使用deepseek进行Oracle恢复,引起重大故障

有一个恢复case,查询数据库是open状态,有一个数据文件处于offline,删除表空间报offline的文件不能读写
2
3
1


根据经验,这个是一个小问题,可能就是由于datafile 5 offline了,而这个文件是undo表空间的所以出现这样的情况,想着屏蔽下异常回滚段,或者强制online下文件就可以解决该问题.先进行第一个尝试,屏蔽异常回滚段,由于库是open状态,直接查询数据库是否有异常回滚段
undo_seg

无法查询到异常回滚段,这个有点不太符合常规认知,进一步核实文件和表空间信息
4
5
6
到这一步就发现了异常:
1. v$tablespace里面有两个undotbs1的表空间(这个肯定不对,是ctl和ts$不一致)
2. ts$中只有一个而且ts#=9没有ts#=2
3. file$中有ts#=2,这样导致ts$和file$信息不匹配,也不对
基于上述这样信息,我怀疑有人对底层字典进行了操作delete了ts$这个表记录.让现场技术人员再次确认这个库的所有操作,最后确认在他不知情的情况下,有另外的技术人员上来进行了类似操作
delete_ts

根据他们提供的聊天记录,以及当前数据库情况,进一步确认他们应该是执行了

delete from ts$ where name='UNDOTB1';
delete from seg$ where ts#=2;

没有对file$进行delete操作.对于这样的情况,人工删除字典,明显没有处理干净.导致数据库的任何操作都会去检查异常事务.
seg


通过清理这些异常事务,数据库可以正常操作,数据也导出成功
expdp

后续和当时直接进行delete 字典操作的人员沟通,他那边是根据deepseek提供的建议进行处理的
deepseek

在这里温馨提醒,虽然现在的ai比较发达,很多问题可以直接在上面问出来答案,但是需要对这些答案有一个判断能力,不能他说啥你就执行啥,特别是数据库非常规恢复这种不可逆而且可能引起重大事故的高风险性操作需要谨慎和做好回退方案.

发表在 Oracle备份恢复 | 标签为 , , | 评论关闭

接手一个只差临门一脚的数据库恢复

刚刚进行了一个数据库恢复case,故障处理起来非常简单(设置一个参数启动库即可),大概回溯下故障经过,反馈是由于客户在没有停机的情况下复制虚拟机,然后启动虚拟机之后,数据库无法正常启动报ORA-00314 ORA-00312错误

Wed Jun 17 10:24:02 2026
ALTER DATABASE   MOUNT
Successful mount of redo thread 1, with mount id 3689345602
Database mounted in Exclusive Mode
Lost write protection disabled
Completed: ALTER DATABASE   MOUNT
Wed Jun 17 10:24:06 2026
ALTER DATABASE OPEN
Beginning crash recovery of 1 threads
 parallel recovery started with 32 processes
Started redo scan
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_60458.trc:
ORA-00314: log 2 of thread 1, expected sequence# 12271 doesn't match 9443
ORA-00312: online log 2 thread 1: '/oradata/orcl/redo02.log'
Aborting crash recovery due to error 314
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_60458.trc:
ORA-00314: log 2 of thread 1, expected sequence# 12271 doesn't match 9443
ORA-00312: online log 2 thread 1: '/oradata/orcl/redo02.log'
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_60458.trc:
ORA-00314: log 2 of thread 1, expected sequence# 12271 doesn't match 9443
ORA-00312: online log 2 thread 1: '/oradata/orcl/redo02.log'
ORA-314 signalled during: ALTER DATABASE OPEN...
Wed Jun 17 10:24:07 2026
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_m000_61170.trc:
ORA-00314: log 1 of thread 1, expected sequence# 12269 doesn't match 9441
ORA-00312: online log 1 thread 1: '/oradata/orcl/redo01.log'
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_m000_61170.trc:
ORA-00314: log 2 of thread 1, expected sequence# 12271 doesn't match 9443
ORA-00312: online log 2 thread 1: '/oradata/orcl/redo02.log'
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_m000_61170.trc:
ORA-00314: log 3 of thread 1, expected sequence# 12265 doesn't match 9444
ORA-00312: online log 3 thread 1: '/oradata/orcl/redo03.log'
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_m000_61170.trc:
ORA-00314: log 4 of thread 1, expected sequence# 12266 doesn't match 9438
ORA-00312: online log 4 thread 1: '/oradata/orcl/redo04.log'
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_m000_61170.trc:
ORA-00314: log 5 of thread 1, expected sequence# 12267 doesn't match 9439
ORA-00312: online log 5 thread 1: '/oradata/orcl/redo05.log'
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_m000_61170.trc:
ORA-00314: log 6 of thread 1, expected sequence# 12268 doesn't match 9440
ORA-00312: online log 6 thread 1: '/oradata/orcl/redo06.log'

