ORA-600 ktubko_1 恢复

oracle 12.2的rac库,pdb在open成功之后,没过多久会自动crash掉,主要报错ORA-600 ktubko_1

2022-10-08T16:00:17.874444+08:00
XFF(5):Endian type of dictionary set to little
2022-10-08T16:00:18.602483+08:00
XFF(5):[218515] Successfully onlined Undo Tablespace 26.
XFF(5):Undo initialization finished serial:0 start:73483625 end:73484200 diff:575 ms (0.6 seconds)
XFF(5):Database Characterset for XFF is ZHS16GBK
2022-10-08T16:00:19.340271+08:00
Buffer Cache Full DB Caching mode changing from FULL CACHING ENABLED to FULL CACHING DISABLED 
Full DB Caching disabled: DEFAULT_CACHE_SIZE should be at least 1394670 MBs bigger than current size. 
2022-10-08T16:00:21.308122+08:00
XFF(5):Opening pdb with no Resource Manager plan active
2022-10-08T16:00:22.655433+08:00
Pluggable database XFF opened read write
Completed:  ALTER PLUGGABLE DATABASE ALL OPEN
2022-10-08T16:00:36.419719+08:00
XFF(5):Setting Resource Manager plan SCHEDULER[0x4AC8]:DEFAULT_MAINTENANCE_PLAN via scheduler window
XFF(5):Setting Resource Manager plan DEFAULT_MAINTENANCE_PLAN via parameter
2022-10-08T16:00:57.054295+08:00
XFF(5):minact-scn: got error during useg scan e:1555 usn:57
XFF(5):minact-scn: useg scan erroring out with error e:1555
2022-10-08T16:01:41.527943+08:00
Errors in file /u01/app/db/diag/rdbms/orcl/orcl1/trace/orcl1_smon_218039.trc  (incident=737693) (PDBNAME=XFF):
ORA-00600: internal error code, arguments: [ktubko_1], [], [], [], [], [], [], [], [], [], [], []
XFF(5):Incident details in: /u01/app/db/diag/rdbms/orcl/orcl1/incident/incdir_737693/orcl1_smon_218039_i737693.trc
XFF(5):Use ADRCI or Support Workbench to package the incident.
See Note 411.1 at My Oracle Support for error and packaging details.
2022-10-08T16:01:41.530481+08:00
XFF(5):*****************************************************************
XFF(5):An internal routine has requested a dump of selected redo.
XFF(5):This usually happens following a specific internal error, when
XFF(5):analysis of the redo logs will help Oracle Support with the
XFF(5):diagnosis.
XFF(5):It is recommended that you retain all the redo logs generated (by
XFF(5):all the instances) during the past 12 hours, in case additional
XFF(5):redo dumps are required to help with the diagnosis.
XFF(5):*****************************************************************
2022-10-08T16:01:42.611317+08:00
XFF(5):*****************************************************************
XFF(5):An internal routine has requested a dump of selected redo.
XFF(5):This usually happens following a specific internal error, when
XFF(5):analysis of the redo logs will help Oracle Support with the
XFF(5):diagnosis.
XFF(5):It is recommended that you retain all the redo logs generated (by
XFF(5):all the instances) during the past 12 hours, in case additional
XFF(5):redo dumps are required to help with the diagnosis.
XFF(5):*****************************************************************
XFF(5):ORACLE Instance orcl1 (pid = 44) - Error 600 encountered while recovering transaction (12, 1) on object 50.
2022-10-08T16:01:42.611961+08:00
XFF(5):Errors in file /u01/app/db/diag/rdbms/orcl/orcl1/trace/orcl1_smon_218039.trc:
ORA-00600: internal error code, arguments: [ktubko_1], [], [], [], [], [], [], [], [], [], [], []
2022-10-08T16:01:42.849438+08:00
Errors in file /u01/app/db/diag/rdbms/orcl/orcl1/trace/orcl1_smon_218039.trc  (incident=737694) (PDBNAME=XFF):
ORA-00600: internal error code, arguments: [ktubko_1], [], [], [], [], [], [], [], [], [], [], []
XFF(5):Incident details in: /u01/app/db/diag/rdbms/orcl/orcl1/incident/incdir_737694/orcl1_smon_218039_i737694.trc
…………
2022-10-08T16:01:55.212368+08:00
Instance Critical Process (pid: 44, ospid: 218039, SMON) died unexpectedly
PMON (ospid: 217933): terminating the instance due to error 474
2022-10-08T16:01:55.379857+08:00
System state dump requested by (instance=1, osid=217933 (PMON)), summary=[abnormal instance termination].
System State dumped to trace file /u01/app/db/diag/rdbms/orcl/orcl1/trace/orcl1_diag_217966_20221008160155.trc
2022-10-08T16:01:56.417514+08:00
ORA-1092 : opitsk aborting process

因为有smon报的ORACLE Instance orcl1 (pid = 44) – Error 600 encountered while recovering transaction (12, 1) on object xxx这种比较明显错误,基本上可以定位是undo问题.对undo异常事务进行处理,数据库顺利open,并且稳定不再crash,然后对异常对象进行处理(当然也可以逻辑迁移)

---

在oracle 12.2到18.14的rac环境的cdb库中,如果节点sga大小不一致,而且有一个节点sga大于128G,就可能出现该问题,敬请注意

---

Bug 32347014: ORA-600[4506], ORA-600[KTUBKO_1] OCCUR AND INSTANCE CRASHES