ORACLE 8.1.7 数据库ORA-600 4194故障恢复

发表于 2014 年 7 月 28 日 由 惜分飞

一个817数据库报ORA-600 4194 无法正常启动

Fri Jul 25 10:49:47 2014
Database mounted in Exclusive Mode.
Completed: ALTER DATABASE   MOUNT
Fri Jul 25 10:49:58 2014
ALTER DATABASE RECOVER  database  
Fri Jul 25 10:49:58 2014
Media Recovery Start
Media Recovery Log 
Recovery of Online Redo Log: Thread 1 Group 2 Seq 3320 Reading mem 0
  Mem# 0 errs 0: D:\ORACLE\ORADATA\ORCL\REDO02.LOG
Media Recovery Complete
Completed: ALTER DATABASE RECOVER  database  
Fri Jul 25 10:50:09 2014
alter database open

Beginning crash recovery of 1 threads
Fri Jul 25 10:50:09 2014
Thread recovery: start rolling forward thread 1
Recovery of Online Redo Log: Thread 1 Group 2 Seq 3320 Reading mem 0
  Mem# 0 errs 0: D:\ORACLE\ORADATA\ORCL\REDO02.LOG
Fri Jul 25 10:50:09 2014
Thread recovery: finish rolling forward thread 1
Thread recovery: 0 data blocks read, 0 data blocks written, 3 redo blocks read
Crash recovery completed successfully
Fri Jul 25 10:50:09 2014
Thread 1 advanced to log sequence 3321
Thread 1 opened at log sequence 3321
  Current log# 3 seq# 3321 mem# 0: D:\ORACLE\ORADATA\ORCL\REDO01.LOG
Successful open of redo thread 1.
Fri Jul 25 10:50:09 2014
SMON: enabling cache recovery
Fri Jul 25 10:50:10 2014
Errors in file D:\oracle\admin\ORCL\udump\ORA03216.TRC:
ORA-00600: ??????????: [4194], [12], [37], [], [], [], [], []

Fri Jul 25 10:50:10 2014
Recovery of Online Redo Log: Thread 1 Group 3 Seq 3321 Reading mem 0
  Mem# 0 errs 0: D:\ORACLE\ORADATA\ORCL\REDO01.LOG
Fri Jul 25 10:50:10 2014
SMON: disabling cache recovery
Fri Jul 25 10:50:10 2014
ORA-600 signalled during: alter database open

ORA-600 4194这个错误在数据库异常恢复中非常常见，因为库不是很重要,因此就是直接屏蔽掉故障回滚段,然后强制拉库,该库的恢复过程中,也直接使用隐含参数屏蔽回滚段
_corrupted_rollback_segments= RBS0, RBS1, RBS2, RBS3, RBS4, RBS5, RBS6, RBS_HDSYS,数据库依然无法open,进一步分析trace文件

Fri Jul 25 11:26:07 2014
ORACLE V8.1.7.0.0 - Production vsnsta=0
vsnsql=e vsnxtr=3
Windows 2000 Version 5.2 Service Pack 2, CPU type 586
Oracle8i Release 8.1.7.0.0 - Production
JServer Release 8.1.7.0.0 - Production
Windows 2000 Version 5.2 Service Pack 2, CPU type 586
Instance name: orcl

Redo thread mounted by this instance: 1

Oracle process number: 14

Windows thread id: 3648, image: ORACLE.EXE


*** SESSION ID:(11.1) 2014-07-25 11:26:07.843
*** 2014-07-25 11:26:07.843
ksedmp: internal or fatal error
ORA-00600: ??????????: [4194], [12], [37], [], [], [], [], []
Current SQL statement for this session:
update undo$ set name=:2,file#=:3,block#=:4,status$=:5,user#=:6,undosqn=:7,xactsqn=:8,
scnbas=:9,scnwrp=:10,inst#=:11,ts#=:12 where us#=:1
----- Call Stack Trace -----

这里很明显看出来，数据库是在open过程中，update undo$表遭遇到ORA-600 4194,因为该过程需要使用系统回滚段，但是由于其所对应的undo和redo信息不一致，所以无法正常启动数据库.继续读trace文件

