ORA-07445 lmebucp

xifenfei

有一个客户找到我们,说他们是数据库启动之时报的错误和数据库不能open 报ORA-7445lmebucp错类似,让我们对其进行恢复支持,通过分析确定客户数据库版本为12.2.0.1

Oracle Database  12c Enterprise Edition Release 12.2.0.1.0 - 64bit Productio    PL/SQL Release  12.2.0.1.0 - Production      CORE 12.2.0.1.0  Production      TNS for Linux: Version  12.2.0.1.0 - Production      NLSRTL Version  12.2.0.1.0 - Production

alert日志报错

--最初报错    2020-05-11T03:43:06.787164+08:00    Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_m004_3253.trc   (incident=639048):    ORA-00600: 内部错误代码, 参数:  [kkpo_rcinfo_defstg:delseg], [72280], [], [], [], [], [], [], [], [], [], []    Incident details  in: /u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_639048/orcl_m004_3253_i639048.trc    Use ADRCI or  Support Workbench to package the incident.    See Note 411.1  at My Oracle Support for error and packaging details.    2020-05-11T03:43:14.250993+08:00    Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_m004_3253.trc   (incident=639049):    ORA-00600: 内部错误代码, 参数:  [kkpo_rcinfo_defstg:delseg], [72280], [], [], [], [], [], [], [], [], [], []    Incident details  in: /u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_639049/orcl_m004_3253_i639049.trc    2020-05-11T03:43:21.286310+08:00    Use ADRCI or  Support Workbench to package the incident.    See Note 411.1  at My Oracle Support for error and packaging details.    Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_m004_3253.trc   (incident=639050):    ORA-00600: 内部错误代码, 参数:  [kkpo_rcinfo_defstg:delseg], [72280], [], [], [], [], [], [], [], [], [], []    ORA-06512: 在 line 2    Incident details  in:  /u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_639050/orcl_m004_3253_i639050.trc    2020-05-11T03:43:28.059048+08:00    Use ADRCI or  Support Workbench to package the incident.    See Note 411.1  at My Oracle Support for error and packaging details.    2020-05-11T03:43:28.074681+08:00    Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_m004_3253.trc:    ORA-00600: 内部错误代码, 参数:  [kkpo_rcinfo_defstg:delseg], [72280], [], [], [], [], [], [], [], [], [], []    ORA-06512: 在 line 2    2020-05-11T08:31:22.416087+08:00    Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_j000_16511.trc:    ORA-12012: 自动执行作业 141 出错    ORA-30732: 表中不包含用户可见的列    ORA-06512: 在 line 1          ---关闭数据库之后重启报错    2020-05-11T09:42:43.234769+08:00    ALTER DATABASE  OPEN    2020-05-11T09:42:44.353085+08:00    Ping without log  force is disabled:      instance  mounted in exclusive mode.    Endian type of dictionary  set to little    2020-05-11T09:42:44.660388+08:00    TT00: Gap  Manager starting (PID:31134)    2020-05-11T09:42:45.180876+08:00    Thread 1 opened  at log sequence 78596      Current  log# 2 seq# 78596 mem# 0: /u01/app/oracle/oradata/orcl/redo02.log    Successful open of redo thread 1    2020-05-11T09:42:45.181357+08:00    MTTR advisory is  disabled because FAST_START_MTTR_TARGET is not set    Exception [type:  SIGSEGV, Address not mapped to object][ADDR:0x0][PC:0x10F4C112, lmebucp()+34][flags:  0x0, count: 1]    Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_31125.trc   (incident=646445):    ORA-07445:  exception encountered:core  dump[lmebucp()+34][SIGSEGV][ADDR:0x0][PC:0x10F4C112][Address not mapped to  object]    Incident details  in:  /u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_646445/orcl_ora_31125_i646445.trc    Use ADRCI or  Support Workbench to package the incident.    See Note 411.1  at My Oracle Support for error and packaging details.    2020-05-11T09:42:46.070049+08:00    *****************************************************************    An internal  routine has requested a dump of selected redo.    This usually  happens following a specific internal error, when    analysis of the  redo logs will help Oracle Support with the    diagnosis.    It is  recommended that you retain all the redo logs generated (by    all the  instances) during the past 12 hours, in case additional    redo dumps are  required to help with the diagnosis.    *****************************************************************    2020-05-11T09:42:53.344955+08:00    Instance  Critical Process (pid: 41, ospid: 31125) died unexpectedly    PMON (ospid:  31026): terminating the instance due to error 12752    2020-05-11T09:42:53.377209+08:00    System state  dump requested by (instance=1, osid=31026 (PMON)), summary=[abnormal instance  termination].    System State  dumped to trace file /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_diag_31046_20200511094253.trc    2020-05-11T09:42:56.690224+08:00    Instance  terminated by PMON, pid = 31026

