问题描述:
开发人员报告数据库时快时慢,下午应用报错,系统死掉,现场重启库。
分析:
alert日志:
Thu Nov 01 08:39:12 2007
Errors in file e:/oracle/admin/gbbcrm/udump/gbbcrm_s000_3688.trc:
ORA-00600: 内部错误代码,参数: [1236], [], [], [], [], [], [], []
Thu Nov 01 08:40:00 2007
Errors in file e:/oracle/admin/gbbcrm/udump/gbbcrm_s000_3688.trc:
ORA-00600: 内部错误代码,参数: [1236], [], [], [], [], [], [], []
Errors in file e:/oracle/admin/gbbcrm/udump/gbbcrm_s002_4232.trc:
ORA-00600: 内部错误代码,参数: [16365], [0x1D6F1A30], [0], [4], [], [], [], []
Fri Nov 02 13:41:29 2007
Error 600 trapped in 2PC on transaction 20.0.2. Cleaning up.
Error stack returned to user:
ORA-00600: 内部错误代码,参数: [16365], [0x1D6F1A30], [0], [4], [], [], [], []
应用报错:
ORA-04031: 无法分配 8132 字节的共享内存 (”large pool”,”unknown object”,”cursor work he”,”qesaQBInit:buffer”)
metalink中用600那个工具搜索了一下,找到两篇相关文档:
138354.1
106607.1
说的基本都是关于共享服务器连接模式下的错误,再看看应用抛的04031,问题确认基本出在应用连数据库的模式问题上了,用lsnrct service确认,果然共享连接占大比例,目前把应用修改成专用的再观察一段。