  Extent Control Header
  -----------------------------------------------------------------
  Extent Header:: spare1: 0      space2: 0      #extents: 5      #blocks: 49    
                  last map  0x00000000  #maps: 0      offset: 4128  
      Highwater::  0x00400006  ext#: 0      blk#: 3      ext size: 9     
  #blocks in seg. hdr's freelists: 0     
  #blocks below: 0     
  mapblk  0x00000000  offset: 0     
                   Unlocked
     Map Header:: next  0x00000000  #extents: 5    obj#: 0      flag: 0x40000000
  Extent Map
  -----------------------------------------------------------------
   0x00400003  length: 9     
   0x0040000c  length: 10    
   0x0040008f  length: 10    
   0x00400099  length: 10    
   0x004000a3  length: 10    
  
  TRN CTL:: seq: 0x003c chd: 0x004e ctl: 0x0050 inc: 0x00000000 nfb: 0x0000
            mgc: 0x8002 xts: 0x0068 flg: 0x0001 opt: 2147483646 (0x7ffffffe)
            uba: 0x00400006.003c.25 scn: 0x0000.009a4009
Version: 0x01
  FREE BLOCK POOL::
    uba: 0x00000000.003c.24 ext: 0x0  spc: 0x196   
    uba: 0x00000000.001f.14 ext: 0x1  spc: 0x16f6  
    uba: 0x00000000.0018.02 ext: 0x4  spc: 0x1f1a  
    uba: 0x00000000.0000.00 ext: 0x0  spc: 0x0     
    uba: 0x00000000.0000.00 ext: 0x0  spc: 0x0     
  TRN TBL::

通过这里可以看出来,数据库在启动的时候，使用system undo的block为为0×00400006,使用bbed清除掉该uba记录,让数据库启动的时候重新分配system undo block给数据库执行update undo$使用,数据库open成功

BBED> m /x 0x00000000
 File: D:\ORACLE\ORADATA\ORCL\SYSTEM01.DBF (0)
 Block: 2                Offsets: 4188 to 4192           Dba:0x00000000
------------------------------------------------------------------------
 00000000 3c002400 00009601 00000000 1f001400 0100f616 00000000 18000200

BBED> m /x 0x0000
 File: D:\ORACLE\ORADATA\ORCL\SYSTEM01.DBF (0)
 Block: 2                Offsets: 4028 to 4032           Dba:0x00000000
------------------------------------------------------------------------
 00000000 00000000 3c005000 02800100 68000000 feffff7f 06004000 3c002400

Sat Jul 26 12:09:21 2014
Thread recovery: start rolling forward thread 1
Recovery of Online Redo Log: Thread 1 Group 2 Seq 3326 Reading mem 0
  Mem# 0 errs 0: D:\ORACLE\ORADATA\ORCL\REDO02.LOG
Sat Jul 26 12:09:21 2014
Thread recovery: finish rolling forward thread 1
Thread recovery: 0 data blocks read, 0 data blocks written, 3 redo blocks read
Crash recovery completed successfully
Sat Jul 26 12:09:22 2014
Thread 1 advanced to log sequence 3327
Thread 1 opened at log sequence 3327
  Current log# 3 seq# 3327 mem# 0: D:\ORACLE\ORADATA\ORCL\REDO01.LOG
Successful open of redo thread 1.
Sat Jul 26 12:09:22 2014
SMON: enabling cache recovery
SMON: enabling tx recovery
Sat Jul 26 12:09:39 2014
Completed: alter database open
发表在 非常规恢复 | 标签为 | 评论关闭

数据库启动ORA-08103故障恢复

发表于 2014 年 7 月 21 日 由 惜分飞

数据库在open过程报ORA-08103错误导致数据库无法正确启动

Fri Jul 18 22:02:51 2014
SMON: enabling tx recovery
Fri Jul 18 22:02:51 2014
Errors in file d:\oracle\product\10.2.0\admin\kemu3\udump\kemu3_ora_29788.trc:
ORA-00604: ?? SQL ?? 1 ????
ORA-08103: ??????