对启动过程做10046跟踪

PARSING IN  CURSOR #139821999713040 len=65 dep=1 uid=0 oct=3 lid=0    tim=2200910467  hv=1762642493 ad='1bfd79130'sqlid='aps3qh1nhzkjx'    select line#, sql_text  from bootstrap$ where obj# not in (:1, :2)    END OF STMT    PARSE  #139821999713040:c=378,e=378,p=0,cr=0,cu=0,mis=1,r=0,dep=1,og=4,plh=0,tim=2200910467    BINDS  #139821999713040:       Bind#0      oacdty=02  mxl=22(22) mxlc=00 mal=00 scl=00 pre=00      oacflg=08  fl2=1000001 frm=00 csi=00 siz=24 off=0      kxsbbbfp=7f2ad89fe6c8   bln=22  avl=02  flg=05      value=59    Bind#1      oacdty=02  mxl=22(22) mxlc=00 mal=00 scl=00 pre=00      oacflg=08  fl2=1000001 frm=00 csi=00 siz=24 off=0      kxsbbbfp=7f2ad89fe698   bln=24  avl=06  flg=05      value=4294967295    EXEC  #139821999713040:c=219,e=658,p=0,cr=0,cu=0,mis=1,r=0,dep=1,og=4,plh=867914364,tim=2200911218    WAIT  #139821999713040: nam='db file sequential read' ela= 9 file#=1 block#=520  blocks=1 obj#=59 tim=2200911300    WAIT  #139821999713040: nam='db file scattered read' ela= 19 file#=1 block#=521  blocks=3 obj#=59 tim=2200911559    FETCH  #139821999713040:c=404,e=404,p=4,cr=5,cu=0,mis=0,r=0,dep=1,og=4,plh=867914364,tim=2200911646    STAT  #139821999713040 id=1 cnt=0 pid=0 pos=1 obj=59 op='TABLE ACCESS FULL  BOOTSTRAP$ (cr=5 pr=4 pw=0 str=1 time=406 us)'       ***  2020-05-11T12:25:30.977832+08:00    Exception [type:  SIGSEGV, Address not mapped to object] [ADDR:0x0] [PC:0x10F4C112,  lmebucp()+34] [flags: 0x0, count: 1]    2020-05-11T12:25:31.026914+08:00    Incident 718445  created, dump file:/u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_718445/orcl_ora_14324_i718445.trc    ORA-07445:  exception encountered:core dump  [lmebucp()+34][SIGSEGV][ADDR:0x0][PC:0x10F4C112][Address not mapped to  object]

根据该报错,可以大概定位数据库重启之后报ORA-07445 lmebucp()+34错误不能正常启动是由于bootstrap$表异常导致.

BBED> p  kcvfhrdb    ub4  kcvfhrdb                                 @96       0x00400208       BBED> set block 523            BLOCK#           523       BBED> map    File:  F:\temp\12.2\1\system01.dbf (0)    Block:  523                                    Dba:0x00000000    ------------------------------------------------------------    KTB Data  Block (Table/Cluster)       struct  kcbh, 20  bytes                       @0       struct  ktbbh, 48  bytes                      @20       struct  kdbh, 14  bytes                       @68       struct  kdbt[1], 4  bytes                     @82       sb2  kdbr[20]                                @86       ub1  freespace[1015]                         @126       ub1  rowdata[7047]                           @1141       ub4  tailchk                                 @8188       BBED> p  *kdbr[1]    rowdata[6431]    -------------    ub1  rowdata[6431]                            @7572     0x3c       BBED> x /rnnc    rowdata[6431]                                @7572    -------------    flag@7572: 0x3c  (KDRHFL, KDRHFF, KDRHFD, KDRHFH)    lock@7573: 0x01    cols@7574:     0

通过bbed定位rootdba,然后dump相关block,随机找一条记录,确认bootstrap表无后效记录.但是该数据库在重启之前已经报了ORA-600 kkpo_rcinfo_defstg:delseg和ORA-30732错误,很可能还有其他的基表异常.通过先修复bootstrap$记录,然后根据该表中记录分析其他相关表记录,最终确定tab$中记录也异常,通过bbed 批量循环修复方法,对其进行恢复,open数据库,可以验证数据没有问题.至此该问题解决,但是没有找出来故障原因(是人为破坏【直接人工删除】,还是某种工具带入恶意脚本导致,类似:plsql dev引起的数据库被黑勒索比特币实现原理分析和解决方案,亦或者是数据库安装介质带有恶意程序,类似:警告：互联网中有oracle介质被注入恶意程序导致—ORA-60016703)