Oracle undo表空间文件丢失案例

场景1：
undo表空间文件丢失，有undo数据文件备份，日志存在且可用。
如果备份文件为RMAN备份：
restore
recover
如果为手工备份：
cp
recover

场景2:
undo表空间文件丢失，无undo数据文件备份，日志存在且可用。
mount模式执行：
alter database create datafile 3 as ‘/u03/app/oracle/oradata/prod/undotbs01.dbf’ size 50m;

recover进行读取日志恢复。

场景3：
undo表空间文件丢失，无undo数据文件备份，undo段头存在未提交的事务need recovery。

案例模拟：
查看当前回滚段状态
SYS@prod>select segment_name,status from dba_rollback_segs;

SEGMENT_NAME STATUS

SYSTEM ONLINE
_SYSSMU10_1197734989$ ONLINE
_SYSSMU9_1650507775$ ONLINE
_SYSSMU8_517538920$ ONLINE
_SYSSMU7_2070203016$ ONLINE
_SYSSMU6_1263032392$ ONLINE
_SYSSMU5_898567397$ ONLINE
_SYSSMU4_1254879796$ ONLINE
_SYSSMU3_1723003836$ ONLINE
_SYSSMU2_2996391332$ ONLINE
_SYSSMU1_3724004606$ ONLINE

执行一个事务：
HR@prod>update employees set salary = 100;

107 rows updated.
破坏undo
[oracle@service1 prod]$ cp /etc/passwd undotbs01.dbf

SYS@prod>shutdown abort;
ORACLE instance shut down.
SYS@prod>startup;
ORACLE instance started.

Total System Global Area 835104768 bytes
Fixed Size 2257840 bytes
Variable Size 541068368 bytes
Database Buffers 285212672 bytes
Redo Buffers 6565888 bytes
Database mounted.
ORA-01157: cannot identify/lock data file 3 - see DBWR trace file
ORA-01110: data file 3: ‘/u01/app/oracle/oradata/prod/undotbs01.dbf’

SYS@prod>alter system set undo_management=manual scope=spfile;
SYS@prod>alter database datafile 3 offline;
Database altered.

SYS@prod>select name,status from v$datafile;
/u01/app/oracle/oradata/prod/undotbs01.dbf
RECOVER
undo段上存在事务，需要recover。
但是我们没有备份无法恢复。

SYS@prod>alter database open;(不能进行DML操作)
Database altered.

创建一个undo表空间：
SYS@prod>create undo tablespace undotbs02 datafile ‘/u01/app/oracle/oradata/prod/untbs02.dbf’ size 10M autoextend on;

修改undo_tablespace/undo_management参数：
SYS@prod>alter system set undo_tablespace=undotbs2 scope=spfile;
SYS@prod>alter system set undo_management=auto scope=spfile;
SYS@prod>startup mount force;

SYS@prod>drop tablespace undotbs1 including contents and datafiles;
error：
ORA-01548: 已找到活动回退段’_SYSSMU1$’，终止删除表空间

查看回滚段状态：
SYS@prod>select segment_name,status from dba_rollback_segs;

SEGMENT_NAME STATUS

SYSTEM ONLINE
_SYSSMU10_1197734989$ NEEDS RECOVERY
_SYSSMU9_1650507775$ NEEDS RECOVERY
_SYSSMU8_517538920$ NEEDS RECOVERY
_SYSSMU7_2070203016$ NEEDS RECOVERY
_SYSSMU6_1263032392$ NEEDS RECOVERY
_SYSSMU5_898567397$ NEEDS RECOVERY
_SYSSMU4_1254879796$ NEEDS RECOVERY
_SYSSMU3_1723003836$ NEEDS RECOVERY
_SYSSMU2_2996391332$ NEEDS RECOVERY
_SYSSMU1_3724004606$ NEEDS RECOVERY

修改隐藏参数：
SYS@prod>alter system set “_offline_rollback_segments”=true scope=spfile;

SYS@prod>
alter system set “_corrupted_rollback_segments”=’_SYSSMU1_3724004606$ ‘,’_SYSSMU2_2996391332$ ‘,’_SYSSMU3_1723003836$ ‘,’_SYSSMU4_1254879796$ ‘,’_SYSSMU5_898567397$ ‘,’_SYSSMU6_1263032392$ ‘,’_SYSSMU7_2070203016$ ‘,’_SYSSMU8_517538920$ ‘,’_SYSSMU9_1650507775$ ‘,’_SYSSMU10_1197734989$’ scope=spfile

SYS@prod>drop tablespace undotbs1 including contents and datafiles;

Tablespace dropped.

再次查看回滚段状态：
SYS@prod>select segment_name,status from dba_rollback_segs;

SEGMENT_NAME STATUS

SYSTEM ONLINE
_SYSSMU30_260568829$ ONLINE
_SYSSMU29_1074352905$ ONLINE
_SYSSMU28_3910680688$ ONLINE
_SYSSMU27_3692739459$ ONLINE
_SYSSMU26_345096822$ ONLINE
_SYSSMU25_2172562784$ ONLINE
_SYSSMU24_109115100$ ONLINE
_SYSSMU23_3673005684$ ONLINE
_SYSSMU22_2483286186$ ONLINE
_SYSSMU21_2267006838$ ONLINE

可以校验数据，数据会被默认提交。
结论 open状态下，undo数据文件损坏，存在未提交活动的事务，恢复后，数据默认会提交。

场景4：
undo表空间文件丢失，无undo数据文件备份，不存在未提交的事务，无需recover情况恢复。
简单描述：
undo数据文件offline后，直接变为offline状态，不是recover状态。
直接创建新的undo表空间，修改undo_tablespace参数替换即可。