这个错误的redo sequence差距有点大,个人感觉可能不是简单的复制引起的,由于没有第一现场不好溯源,不乱做评论,姑且认为是由于虚拟机复制引起的问题
客户使用隐含参数强制打开数据库,报ORA-1555错误

Wed Jun 17 10:42:50 2026
ALTER DATABASE OPEN RESETLOGS
RESETLOGS is being done without consistancy checks. This may result
in a corrupted database. The database should be recreated.
RESETLOGS after incomplete recovery UNTIL CHANGE 1838398816
Resetting resetlogs activation ID 3463997979 (0xce786a1b)
ORA-344 signalled during: ALTER DATABASE OPEN RESETLOGS...
Wed Jun 17 10:44:30 2026
ALTER DATABASE OPEN RESETLOGS
RESETLOGS is being done without consistancy checks. This may result
in a corrupted database. The database should be recreated.
RESETLOGS after incomplete recovery UNTIL CHANGE 1838398816
Resetting resetlogs activation ID 3463997979 (0xce786a1b)
Wed Jun 17 10:45:13 2026
Setting recovery target incarnation to 3
Wed Jun 17 10:45:13 2026
Assigning activation ID 3689297678 (0xdbe6370e)
Thread 1 opened at log sequence 1
  Current log# 1 seq# 1 mem# 0: /oradata/orcl/redo01.log
Successful open of redo thread 1
Wed Jun 17 10:45:14 2026
MTTR advisory is disabled because FAST_START_MTTR_TARGET is not set
Wed Jun 17 10:45:14 2026
SMON: enabling cache recovery
ORA-01555 caused by SQL statement below (SQL ID: 4krwuz0ctqxdt, SCN: 0x0000.6d93bd66):
select ctime, mtime, stime from obj$ where obj# = :1
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_18038.trc:
ORA-00704: bootstrap process failure
ORA-00704: bootstrap process failure
ORA-00604: error occurred at recursive SQL level 1
ORA-01555: snapshot too old: rollback segment number 35 with name "_SYSSMU35_3782695576$" too small
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_18038.trc:
ORA-00704: bootstrap process failure
ORA-00704: bootstrap process failure
ORA-00604: error occurred at recursive SQL level 1
ORA-01555: snapshot too old: rollback segment number 35 with name "_SYSSMU35_3782695576$" too small
Error 704 happened during db open, shutting down database
USER (ospid: 18038): terminating the instance due to error 704
Instance terminated by USER, pid = 18038
ORA-1092 signalled during: ALTER DATABASE OPEN RESETLOGS...
opiodr aborting process unknown ospid (18038) as a result of ORA-1092
Wed Jun 17 10:45:16 2026
ORA-1092 : opitsk aborting process

这个是一个比较经典的错误,以往的文章中总结了这类错误可能涉及的具体sql语句
在数据库open过程中常遇到ORA-01555汇总
数据库open过程遭遇ORA-1555对应sql语句补充
客户尝试多次重启之后,数据库报ORA-600 2662错误

