年终难终 - 记年底前的几则数据恢复案例 - Oracle Life

« Oracle数据恢复 - Linux / Unix 误删除的文件恢复 | Blog首页 | Delete sdo_geor_ddl__table$$ 的异常删除操作 »

年终难终 - 记年底前的几则数据恢复案例

每到年底，很多数据库就开始动荡不安起来，这是一个数据库故障的多发期，大家都应当提高警惕。

以下收录了最近我在微博上发布的关于数据库恢复的案例，供大家参考警示。

【数据恢复】刚刚又接到一个用户电话，Linux系统，Oracle Database 11g，因为某些故障，去fsck修复文件系统，结果数据库崩溃，主机也无法正常启动，数据库服务中断。fsck 是一个相对危险的操作，对于无备份的数据库系统，任何一个块的修复和变更都可能导致数据库故障，慎用慎用！

【数据恢复】上午接到用户电话，一个目录，24个文件，直接被rm删除掉了，之前用户的空间紧张，为了保证运行，经常性四处倒腾空间，删除、转移文件，为了一块硬盘，损失了整个数据库，这就是得不偿失！

【数据恢复】刚刚收到用户电话，某集成商在帮助用户恢复REDO损坏故障时，强制重置日志，启动数据库，丢失了重要数据，结果遭到用户质疑，难以收尾。这个案例给我们的启示是：在恢复数据库时，一定要充分考虑数据的重要性，精心设计方案，并取得用户认同。任何一条数据对于用户都可能无比重要。

今天在ITPUB论坛上又看到两则严重的数据库故障，收录信息如下：
1.SYSTEM表空间损坏
类似的案例本站之前描述过：

Mon Dec 19 12:00:16 2011
ALTER DATABASE OPEN
Mon Dec 19 12:00:16 2011
Beginning crash recovery of 1 threads
Mon Dec 19 12:00:16 2011
Started first pass scan
Mon Dec 19 12:00:17 2011
Completed first pass scan
3953 redo blocks read, 365 data blocks need recovery
Mon Dec 19 12:00:17 2011
Started recovery at
Thread 1: logseq 2866, block 2, scn 0.288668733
Recovery of Online Redo Log: Thread 1 Group 3 Seq 2866 Reading mem 0
Mem# 0 errs 0: D:\ORACLE\ORADATA\TH\REDO03.LOG
RECOVERY OF THREAD 1 STUCK AT BLOCK 9 OF FILE 1
ORA-1172 signalled during: ALTER DATABASE OPEN...
Sat Dec 24 10:05:35 2011
alter database open
Sat Dec 24 10:05:35 2011
Beginning crash recovery of 1 threads
Sat Dec 24 10:05:35 2011
Started first pass scan
Sat Dec 24 10:05:36 2011
Completed first pass scan
3953 redo blocks read, 365 data blocks need recovery
Sat Dec 24 10:05:36 2011
Started recovery at
Thread 1: logseq 2866, block 2, scn 0.288668733
Recovery of Online Redo Log: Thread 1 Group 3 Seq 2866 Reading mem 0
Mem# 0 errs 0: D:\ORACLE\ORADATA\TH\REDO03.LOG
RECOVERY OF THREAD 1 STUCK AT BLOCK 9 OF FILE 1
ORA-1172 signalled during: alter database open...
Sat Dec 24 10:05:59 2011
ALTER DATABASE RECOVER datafile 1
Sat Dec 24 10:05:59 2011
Media Recovery Datafile: 1
Media Recovery Start
Starting datafile 1 recovery in thread 1 sequence 2866
Datafile 1: 'D:\ORACLE\ORADATA\TH\SYSTEM01.DBF'
Media Recovery Log
Recovery of Online Redo Log: Thread 1 Group 3 Seq 2866 Reading mem 0
Mem# 0 errs 0: D:\ORACLE\ORADATA\TH\REDO03.LOG
Sat Dec 24 10:06:08 2011
Errors in file d:\oracle\admin\th\udump\th_ora_3672.trc:
ORA-00600: 内部错误代码，参数: [3020], [4194313], [1], [2866], [3], [16], [], []
ORA-10567: Redo is inconsistent with data block (file# 1, block# 9)
ORA-10564: tablespace SYSTEM
ORA-01110: 数据文件 1: 'D:\ORACLE\ORADATA\TH\SYSTEM01.DBF'
ORA-10560: block type 'KTU UNDO HEADER W/UNLIMITED EXTENTS'
Errors with log .
Media Recovery failed with error 600
ORA-283 signalled during: ALTER DATABASE RECOVER datafile 1 ...