ORA-07445错误

今天发现数据库里的cdump目录下产生了一对的CORE文件。

查看alert日志发现都是

…… Wed Oct 12 09:58:51 2011 Restarting dead background process CJQ0 Wed Oct 12 09:58:52 2011 CJQ0 started with pid=43, OS id=5247 Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x130000001E] [PC:0x6E678E4, kslgetl()+40] Errors in file /home/oracle/diag/rdbms/orarpt/orarpt/trace/orarpt_cjq0_5247.trc (incident=276255): ORA-07445: exception encountered: core dump [kslgetl()+40] [SIGSEGV] [ADDR:0x130000001E] [PC:0x6E678E4] [Address not mapped to object] [] Wed Oct 12 09:58:59 2011 Sweep Incident[276255]: completed

接着查看相应的trace文件

oracle@linux-35:~/diag/rdbms/orarpt/orarpt/trace> more /home/oracle/diag/rdbms/orarpt/orarpt/trace/orarpt_cjq0_5247.trc Trace file /home/oracle/diag/rdbms/orarpt/orarpt/trace/orarpt_cjq0_5247.trc Oracle Database 11g Enterprise Edition Release 11.1.0.6.0 - 64bit Production With the Partitioning option ORACLE_HOME = /home/oracle/product/11.1.0/db_1 System name: Linux Node name: linux-35 Release: 2.6.16.21-0.8-smp Version: #1 SMP Mon Jul 3 18:25:39 UTC 2006 Machine: x86_64 Instance name: orarpt Redo thread mounted by this instance: 1 Oracle process number: 43 Unix process pid: 5247, image: oracle@linux-35 (CJQ0) *** 2011-10-12 09:58:55.533 *** SESSION ID:(1197.56623) 2011-10-12 09:58:55.533 *** CLIENT ID:() 2011-10-12 09:58:55.533 *** SERVICE NAME:(SYS$BACKGROUND) 2011-10-12 09:58:55.533 *** MODULE NAME:() 2011-10-12 09:58:55.533 *** ACTION NAME:() 2011-10-12 09:58:55.533 Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x130000001E] [PC:0x6E678E4, Exception [type: SIGSEGV, Address not map ped to object] [ADDR:0x130000001E] [PC:0x6E678E4, kslgetl()+40] DDE: Problem Key 'ORA 7445 [kslgetl()+40]' was flood controlled (0x6) (incident: 276255) ORA-07445: exception encountered: core dump [kslgetl()+40] [SIGSEGV] [ADDR:0x130000001E] [PC:0x6E678E4] [Address not mapped to objec t] [] ssexhd: crashing the process... Background_Core_Dump = PARTIAL

感觉应该是job进程有问题，同事也反映，从11号开始job都不能正常运行。

上网查了资料，也没找到相关原因，大部分都说是oracle BUG。无奈只得重启下数据库试试，重启后发现job进程正常了，所有job都能正常执行。

具体原因不明，先在此做个记录，后续有相关答案再写。