ORA-07445: core dump kslgetl

 昨晚一套核心库的一个节点宕掉，然后reboot了，

在alert里面发现如下信息：

Thu Jul  5 03:03:50 2012
Errors in file /u01/oracle/app/oracle/admin/crm/bdump/crm1_dbw9_14426.trc:
ORA-07445: exception encountered: core dump [kslgetl()+32] [SIGSEGV] [Address not mapped to object] [0xF27D6E41F020E369] [] []
Thu Jul  5 03:03:54 2012
Errors in file /u01/oracle/app/oracle/admin/crm/bdump/crm1_pmon_14238.trc:
ORA-00471: DBWR process terminated with error
Thu Jul  5 03:03:54 2012
Errors in file /u01/oracle/app/oracle/admin/crm/bdump/crm1_ckpt_14549.trc:
ORA-00471: DBWR process terminated with error
Thu Jul  5 03:03:54 2012
Errors in file /u01/oracle/app/oracle/admin/crm/bdump/crm1_lgwr_14540.trc:
ORA-00471: DBWR process terminated with error
Thu Jul  5 03:03:54 2012
Errors in file /u01/oracle/app/oracle/admin/crm/bdump/crm1_lms0_14351.trc:
ORA-00471: DBWR process terminated with error
Thu Jul  5 03:03:54 2012
PMON: terminating instance due to error 471
Thu Jul  5 03:03:58 2012
Shutting down instance (abort)
License high water mark = 3882
Thu Jul  5 03:04:00 2012
Instance terminated by PMON, pid = 14238
Thu Jul  5 03:04:03 2012
Instance terminated by USER, pid = 14287
Thu Jul  5 03:04:30 2012
Starting ORACLE instance (normal)

出现ORA-07445，第一反应就是去看trace，以下是call stack trace：
kslgetl>kclmvreqbg>kclrwrite>kcbbxsv>kcbb_coalesce>kcbbwlru>
在metalink上发现在10.2.0.3上的类似bug，5932514,5879114，最后根据call stack trace信息发现和unpublished的bug:4637902的描述更靠近。