYSTEM DUMP LOGFILE 'filename' TIME MIN value TIMEMAX value;

如：
ALTER SYSTEM DUMP LOGFILE '+data/anqing/onlinelog/redo02.log' TIME MIN 299425687 TIMEMAX 299458800;
注意這裡的時間是用redo dump time。關於redo time 和date 的轉換，網上有個指令碼。據說這個指令碼出自metalink，但是我在MOS 上搜了一下，沒有看到這個指令碼的內容。

執行SQL> select scn,to_char(time_dp,'YYYY-MM-DD HH24:MI:SS') from sys.smon_scn_time
2 where to_char(time_dp,'YYYY-MM-DDHH24')='2011-08-2616';
SCN和TIMESTAMP抓換函式
timestamp_to_scn(sysdata) ,scntotimestamp(scn)

指令碼1： date to redodump time
/* Formatted on2011/8/8 23:00:53 (QP5 v5.163.1008.3004) */
SET ECHO OFF
REM NAME:   TFSTM2RD.SQL
REMUSAGE:"@path/tfstm2rd"
REM------------------------------------------------------------------------
REM REQUIREMENTS:
REM   None.
REM------------------------------------------------------------------------
REM AUTHOR:
REM   Anonymous
REM   Copyright 1996, Orqacle Corporation
REM------------------------------------------------------------------------
REM PURPOSE:
REM   Converts a standard date into redo dump timeformat.
REM------------------------------------------------------------------------
REM EXAMPLE:
REM   Enter day (DD/MM/YYYY) ?08/07/1996
REM   Enter time (HH24:MI:SS) ? 12:05:05
REM
REM   REDO_YEAR REDO_MONTH REDO_DAY REDO_HOURREDO_MIN REDO_SEC
REM   --------- ---------- -------- ----------------- --------
REM        1996       7      8       12       5      5
REM
REM   EDO_TIME
REM   ----------
REM    273845105
REM
REM------------------------------------------------------------------------
REM DISCLAIMER:
REM   This script. is provided for educational purposesonly. It is NOT
REM   supported by Oracle World Wide TechnicalSupport.
REM   The script. has been tested and appears towork as intended.
REM   You should always run new scripts on a testinstance initially.
REM------------------------------------------------------------------------
REM Main text ofscript. follows:
UNDEFINE redo_day
UNDEFINE redo_hhmiss
ACCEPT redo_day PROMPT "Enter day (DD/MM/YYYY) ? "
ACCEPT redo_hhmiss PROMPT "Enter time (HH24:MI:SS) ? "
COLUMN redo_year NEW_VALUE redo_year FORMAT 9999
COLUMN redo_month NEW_VALUE redo_month FORMAT 9999
COLUMN redo_day NEW_VALUE redo_day FORMAT9999
COLUMN redo_hour NEW_VALUE redo_hour FORMAT 9999
COLUMN redo_min NEW_VALUE redo_min FORMAT9999
COLUMN redo_sec NEW_VALUE redo_sec FORMAT9999
COLUMN redo_time NEW_VALUE redo_time
SETVERIFY OFF
SELECT TO_NUMBER (
          TO_CHAR (
             TO_DATE ('&redo_day&redo_hhmiss', 'DD/MM/YYYY HH24:MI:SS'),
             'YYYY'))
         redo_year,
       TO_NUMBER (     TO_CHAR (