ORA-600 ktfbhget-4

有客户和我们反馈,他们数据库现在使用异常,查询dba_data_files报ORA-600 ktfbhget-4

*** 2019-09-21 13:56:42.944
ksedmp: internal or fatal error
ORA-00600: internal error code, arguments: [ktfbhget-4], [384], [16], [], [], [], [], []
Current SQL statement for this session:
select tablespace_name from dba_data_files
----- Call Stack Trace -----
calling              call     entry                argument values in hex      
location             type     point                (? means dubious value)     
-------------------- -------- -------------------- ----------------------------
ksedmp+0148          bl       ksedst               1028F23AC ?
ksfdmp+0018          bl       01FD7F4C             
kgerinv+00e8         bl       _ptrgl               
kgesinv+0020         bl       kgerinv              000000000 ? 000000000 ?
                                                   FFFFFFFFFFF8920 ? 000400002 ?
                                                   000000000 ?
ksesin+005c          bl       kgesinv              7000001637868E8 ? 110359FD8 ?
                                                   000000100 ? 000007FFF ?
                                                   000007FFF ?
ktfbhget+047c        bl       ksesin               1029B1D18 ? 200000002 ?
                                                   000000000 ? 000000180 ?
                                                   000000000 ? 000000010 ?
                                                   000000008 ? 000000000 ?
ktfbhcf+03c0         bl       ktfbhget             FFFFFFFFFFF8DA0 ? 1103589D8 ?
                                                   201035A57C ? 1C610005A40 ?
qerfxFetch+0c94      bl       01FD7AC4             
qerjoFetch+037c      bl       _ptrgl               
rwsfcd+0060          bl       _ptrgl               
qersoFetch+0108      bl       _ptrgl               
qerjoFetch+037c      bl       _ptrgl               
qerjoFetch+037c      bl       _ptrgl               
rwsfcd+0060          bl       _ptrgl               
qeruaFetch+0140      bl       01FD7AC4             
qervwFetch+008c      bl       01FD7AC4             
opifch2+0c68         bl       01FD7AC4             
opiall0+1244         bl       opifch2              7000001626B2F28 ? 100000001 ?
                                                   FFFFFFFFFFFA3D8 ?
kpoal8+0a68          bl       opiall0              5E1000C818 ? 2200000014 ?
                                                   FFFFFFFFFFFAAD8 ? 000000000 ?
                                                   FFFFFFFFFFFAA28 ? 11028C2C8 ?
                                                   080000000 ? 4000000007FFF ?
opiodr+08e8          bl       _ptrgl               
ttcpip+0c54          bl       _ptrgl               
opitsk+0c28          bl       ttcpip               11000C818 ? 000000000 ?
                                                   000000000 ? 000000000 ?
                                                   000000000 ? 000000000 ?
                                                   000000000 ? 000000000 ?
opiino+0798          bl       opitsk               000000000 ? 000000000 ?
opiodr+08e8          bl       _ptrgl               
opidrv+032c          bl       opiodr               3C00000018 ? 4101F5E80 ?
                                                   FFFFFFFFFFFF4C0 ? 0A000FD98 ?
sou2o+0028           bl       opidrv               3C0C000000 ? 400000000 ?
                                                   FFFFFFFFFFFF4C0 ?
main+0138            bl       01FD758C             
__start+0098         bl       main                 000000000 ? 000000000 ?

通过询问知道由于file#=17最初由于裸设备权限异常,导致无法访问,执行了delete from file$ where file#=17,然后遭遇异常又人工插入了一条file#=17的记录到file$中,但是由于不知道具体值,可能是由于部分值插入错误引起现在问题.

---

通过分析file$表定义

create table file$                                             /* file table */
( file#         number not null,                   /* file identifier number */
  status$       number not null,                      /* status (see KTS.H): */
                                               /* 1 = INVALID, 2 = AVAILABLE */
  blocks        number not null,                   /* size of file in blocks */
                                           /* zero for bitmapped tablespaces */
  ts#           number,                         /* tablespace that owns file */
  relfile#      number,                              /* relative file number */
  maxextend     number,                                 /* maximum file size */
  inc           number,                                  /* increment amount */
  crscnwrp      number,                                 /* creation SCN wrap */
  crscnbas      number,                                 /* creation SCN base */
  ownerinstance varchar("M_IDEN"),                    /* Owner instance name */
  spare1        number,      /* tablespace-relative DBA of space file header */
                                   /* NULL for dictionary-mapped tablespaces */
  spare2        number,
  spare3        varchar2(1000),
  spare4        date
)

通过dump文件头(文件创建大小/SCN等),结合一些计算和经验值,获取到相关值重新插入正确值,一切恢复正常

delete from file$ where file#=17 and ts#=384;
insert into sys.file$ (FILE#, STATUS$, BLOCKS, TS#, RELFILE#, MAXEXTEND, INC,
 CRSCNWRP, CRSCNBAS, OWNERINSTANCE, SPARE1, SPARE2, SPARE3, SPARE4)
values (17,2,256000,384,17,0,0,3087,1631091037,null,71303170, null, null, null);
commit;

再次提醒,file$中记录请勿人工修改,以前写过相关casefile$ 删除记录恢复(delete file$ recovery)