Small case for windowsTB non-archive Oracle database recovery

This is a case of friendship to support a friend's database recovery. I came to ask for help during the Christmas Day yesterday, and I had to help solve the problem quickly. First, let's take a look at the alert log information:

Wed Dec 23 15:19:21 2015
SMON: enabling tx recovery
Wed Dec 23 15:19:21 2015
Database Characterset is ZHS16GBK
Opening with internal Resource Manager plan
Where numa pg = 2, CPUs = 8
Wed Dec 23 15:19:22 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ ajzq \ udump \ ajzq_ora_1712.trc:
ORA-00600: internal error code, arguments: [4194], [45], [45], [], [], [], [], [], []
 
Doing block recovery for file 2 block 1618943
Block recovery from logseq 4, block 63 to scn 7662329350
Wed Dec 23 15:19:23 2015
Recovery of Online Redo Log: Thread 1 Group 1 Seq 4 Reading mem 0
Mem #0: D: \ ORACLE \ PRODUCT \ 10.2.0 \ ORADATA \ xxxx \ REDO01.LOG
Block recovery completed at rba 4.4749.16, scn 1.3367362055
Doing block recovery for file 2 block 121
Block recovery from logseq 4, block 63 to scn 7662329346
Wed Dec 23 15:19:23 2015
Recovery of Online Redo Log: Thread 1 Group 1 Seq 4 Reading mem 0
Mem #0: D: \ ORACLE \ PRODUCT \ 10.2.0 \ ORADATA \ xxxx \ REDO01.LOG
Block recovery completed at rba 4.2.16.16, scn 1.3367362051
Wed Dec 23 15:19:24 2015
SMON: Restarting fast_start parallel rollback
Wed Dec 23 15:19:24 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ udump \ xxxx_ora_1712.trc:
ORA-00600: internal error code, arguments: [4194], [45], [45], [], [], [], [], [], []
 
Wed Dec 23 15:19:24 2015
DEBUG: Replaying xcb 0xf0adfbe0, pmd 0x20588f48 for failed op 8
Doing block recovery for file 2 block 1618943
Block recovery from logseq 4, block 63 to scn 7662329350
Wed Dec 23 15:19:24 2015
Recovery of Online Redo Log: Thread 1 Group 1 Seq 4 Reading mem 0
Mem #0: D: \ ORACLE \ PRODUCT \ 10.2.0 \ ORADATA \ xxxx \ REDO01.LOG
Block recovery completed at rba 4.4749.16, scn 1.3367362055
Wed Dec 23 15:21:28 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ udump \ xxxx_ora_1712.trc:
ORA-00600: internal error code, arguments: [4194], [45], [45], [], [], [], [], [], []
ORA-00600: internal error code, arguments: [4194], [45], [45], [],
.....
Wed Dec 23 16:47:26 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_p005_8980.trc:
ORA-00600: internal error code, arguments: [2037], [141460651], [141460651], [162], [6], [1], [3367403517], [3141222167]
Wed Dec 23 16:47:26 2015
Hex dump of (file 92,block 3926982) in trace file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_dbw0_8864.trc
Unzip UPT block relative dba: 0x173bebc6 (file 92, block 3926982)
Bad header found during preparing block for write
Data in bad block:
Type: 23 format: 7 rdba: 0xbd3b3f17
Last change scn: 0x0001. c8b6be51 seq: 0x5 flg: 0x00
Spare1: 0x3b spare2: 0xbc spare3: 0xc
Consistency value in tail: 0xbe0000705
Check value in block header: 0x0
Block checksum disabled
Wed Dec 23 16:47:26 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_p005_8980.trc:
ORA-00600: internal error code, arguments: [2037], [141460651], [141460651], [162], [6], [1], [3367403517], [3141222167]
 
Wed Dec 23 16:47:27 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_dbw0_8864.trc:
ORA-00600: internal error code, arguments: [kcbzpbuf_1], [4], [1], [], [], [], [], [], []
 
Wed Dec 23 16:47:27 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_dbw0_8864.trc:
ORA-00600: internal error code, arguments: [kcbzpbuf_1], [4], [1], [], [], [], [], [], []
 
DBW0: terminating instance due to error 471
We can see that this database has been down for several days and cannot be opened on the 23rd. The above results are also common and simple. After processing, the user successfully opened the database. However, after opening the database, the user created a new undo tablespace and then tried to restart the tablespace, and then started the database again. The following error is reported:

