oracle 10.2.4 遭遇bug 宕机

2012-1-29号下午15点30分左右。

一台核心数据库 遇到bug  ，数据库宕机。
解决方案  重启数据库.



这套oracle 系统已经启动了 oracle 转PG 数据库的程序了。
预计3月份上线。 


以下为aler_log　的错误记录：
Tue Jan 29 15:24:15 2013
Errors in file /data/oracle/admin/cmsdb/udump/cmsdb_ora_20568.trc:
ORA-00600: internal error code, arguments: [17018], [0], [], [], [], [], [], []
Tue Jan 29 15:24:18 2013
Errors in file /data/oracle/admin/cmsdb/udump/cmsdb_ora_20568.trc:
ORA-00600: internal error code, arguments: [1115], [], [], [], [], [], [], []
ORA-00600: internal error code, arguments: [17018], [0], [], [], [], [], [], []
Tue Jan 29 15:24:22 2013
Errors in file /data/oracle/admin/cmsdb/bdump/cmsdb_pmon_4759.trc:
ORA-00600: internal error code, arguments: [1115], [], [], [], [], [], [], []
Tue Jan 29 15:24:23 2013
Errors in file /data/oracle/admin/cmsdb/bdump/cmsdb_pmon_4759.trc:
ORA-00600: internal error code, arguments: [1115], [], [], [], [], [], [], []
Tue Jan 29 15:24:23 2013
PMON: terminating instance due to error 472
Tue Jan 29 15:24:23 2013
Errors in file /data/oracle/admin/cmsdb/bdump/cmsdb_lns2_11584.trc:
ORA-00472: PMON  process terminated with error
Tue Jan 29 15:24:23 2013
Errors in file /data/oracle/admin/cmsdb/bdump/cmsdb_lns1_4820.trc:
ORA-00472: PMON  process terminated with error
Tue Jan 29 15:24:23 2013
Errors in file /data/oracle/admin/cmsdb/bdump/cmsdb_psp0_4761.trc:
ORA-00472: PMON  process terminated with error

ORA-600 [1115] In Alertlog With PMON Terminating Instance Due To Error 472 [ID 1078393.1]     … 
  

一下为 metalink 的说明
--------------------------------------------------------------------------------

修改时间:2011-10-12类型:PROBLEM状态:PUBLISHED优先级:3        注释 (0) 转到底部



In this Document
  Symptoms
  Cause
  Solution
  References



--------------------------------------------------------------------------------



Applies to:
Oracle Server - Enterprise Edition - Version: 10.2.0.3 to 11.1.0.6 - Release: 10.2 to
 11.1
Information in this document applies to any platform.
***Checked for relevance on 11-10-2011***
Symptoms
The following error is experienced in the alertlog:

ORA-00600: internal error code, arguments: [1115], [], [], [], [], [], [], []

With PMON terminating the instance with error 472

The ORA-600 [1115] stack trace will resemble the following:

ksedmp



Cause
This is caused by Bug 7566075 which was closed as a duplicate of unpublished bug 7131656 where an ORA-600 [16605] can occur when loading a library cache object.

This is fixed in the 11.1.0.7 and 10.2.0.5 patchsets.




Solution
1.  If on 11.1, apply the 11.1.0.7 patchset or higher.

2. If on 10.2, apply the 10.2.0.5 patchset or higher
    **(Note that this patch was not available at time of writing (Mar 2010))**

OR

3.  If available for your platform. and version, download and apply Patch 7131656


References
BUG:7566075 - DATABASE CRASHED WITH ORA-7445 [KGLPIM [KGSSDH() ORA-600 [1115]
NOTE:7131656.8 - Bug 7131656 - OERI[16605] loading a library cache object
NOTE:756746.1 - ORA-600 [16605]

---下面是一段trace 文件
call pin=(nil) session pin=(nil) hpc=0000 hlc=0000
      htl=0x4b3310fb0[0x4cbda9be0,0x4b33114b8] htb=0x4cbda9be0 ssga=0x4cbda8c68
      user=55ce58e48 session=55ce58e48 count=1 flags=[0000] savepoint=0x510778de
      LIBRARY OBJECT HANDLE: handle=4e6d11bd0 mtx=0x4e6d11d00(1) cdp=1
      name=select id, title, pub_date, author, inputuserid, pub_url, comment_script, iswap, wap_id, keyword, channel_id from tblarticle where id = :1
      hash=5112109de7bdb0faa9ffa1ea280650e6 timestamp=12-21-2012 10:26:03
      namespace=CRSR flags=RON/KGHP/TIM/PN0/MED/KST/DBN/MTX/[500100d0]
      kkkk-dddd-llll=0000-0001-0001 lock=N pin=0 latch#=3 hpc=27b8 hlc=27b8
      lwt=0x4e6d11c78[0x4e6d11c78,0x4e6d11c78] ltm=0x4e6d11c88[0x4e6d11c88,0x4e6d11c88]
      pwt=0x4e6d11c40[0x4e6d11c40,0x4e6d11c40] ptm=0x4e6d11c50[0x4e6d11c50,0x4e6d11c50]
      ref=0x4e6d11ca8[0x4e6d11ca8,0x4e6d11ca8] lnd=0x4e6d11cc0[0x4e6d11cc0,0x4e6d11cc0]
        LIBRARY OBJECT: bject=510cd5738
        type=CRSR flags=EXS[0001] pflags=[0000] status=VALD load=0
        CHILDREN: size=16
        child#    table reference   handle
        ------ -------- --------- --------
             0 510cd5200 510cd4e70 4e612e5a0
        DATA BLOCKS:
        data#     heap  pointer    status pins change whr
        ----- -------- -------- --------- ---- ------ ---
            0 4e6d11b10 510cd5850 I/P/A/-/-    0 NONE   00
      ----------------------------------------
      SO: 0x4b33115a8, type: 53, owner: 0x55ce58e48, flag: INIT/-/-/0x00
      LIBRARY OBJECT LOCK: lock=4b33115a8 handle=55da60e40 mode=N
      call pin=(nil) session pin=(nil) hpc=0000 hlc=0000
      htl=0x4b3311628[0x4ccf9d0e0,0x4ccf9efc8] htb=0x4cbda92b0 ssga=0x4cbda8c68
      user=55ce58e48 session=55ce30c18 count=0 flags=LRU/[4000] savepoint=0x79

有可能ORACLE内部在请求 library object lock 的时候出发了bug
导致pmon 进程被终止，从而引发数据库宕机

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/133735/viewspace-753623/，如需转载，请注明出处，否则将追究法律责任。

转载于:http://blog.itpub.net/133735/viewspace-753623/