使用PlateSpin复制出来的一数据库服务器（Oracle 10g）在启动数据库实例时遇到“ORA-00600: internal error code, arguments: [4194].....”错误，实例在启动后，会自然Down掉。具体情况如下所示：
Successfully onlined Undo Tablespace 54.
Mon Nov 21 11:34:03 2016
SMON: enabling tx recovery
Mon Nov 21 11:34:03 2016
Errors in file /u01/app/oracle/admin/epps/bdump/epps_smon_7522.trc:
ORA-00600: internal error code, arguments: [4097], [], [], [], [], [], [], []
Mon Nov 21 11:34:03 2016
Database Characterset is UTF8
Mon Nov 21 11:34:03 2016
Errors in file /u01/app/oracle/admin/epps/udump/epps_ora_7548.trc:
ORA-00600: internal error code, arguments: [4194], [59], [40], [], [], [], [], []
Mon Nov 21 11:34:04 2016
Non-fatal internal error happenned while SMON was doing temporary segment drop.
SMON encountered 1 out of maximum 100 non-fatal internal errors.
Mon Nov 21 11:34:04 2016
Errors in file /u01/app/oracle/admin/epps/bdump/epps_smon_7522.trc:
ORA-00600: internal error code, arguments: [4097], [], [], [], [], [], [], []
................................................................................
................................................................................
ORA-00600: internal error code, arguments: [4194], [52], [29], [], [], [], [], []
ORA-00600: internal error code, arguments: [4194]Non-fatal internal error happenned while SMON was doing temporary segment drop.
Mon Nov 21 11:35:04 2016
Flush retried for xcb 0x4419143b0, pmd 0x4401e3c90
Mon Nov 21 11:35:04 2016
SMON encountered 6 out of maximum 100 non-fatal internal errors.
Mon Nov 21 11:35:04 2016
Doing block recovery for file 2 block 1007
No block recovery was needed
Mon Nov 21 11:35:05 2016
Errors in file /u01/app/oracle/admin/epps/bdump/epps_pmon_7510.trc:
ORA-00600: internal error code, arguments: [4194], [52], [29], [], [], [], [], []
Mon Nov 21 11:35:05 2016
Errors in file /u01/app/oracle/admin/epps/bdump/epps_pmon_7510.trc:
ORA-00600: internal error code, arguments: [4194], [52], [29], [], [], [], [], []
PMON: terminating instance due to error 472
Instance terminated by PMON, pid = 7510
这个错误，官方文档Step by step to resolve ORA-600 4194 4193 4197 on database crash (文档 ID 1428786.1)有提供详细的介绍。建议处理前，建议先查阅次文档，错误出现的主要场景：
This issue generally occurs when there is a power outage or hardware failure that initially crashes the database. On startup, the database does the normal roll forward (redo) and then rollback (undo), this is where the error is generated on the rollback.
具体操作如下所示
Step 1: 通过spfile创建生成pfile
SQL> create pfile from spfile;
File created.
Step 2: 关闭数据库实例。
Step 3：修改pfile中undo_management为MANUAL
[oracle@DB-Server dbs]$ grep undo initepps.ora
*.undo_management='AUTO'
*.undo_tablespace='UNDOTBS'
[oracle@DB-Server dbs]$ vi initepps.ora
Step 4: 使用PFILE启动数据库
[oracle@DB-Server dbs]$ sqlplus / as sysdba
SQL*Plus: Release 10.2.0.4.0 - Production on Mon Nov 21 11:51:59 2016
Copyright (c) 1982, 2007, Oracle. All Rights Reserved.
Connected to an idle instance.
SQL> startup restrict pfile='initepps.ora'
ORACLE instance started.
Total System Global Area 1.0737E+10 bytes
Fixed Size 2101808 bytes
Variable Size 6476008912 bytes
Database Buffers 4244635648 bytes
Redo Buffers 14671872 bytes
Database mounted.
Database opened.
Step 5：This is critical - we are looking for all undo segments to be offline - System will always be online.
If any are 'PARTLY AVAILABLE' or 'NEEDS RECOVERY' - Please open an issue with Oracle Support or update the current SR. There are many options from this moment and Oracle Support Analyst can offer different solutions for the bad undo segments.
If all offline then continue to the next step
SQL> select tablespace_name, status, segment_name from dba_rollback_segs where status != 'OFFLINE';
TABLESPACE_NAME STATUS SEGMENT_NAME
------------------------------ ---------------- ------------------------------
SYSTEM ONLINE SYSTEM
Step 6： 新建一个UNDO表空间
SQL> create undo tablespace UNDOTBS1
2 datafile '/u04/epps/oradata/undotbs_01.dbf'
3 size 4G;
Tablespace created.
Step 7： 删除旧的UNDO表空间
SQL> drop tablespace UNDOTBS including contents and datafiles;
Tablespace dropped.
Step 8：关闭数据库实例
SQL> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
SQL> exit
Disconnected from Oracle Database 10g Release 10.2.0.4.0 - 64bit Production
Step 9：启动数据库实例到NOMOUNT状态
SQL> startup nomount;
ORACLE instance started.
Total System Global Area 1.6777E+10 bytes
Fixed Size 2113368 bytes
Variable Size 9982443688 bytes
Database Buffers 6777995264 bytes
Redo Buffers 14663680 bytes
Step 10：修改spfile中的undo_tablespace参数
SQL> alter system set undo_tablespace='UNDOTBS1' scope=spfile;
System altered.
Setp 11：关闭数据库实例。
Step 12：启动数据库实例（使用spfile）
SQL> startup;
ORACLE instance started.
Total System Global Area 1.6777E+10 bytes
Fixed Size 2113368 bytes
Variable Size 9982443688 bytes
Database Buffers 6777995264 bytes
Redo Buffers 14663680 bytes
Database mounted.
Database opened.