Fri Jul 18 22:02:51 2014
Database Characterset is ZHS16GBK
Fri Jul 18 22:02:51 2014
Errors in file d:\oracle\product\10.2.0\admin\kemu3\bdump\kemu3_smon_29704.trc:
ORA-00604: error occurred at recursive SQL level 1
ORA-08103: object no longer exists

Fri Jul 18 22:02:51 2014
Errors in file d:\oracle\product\10.2.0\admin\kemu3\bdump\kemu3_smon_29704.trc:
ORA-00604: error occurred at recursive SQL level 1
ORA-08103: object no longer exists

Fri Jul 18 22:02:51 2014
Errors in file d:\oracle\product\10.2.0\admin\kemu3\bdump\kemu3_smon_29704.trc:
ORA-00604: error occurred at recursive SQL level 1
ORA-08103: object no longer exists

Fri Jul 18 22:02:52 2014
Errors in file d:\oracle\product\10.2.0\admin\kemu3\bdump\kemu3_smon_29704.trc:
ORA-00604: error occurred at recursive SQL level 1
ORA-08103: object no longer exists

replication_dependency_tracking turned off (no async multimaster replication found)
Starting background process QMNC
QMNC started with pid=18, OS id=29876
Fri Jul 18 22:02:53 2014
Errors in file d:\oracle\product\10.2.0\admin\kemu3\bdump\kemu3_smon_29704.trc:
ORA-00604: error occurred at recursive SQL level 1
ORA-08103: object no longer exists

Fri Jul 18 22:02:54 2014
ORA-604 signalled during: alter database open...

对数据库启动过程做10046

PARSING IN CURSOR #22 len=210 dep=2 uid=0 oct=3 lid=0 tim=20960424464 hv=864012087 ad='3063f0b4'
select /*+ rule */ bucket_cnt, row_cnt, cache_cnt, null_cnt, timestamp#, sample_size, minimum, 
maximum, distcnt, lowval, hival, density, col#, spare1, spare2, avgcln from hist_head$ where obj#=:1 and intcol#=:2
END OF STMT
EXEC #22:c=0,e=80,p=0,cr=0,cu=0,mis=0,r=0,dep=2,og=3,tim=20960424461
WAIT #22: nam='db file sequential read' ela= 5452 file#=1 block#=60213 blocks=1 obj#=4586 tim=20960429962
FETCH #22:c=0,e=5967,p=1,cr=1,cu=0,mis=0,r=0,dep=2,og=3,tim=20960430462
*** KEWUXS - encountered error: (ORA-00604: 递归 SQL 级别 2 出现错误
ORA-08103: 对象不再存在
)  
*** kewrwdbi_1: Error=13515 encountered during run_once
BINDS #21:
kkscoacd
 Bind#0
  oacdty=02 mxl=22(22) mxlc=00 mal=00 scl=00 pre=00
  oacflg=08 fl2=0001 frm=00 csi=00 siz=24 off=0
  kxsbbbfp=0a217744  bln=22  avl=01  flg=05
  value=0
 Bind#1
  oacdty=01 mxl=32(20) mxlc=00 mal=00 scl=00 pre=00
  oacflg=18 fl2=0001 frm=01 csi=852 siz=32 off=0
  kxsbbbfp=0a217718  bln=32  avl=20  flg=05
  value="WRI$_ADV_DEFINITIONS"
 Bind#2
  oacdty=02 mxl=22(22) mxlc=00 mal=00 scl=00 pre=00
  oacflg=08 fl2=0001 frm=00 csi=00 siz=24 off=0
  kxsbbbfp=0a2176f4  bln=24  avl=02  flg=05
  value=1

这里很明显数据库启动过程,由于hist_head$的file 1 block 60213中的object_id 与 data_object_id 不匹配,从而出现ORA-08103错误,导致数据库无法正常启动,这里的故障的对象为hist_head$,非oracle核心对象,因此直接标记该block 为坏块(模拟普通ORA-08103并解决,模拟极端ORA-08103并解决,rman制造坏块,bbed修复坏块,bbed破坏数据文件),然后启动数据库,备份hist_head$表数据,然后truncate hist_head$,再插入hist_head$,整体完工.
在数据库open过程中,如果遇到ora-8103错误,导致数据库无法正常open,可以对其做10046定位到故障block和对象,然后判断对象是否数据库启动必须的对象,甚至是bootstarp$中对象,然后采取不同的处理方法.

