今天，用友NC的一个数据库突然挂了，查看数据库日志发现，在凌晨数据库已经挂了，今天上午10点被同事启动了，然后中午又挂了，具体告警信息如下：

Sat Mar 02 12:37:30 2019

Exception [type: SIGSEGV, SI_KERNEL(general_protection)] [ADDR:0x0] [PC:0x186016B, opiaba()+639] [flags: 0x0, count: 1]

Errors in file /u01/app/oracle/diag/rdbms/ncdb3/ncdb3/trace/ncdb3_ora_101082.trc  (incident=72265):

ORA-07445: exception encountered: core dump [opiaba()+639] [SIGSEGV] [ADDR:0x0] [PC:0x186016B] [SI_KERNEL(general_protection)] []

Incident details in: /u01/app/oracle/diag/rdbms/ncdb3/ncdb3/incident/incdir_72265/ncdb3_ora_101082_i72265.trc

Use ADRCI or Support Workbench to package the incident.

See Note 411.1 at My Oracle Support for error and packaging details.

Sat Mar 02 12:38:09 2019

Sweep [inc][72265]: completed

Sweep [inc2][72265]: completed

Sat Mar 02 12:38:10 2019

Dumping diagnostic data in directory=[cdmp_20190302123810], requested by (instance=1, osid=101082), summary=[incident=72265].

Sat Mar 02 12:38:43 2019

Errors in file /u01/app/oracle/diag/rdbms/ncdb3/ncdb3/trace/ncdb3_pmon_98334.trc  (incident=72017):

ORA-00600: internal error code, arguments: [17147], [0x9194B8448], [], [], [], [], [], [], [], [], [], []

Incident details in: /u01/app/oracle/diag/rdbms/ncdb3/ncdb3/incident/incdir_72017/ncdb3_pmon_98334_i72017.trc

Use ADRCI or Support Workbench to package the incident.

See Note 411.1 at My Oracle Support for error and packaging details.

Sat Mar 02 12:38:43 2019

Dumping diagnostic data in directory=[cdmp_20190302123843], requested by (instance=1, osid=98334 (PMON)), summary=[incident=72017].

Errors in file /u01/app/oracle/diag/rdbms/ncdb3/ncdb3/trace/ncdb3_pmon_98334.trc:

ORA-00600: internal error code, arguments: [17147], [0x9194B8448], [], [], [], [], [], [], [], [], [], []

PMON (ospid: 98334): terminating the instance due to error 472

System state dump requested by (instance=1, osid=98334 (PMON)), summary=[abnormal instance termination].

System State dumped to trace file /u01/app/oracle/diag/rdbms/ncdb3/ncdb3/trace/ncdb3_diag_98344_20190302123844.trc

Instance terminated by PMON, pid = 98334

查看trace文件，发现可疑SQL，一共34万多行，大概6万多条UPDATE操作，SQL如下：

*** 2019-03-02 12:37:30.547

dbkedDefDump(): Starting a non-incident diagnostic dump (flags=0x3, level=3, mask=0x0)

----- Current SQL Statement for this session (sql_id=4yj2p6s0z2m2q) -----

begin

          UPDATE ic_saleoutdetail t

            SET t.proestdetailid=:1,t.isproest=:2

           WHERE

           t.ID = :3

          ;

          UPDATE ic_saleoutdetail t

            SET t.proestdetailid=:4,t.isproest=:5

           WHERE

           t.ID = :6

          ;

          … …

          … …

          UPDATE ic_saleoutdetail t

            SET t.proestdetailid=:204670,t.isproest=:204671

           WHERE

           t.ID = :204672

          ;end;

从begin到end一共24万多的绑定变量，这个数据库是11.2.0.4.0版本，一个SQL最多只支持65535个绑定变量，超过这个阈值数据库就可能挂掉。通常情况下基本也遇不到这么多绑定变量的情况。

解决方法是拆分上面的SQL，和开发的同事沟通后，他们确定可以拆分，目前开发的同事正在调整程序。虽然这个是BUG，但绝大多数的BUG都是SQL书写不规范触发的。

本文固定链接: http://www.dbdream.com.cn/2019/03/ora-07445opiaba639%e3%80%81ora-0060017147%e9%94%99%e8%af%af%e5%af%bc%e8%87%b4%e6%95%b0%e6%8d%ae%e5%ba%93%e6%8c%82%e6%8e%89/ | 信春哥，系统稳，闭眼上线不回滚！