Fri Dec 25 11:44:17 2015
Recovery of Online Redo Log: Thread 1 Group 3 Seq 6 Reading mem 0
Mem #0: D: \ ORACLE \ PRODUCT \ 10.2.0 \ ORADATA \ xxxx \ REDO03.LOG
Fri Dec 25 11:44:17 2015
Completed redo application
Fri Dec 25 11:44:17 2015
Hex dump of (file 44, block 518322) in trace file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_dbw2_21232.trc
Unzip UPT block relative dba: 0x0b07e8b2 (file 44, block 518322)
Bad header found during preparing block for write
Data in bad block:
Type: 82 format: 2 rdba: 0x1c4ee895
Last change scn: 0x0001. c8bbad54 seq: 0x21 flg: 0xf0
Spare1: 0x52 spare2: 0x4d spare3: 0x895b
Consistency value in tail: 0xad545221
Check value in block header: 0x0
Block checksum disabled
Fri Dec 25 11:44:18 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_dbw2_21232.trc:
ORA-00600: internal error code, arguments: [kcbzpbuf_1], [4], [1], [], [], [], [], [], []
 
Fri Dec 25 11:44:18 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_dbw2_21232.trc:
ORA-00600: internal error code, arguments: [kcbzpbuf_1], [4], [1], [], [], [], [], [], []
 
DBW2: terminating instance due to error 471
Fri Dec 25 11:44:18 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_p014_20704.trc:
ORA-00471: DBWR process terminated with error.
This error is also my first time. According to the kcbzpbuf function, bad blocks are found when dbwr is used to write dirty blocks. Obviously, the dba address 0x0b07e8b2 and rdba address 0x1c4ee895 do not match from the bad block information above, which cannot be written normally.
I searched metalink for this error and found many related documents or even bugs. The processing method is simple. First, add the following two implicit parameters to try to open the database:
*. _ Allow_resetlogs_uption = TRUE
*. _ Allow_error_simulation = TRUE
After incomplete recovery, the following error occurs when you try open resetlogs:

Ri Dec 25 12:20:03 2015
SMON: enabling cache recovery
Fri Dec 25 12:20:04 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ udump \ xxxx_ora_00006.trc:
ORA-00600: internal error code, arguments: [2662], [1], [3367742663], [1], [3367742798], [4194313], [], []
 
Fri Dec 25 12:20:05 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ udump \ xxxx_ora_00006.trc:
ORA-00600: internal error code, arguments: [2662], [1], [3367742663], [1], [3367742798], [4194313], [], []
 
Fri Dec 25 12:20:05 2015
Error 600 happened during db open, shutting down database
USER: terminating instance due to error 600.
Fri Dec 25 12:20:05 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_p015_21472.trc:
ORA-00600: internal error code, arguments: [15784], [600], [], [], [], [], [], [], []
ORA-00600: internal error code, arguments: [], [], [], [], [], [], [], []
 
Fri Dec 25 12:20:05 2015
Errors in file d: \ oracle \ product \ 10.2.0 \ admin \ xxxx \ bdump \ xxxx_p014_9072.trc:
ORA-00600: internal error code, arguments: [15784], [600], [], [], [], [], [], [], []
ORA-00600: internal error code, arguments: [], [], [], [], [], [], [], []
Many people may be confused when we see two errors. Which error did the database fail to open? Here a little attention before and after the error order will know, it is obvious that the previous ora-00600 [2662] error caused by failure to open resetlogs complete.
For the ora-00600 [2662] error, this is not explained here, the SCN problem. Here, we can directly promote scn.
Run the following command to start the database to the mount Status:

SQL> alter system set job_queue_processes = 0;
 
Session altered.
 
SQL> startup mount
ORACLE instance started.
 
Total System Global Area 7516192768 bytes
Fixed Size 2175728 bytes
Variable Size 1232850192 bytes
Database Buffers 6274678784 bytes
Redo Buffers 6488064 bytes
Database mounted.
SQL> alter session set events 'immediate trace name adjust_scn level 10 ';
 
Session altered.
 
SQL> alter database open;
 
Database altered.
The database was opened smoothly. Since a friend has already created a new undo tablespace, I just need to help re-build the undo tablespace and end the support task. An ora-01548 error occurred while trying to drop tablespace undotbs1 including contents and datafiles.
After alter session set "_ smu_debug_mode" = 4; and then drop is not allowed, you can continue to use the following implicit parameters:

_ Offline_rollback_segments
_ Uploupted_rollback_segments
With these two implicit parameters, you can smoothly drop the old undo tablespace, modify the pfile file, and remove unnecessary implicit parameters to open the database smoothly.
Here I end my half-hour support task! In the afternoon of tragedy, I began to optimize the SQL statements of a customer .....
Note:
1) We recommend that you perform a full database check for the database that is forcibly opened due to this exception, including whether there are bad blocks or even whether the data dictionary is consistent (in fact, check the previous alert log findings, in the past, the database reported a lot of bad blocks during the recovery process)
2) in this age, there were 3 TB databases and windows databases. What's even more tragic is that they are not archived and there is no backup. This is not the case.
3) we should consider the disaster recovery construction of the system, even if there is backup.