发表在 Oracle备份恢复 | 标签为 | 评论关闭

数据文件的三个创建SCN一点点探讨

发表于 2014 年 7 月 20 日 由 惜分飞

在给一个朋友数据库恢复的过程中语句该库大量删除表空间,然后创建表空,由于在创建控制文件的时候,列出来不正确文件,导致出现v$datafile_header.error出现WRONG FILE CREATE错误.通过试验重现了该错误,并且进一步测试如果真的需要历史数据文件,该如何狸猫换太子(本实验为了进一步理解数据文件创建scn相关信息)
创建xifenfei表空间,然后删除表空间,但不删除数据文件,然后创建重名表空间

SQL> select to_char(sysdate,'yyyy-mm-dd hh24:mi:ss') today,'www.xifenfei.com' xifenfei from dual;

TODAY               XIFENFEI
------------------- ----------------
2014-07-16 15:54:26 www.xifenfei.com


SQL> create tablespace xifenfei datafile '/u01/app/oracle/oradata/ORCL/xifenfei_old.dbf' size 10m;

Tablespace created.

SQL> select file#,name from v$datafile;

     FILE# NAME
---------- --------------------------------------------------
         1 /u01/app/oracle/oradata/ORCL/system01.dbf
         2 /u01/app/oracle/oradata/ORCL/sysaux01.dbf
         3 /u01/app/oracle/oradata/ORCL/undotbs01.dbf
         4 /u01/app/oracle/oradata/ORCL/users01.dbf
         5 /u01/app/oracle/oradata/ORCL/xifenfei_old.dbf

SQL> select file#,CREATION_CHANGE#,to_char(CREATION_TIME,'yyyy-mm-dd hh24:mi:ss') CREATION_TIME from v$datafile;

     FILE# CREATION_CHANGE# CREATION_TIME
---------- ---------------- -------------------
         1               18 2014-07-14 21:53:05
         2             2338 2014-07-14 21:53:42
         3             3130 2014-07-14 21:53:51
         4            15268 2014-07-14 21:54:25
         5           593520 2014-07-16 16:00:54

SQL> select file#,CREATION_CHANGE#,to_char(CREATION_TIME,'yyyy-mm-dd hh24:mi:ss') CREATION_TIME from v$datafile_header;

     FILE# CREATION_CHANGE# CREATION_TIME
---------- ---------------- -------------------
         1               18 2014-07-14 21:53:05
         2             2338 2014-07-14 21:53:42
         3             3130 2014-07-14 21:53:51
         4            15268 2014-07-14 21:54:25
         5           593520 2014-07-16 16:00:54

SQL> drop tablespace xifenfei;

Tablespace dropped.

SQL>  create tablespace xifenfei datafile '/u01/app/oracle/oradata/ORCL/xifenfei_new.dbf' size 10m;

Tablespace created.

SQL> select file#,CREATION_CHANGE#,to_char(CREATION_TIME,'yyyy-mm-dd hh24:mi:ss') CREATION_TIME from v$datafile;

     FILE# CREATION_CHANGE# CREATION_TIME
---------- ---------------- -------------------
         1               18 2014-07-14 21:53:05
         2             2338 2014-07-14 21:53:42
         3             3130 2014-07-14 21:53:51
         4            15268 2014-07-14 21:54:25
         5           593613 2014-07-16 16:02:45

SQL> select file#,CREATION_CHANGE#,to_char(CREATION_TIME,'yyyy-mm-dd hh24:mi:ss') CREATION_TIME from v$datafile_header;

     FILE# CREATION_CHANGE# CREATION_TIME
---------- ---------------- -------------------
         1               18 2014-07-14 21:53:05
         2             2338 2014-07-14 21:53:42
         3             3130 2014-07-14 21:53:51
         4            15268 2014-07-14 21:54:25
         5           593613 2014-07-16 16:02:45

rename xifenfei表空间数据文件到老数据文件

SQL> alter database datafile 5 offline drop;

Database altered.

SQL> alter database rename file '/u01/app/oracle/oradata/ORCL/xifenfei_new.dbf'
 2   to '/u01/app/oracle/oradata/ORCL/xifenfei_old.dbf';     

Database altered.