Wed Jun 17 10:55:34 2026
Thread 1 advanced to log sequence 3 (thread open)
Thread 1 opened at log sequence 3
  Current log# 3 seq# 3 mem# 0: /oradata/orcl/redo03.log
Successful open of redo thread 1
MTTR advisory is disabled because FAST_START_MTTR_TARGET is not set
Wed Jun 17 10:55:34 2026
SMON: enabling cache recovery
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_6707.trc  (incident=117868):
ORA-00600: internal error code, arguments: [2662], [0], [1838441753], [0], [1838486139], [12583040], 
Incident details in: /u01/oracle/diag/rdbms/orcl/orcl/incident/incdir_117868/orcl_ora_6707_i117868.trc
Wed Jun 17 10:55:45 2026
Use ADRCI or Support Workbench to package the incident.
See Note 411.1 at My Oracle Support for error and packaging details.
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_6707.trc:
ORA-00600: internal error code, arguments: [2662], [0], [1838441753], [0], [1838486139], [12583040], 
Errors in file /u01/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_6707.trc:
ORA-00600: internal error code, arguments: [2662], [0], [1838441753], [0], [1838486139], [12583040], 
Error 600 happened during db open, shutting down database
USER (ospid: 6707): terminating the instance due to error 600
Instance terminated by USER, pid = 6707
ORA-1092 signalled during: alter database open...

客户后续多次重建ctl,强制拉库等操作,一直在ORA-600 2662上面循环,后面终于出现了ORA-600 4193/4194错误,数据库没有正常open成功,至此客户放弃恢复尝试。
我们接手故障之后,设置undo手工管理模式,然后直接启动库成功
open1
然后使用expdp导出数据,完成本次恢复工作

发表在 Oracle备份恢复 | 标签为 , , , , | 评论关闭

硬件故障后数据文件大小不对故障处理—Oracle碎片扫描恢复

有硬件恢复圈朋友找到我,说硬件恢复之后dbv报dbv-00102错误,让我给看看是否可以处理
dbv-00102


这个是oracle dbv中一种常见错误,一般是由于block 0 不对,或者是由于文件大小不对引起,让把恢复文件发给我,进行检查

SQL> select name,bytes/1024/1024/1024 from v$datafile_header;

NAME                                                                             BYTES/1024/1024/1024
-------------------------------------------------------------------------------- --------------------
H:\BAIDUNETDISK\ORADATA\XXXXORCL\SYSTEM01.DBF                                             2.080078125
H:\BAIDUNETDISK\ORADATA\XXXXORCL\SYSAUX01.DBF                                             2.880859375
H:\BAIDUNETDISK\ORADATA\XXXXORCL\UNDOTBS01.DBF                                           9.0087890625
H:\BAIDUNETDISK\ORADATA\XXXXORCL\USERS01.DBF                                          31.993408203125
H:\BAIDUNETDISK\ORADATA\XXXXORCL\USERS02.DBF                                                8.1640625
H:\BAIDUNETDISK\ORADATA\XXXXORCL\USERS03.DBF                                              7.958984375
H:\BAIDUNETDISK\ORADATA\XXXXORCL\USERS04.DBF                                              7.958984375
H:\BAIDUNETDISK\ORADATA\XXXXORCL\USERS05.DBF                                                 7.890625

已选择8行。

确定USER02-USERS05的dbf文件实际大小(数据文件头记录)在8G左右,但是目前恢复出来的文件大小只有4G左右
4g


在恢复工具中直接查看文件大小情况
rs

这里比较明显rs中虽然显示文件状态良好,但是实际大小也不对(得出经验:以后恢复中不能太依赖这个状态),根据反馈现场是三个盘的raid5,中途做了一次强制上线,然后客户也使用win pe拷贝过一次数据,大小和现在一样,也是少了近4G.第一反应可能是由于raid盘弄的不对,但是经过对其他文件的确认,多完全没有问题,排除了盘错误的问题,怀疑是由于文件系统异常导致,对于这种的情况,文件系统层面肯定无法恢复,考虑使用自研的OraScan工具进行扫描(OraScan(Oracle 碎片扫描工具) 使用说明)
ora1
ora2