SQL> alter database datafile 5 online;
alter database datafile 5 online
*
ERROR at line 1:
ORA-01122: database file 5 failed verification check
ORA-01110: data file 5: '/u01/app/oracle/oradata/ORCL/xifenfei_old.dbf'
ORA-01203: wrong incarnation of this file - wrong creation SCN


SQL> select file#,CREATION_CHANGE#,to_char(CREATION_TIME,'yyyy-mm-dd hh24:mi:ss') CREATION_TIME from v$datafile;

     FILE# CREATION_CHANGE# CREATION_TIME
---------- ---------------- -------------------
         1               18 2014-07-14 21:53:05
         2             2338 2014-07-14 21:53:42
         3             3130 2014-07-14 21:53:51
         4            15268 2014-07-14 21:54:25
         5           593613 2014-07-16 16:02:45

SQL> select file#,CREATION_CHANGE#,to_char(CREATION_TIME,'yyyy-mm-dd hh24:mi:ss') CREATION_TIME from v$datafile_header;

     FILE# CREATION_CHANGE# CREATION_TIME
---------- ---------------- -------------------
         1               18 2014-07-14 21:53:05
         2             2338 2014-07-14 21:53:42
         3             3130 2014-07-14 21:53:51
         4            15268 2014-07-14 21:54:25
         5           593520 2014-07-16 16:00:54

SQL> select file#,error from v$datafile_header;

     FILE# ERROR
---------- -----------------------------------------------------------------
         1
         2
         3
         4
         5 WRONG FILE CREATE

至此今天数据库恢复的故障已经模拟出来,就是因为数据文件头的scn和控制文件中scn不一致,从而出现了v$datafile_header.error报WRONG FILE CREATE的现象.

因为控制文件中数据文件scn和数据文件头scn不一致,因此通过重建控制文件来实现两者scn一致

SQL> alter database backup controlfile to trace as '/tmp/ctl';

Database altered.

SQL> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
SQL> STARTUP NOMOUNT
ORACLE instance started.

Total System Global Area  718225408 bytes
Fixed Size                  2292432 bytes
Variable Size             373294384 bytes
Database Buffers          339738624 bytes
Redo Buffers                2899968 bytes
SQL> CREATE CONTROLFILE REUSE DATABASE "ORCL" NORESETLOGS  NOARCHIVELOG
  2      MAXLOGFILES 16
  3      MAXLOGMEMBERS 3
  4      MAXDATAFILES 100
  5      MAXINSTANCES 8
  6      MAXLOGHISTORY 292
  7  LOGFILE
  8    GROUP 1 '/u01/app/oracle/oradata/ORCL/redo01.log'  SIZE 50M BLOCKSIZE 512,
  9    GROUP 2 '/u01/app/oracle/oradata/ORCL/redo02.log'  SIZE 50M BLOCKSIZE 512,
 10    GROUP 3 '/u01/app/oracle/oradata/ORCL/redo03.log'  SIZE 50M BLOCKSIZE 512
 11  DATAFILE
 12    '/u01/app/oracle/oradata/ORCL/system01.dbf',
 13    '/u01/app/oracle/oradata/ORCL/sysaux01.dbf',
 14    '/u01/app/oracle/oradata/ORCL/undotbs01.dbf',
 15    '/u01/app/oracle/oradata/ORCL/users01.dbf',
 16    '/u01/app/oracle/oradata/ORCL/xifenfei_old.dbf'
 17  CHARACTER SET ZHS16GBK
 18  ;

Control file created.

SQL> select file#,CREATION_CHANGE#,to_char(CREATION_TIME,'yyyy-mm-dd hh24:mi:ss') CREATION_TIME from v$datafile_header;

     FILE# CREATION_CHANGE# CREATION_TIME
---------- ---------------- -------------------
         1               18 2014-07-14 21:53:05
         2             2338 2014-07-14 21:53:42
         3             3130 2014-07-14 21:53:51
         4            15268 2014-07-14 21:54:25
         5           593520 2014-07-16 16:00:54

SQL> select file#,CREATION_CHANGE#,to_char(CREATION_TIME,'yyyy-mm-dd hh24:mi:ss') CREATION_TIME from v$datafile;

     FILE# CREATION_CHANGE# CREATION_TIME
---------- ---------------- -------------------
         1               18 2014-07-14 21:53:05
         2             2338 2014-07-14 21:53:42
         3             3130 2014-07-14 21:53:51
         4            15268 2014-07-14 21:54:25
         5           593520 2014-07-16 16:00:54

SQL> select file#,error from v$datafile_header;

     FILE# ERROR
---------- -----------------------------------------------------------------
         1
         2
         3
         4
         5

通过重建控制文件消除了v$datafile_header.error报WRONG FILE CREATE错误,继续尝试online文件

SQL> recover datafile 5;
Media recovery complete.
SQL> alter database datafile 5 online;

Database altered.

SQL> select file#,name from v$datafile;

     FILE# NAME
---------- --------------------------------------------------
         1 /u01/app/oracle/oradata/ORCL/system01.dbf
         2 /u01/app/oracle/oradata/ORCL/sysaux01.dbf
         3 /u01/app/oracle/oradata/ORCL/undotbs01.dbf
         4 /u01/app/oracle/oradata/ORCL/users01.dbf
         5 /u01/app/oracle/oradata/ORCL/xifenfei_old.dbf

SQL> alter database open;
ORA-01092: ORACLE instance terminated. Disconnection forced
ORA-01177: data file does not match dictionary - probably old incarnation
ORA-01110: data file 5: '/u01/app/oracle/oradata/ORCL/xifenfei_old.dbf'
Process ID: 7437
Session ID: 7 Serial number: 5

出现这个错误,是由于数据库中,还有file$中也记录了数据文件创建scn,而这个scn现在和数据文件头和控制文件中的scn不相等,因此无法启动数据库成功.现在需要做的就是在数据库未启动状态下修改file$中的数据文件创建scn相关值,让其和数据文件头(控制文件中记录)一致

使用第三方工具定位file$记录

1|2|89600|0|1|4194302|1280|0|18||4194306|0x004000e9|0
2|2|70400|1|2|4194302|1280|0|2338||8388610|0x004000e9|1
3|2|25600|2|3|4194302|640|0|3130||12582914|0x004000e9|2
4|2|640|4|4|4194302|160|0|15268||16777218|0x004000e9|3
5|2|1280|7|5|0|0|0|593613||20971522|0x004000e9|4   
6|1|3840|||0|0|0|586295||25165826|0x004000e9|5
7|1|3840|||3932160|1280|0|587030||29360130|0x004000e9|6
对应file$结构确定每列含义,以及确定需要修改的列
每行倒数第二列为rdba地址,可以通过转换为file and block,这里对应的就是file 1 block 233
每行最后一列为该条记录在该rdba中的记录顺序

使用工具修改593613为593520,使得file$中的scn与现在控制文件和数据文件头一致,具体参考bbed修改数据内容

修改好file$中数据文件创建scn后,尝试继续操作

SQL> alter database open;
alter database open
*
ERROR at line 1:
ORA-01113: file 5 needs media recovery
ORA-01110: data file 5: '/u01/app/oracle/oradata/ORCL/xifenfei_old.dbf'

SQL> recover datafile 5;
Media recovery complete.
SQL> alter database open;

Database altered.

SQL> select file#,name from v$datafile;

     FILE# NAME
---------- --------------------------------------------------
         1 /u01/app/oracle/oradata/ORCL/system01.dbf
         2 /u01/app/oracle/oradata/ORCL/sysaux01.dbf
         3 /u01/app/oracle/oradata/ORCL/undotbs01.dbf
         4 /u01/app/oracle/oradata/ORCL/users01.dbf
         5 /u01/app/oracle/oradata/ORCL/xifenfei_old.dbf

通过这里的简单测试,发现几个问题
1.v$datafile_header.error报WRONG FILE CREATE错误 不一定就是数据文件异常,而其本质是数据文件头scn和控制文件中scn不一致
2.数据文件online需要file$,v$datafile_header,v$datafile中关于数据文件创建scn都一致
3.通过该分析,证明在一些极端情况下,考虑考虑该替换思路实现删除数据文件重新加入数据库

发表在 Oracle | 标签为 , , | 评论关闭