通过OraScan扫描找到相关block,并提取出来数据文件
file

使用dbv检测文件

C:\Users\XFF>dbv file=H:\BaiduNetdisk\xff\YFKJORCL.USERS.4.7.4.N.DBF

DBVERIFY: Release 11.2.0.4.0 - Production on 星期日 6月 7 18:06:30 2026

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

DBVERIFY - 开始验证: FILE = H:\BAIDUNETDISK\XFF\YFKJORCL.USERS.4.7.4.N.DBF


DBVERIFY - 验证完成

检查的页总数: 1043200
处理的页总数 (数据): 67167
失败的页总数 (数据): 0
处理的页总数 (索引): 37995
失败的页总数 (索引): 0
处理的页总数 (其他): 861109
处理的总页数 (段)  : 0
失败的总页数 (段)  : 0
空的页总数: 76929
标记为损坏的总页数: 0
流入的页总数: 0
加密的总页数        : 0
最高块 SCN            : 347454063 (0.347454063)

把文件拷贝替换掉之前恢复的USERS02-USER05.DBF,然后尝试打开数据库

SQL> recover database ;
完成介质恢复。
SQL> alter database open ;
alter database open
*
第 1 行出现错误:
ORA-03113: 通信通道的文件结尾
进程 ID: 3308
会话 ID: 14 序列号: 3

查看alert日志分析原因

Sun Jun 07 14:43:51 2026
Recovery of Online Redo Log: Thread 1 Group 2 Seq 36464 Reading mem 0
  Mem# 0: H:\BAIDUNETDISK\ORADATA\YFKJORCL\REDO02.LOG
Completed: ALTER DATABASE RECOVER  database   
alter database open 
Beginning crash recovery of 1 threads
 parallel recovery started with 19 processes
Started redo scan
Completed redo scan
 read 2353 KB redo, 0 data blocks need recovery
Started redo application at
 Thread 1: logseq 36464, block 15876
Recovery of Online Redo Log: Thread 1 Group 2 Seq 36464 Reading mem 0
  Mem# 0: H:\BAIDUNETDISK\ORADATA\YFKJORCL\REDO02.LOG
Completed redo application of 0.00MB
Completed crash recovery at
 Thread 1: logseq 36464, block 20582, scn 347475303
 0 data blocks read, 0 data blocks written, 2353 redo k-bytes read
Sun Jun 07 14:43:57 2026
Errors in file c:\app\xff\diag\rdbms\yfkjorcl\o11201\trace\o11201_lgwr_2204.trc:
ORA-00314: ?? 3 (???? 1) ??? sequence# 36462 ? 32025 ???
ORA-00312: ???? 3 ?? 1: 'H:\BAIDUNETDISK\ORADATA\YFKJORCL\REDO03.LOG'
Errors in file c:\app\xff\diag\rdbms\yfkjorcl\o11201\trace\o11201_lgwr_2204.trc:
ORA-00314: ?? 3 (???? 1) ??? sequence# 36462 ? 32025 ???
ORA-00312: ???? 3 ?? 1: 'H:\BAIDUNETDISK\ORADATA\YFKJORCL\REDO03.LOG'
Errors in file c:\app\xff\diag\rdbms\yfkjorcl\o11201\trace\o11201_ora_3308.trc:
ORA-00314: 日志 1 (用于线程 ) 要求的 sequence#  与  不匹配
ORA-00312: 联机日志 3 线程 1: 'H:\BAIDUNETDISK\ORADATA\YFKJORCL\REDO03.LOG'
USER (ospid: 3308): terminating the instance due to error 314
Sun Jun 07 14:44:02 2026
Instance terminated by USER, pid = 3308

由于redo group 异常导致库无法正常open,但是由于已经recover database成功,因此大概率可以clear该redo 组

SQL> select group#,status from v$log;

    GROUP# STATUS
---------- ----------------
         1 INACTIVE
         3 INACTIVE
         2 CURRENT

SQL> alter database clear logfile group 3;

数据库已更改。

SQL> alter database open;

数据库已更改。

数据库open成功,然后使用expdp导出数据,完成本次恢复任务.

发表在 非常规恢复 | 标签为 , , , | 评论关闭