【1】问题背景：Oracle数据库版本为11.2.0.1，操作系统CentOS release 5.9，详细的报错信息如下：
Dump file /data/oracle/diag/rdbms/db01/db01/incident/incdir_70215/db01_ora_17588_i70215.trc
Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
ORACLE_HOME = /data/oracle/oracle11
System name: Linux
Node name: db01
Release: 2.6.18-348.el5
Version: #1 SMP Tue Jan 8 17:53:53 EST 2013
Machine: x86_64
Instance name: db01
Redo thread mounted by this instance: 1
Oracle process number: 52
Unix process pid: 17588, image: oracle@db01
*** 2014-06-20 22:33:42.577
*** SESSION ID:(204.33501) 2014-06-20 22:33:42.577
*** CLIENT ID:() 2014-06-20 22:33:42.577
*** SERVICE NAME:(SYS$USERS) 2014-06-20 22:33:42.577
*** MODULE NAME:() 2014-06-20 22:33:42.577
*** ACTION NAME:() 2014-06-20 22:33:42.577
Dump continued from file: /data/oracle/diag/rdbms/db01/db01/trace/db01_ora_17588.trc
ORA-03137: TTC 协议内部错误: [12333] [32] [49] [50] [] [] [] []
========= Dump for incident 70215 (ORA 3137 [12333]) ========
*** 2014-06-20 22:33:42.600
dbkedDefDump(): Starting incident default dumps (flags=0x2, level=3, mask=0x0)
----- Current SQL Statement for this session (sql_id=6v8rd0sq5a8m3) -----
select count(t.fd_id) as COUNT from sys_notify_todo_done_info t inner join
(select distinct(sysnotifyt1_.fd_subject) FD_SUBJECT,max(sysnotifyt0_.fd_id) as FD_ID
,sysnotifyt1_.fd_app_name,sysnotifyt1_.fd_type,sysnotifyt1_.fd_create_time
from sys_notify_todo_done_info sysnotifyt0_ inner join sys_notify_todo sysnotifyt1_
on sysnotifyt0_.fd_todoid=sysnotifyt1_.fd_id
where (sysnotifyt0_.fd_elementid = :1)
group by sysnotifyt1_.fd_subject,sysnotifyt1_.fd_app_name,sysnotifyt1_.fd_type,sysnotifyt1_.fd_create_time) f on t.fd_id = f.fd_id
----- Call Stack Trace -----
calling call entry argument values in hex
location type point (? means dubious value)
-------------------- -------- -------------------- ----------------------------
*** 2014-06-20 22:33:42.834
skdstdst()+36 call kgdsdst() 000000000 ? 000000000 ?
7FFF5BCF02B8 ? 000000001 ?
7FFF5BCF47B8 ? 000000000 ?
ksedst1()+98 call skdstdst() 000000000 ? 000000000 ?
7FFF5BCF02B8 ? 000000001 ?
000000000 ? 000000000 ?
ksedst()+34 call ksedst1() 000000000 ? 000000001 ?
7FFF5BCF02B8 ? 000000001 ?
【2】问题的解决方法
Metalink上面居然没有发现症状一样的问题，关于ORA-03137的错误，问题有很多，常见的问题有JDBC、绑定变量窥视、ORACLE的BUG等，Metalink也提供了一系列的排查方法，具体如下：
2.1 绑定变量的窥视
解决方法：SQL> alter system set "_optim_peek_user_binds"=false;
弊端：关闭绑定变量的窥视，会导致当SQL语句使用绑定变量的时候，产生的执行计划可能会出现不准，在生产环境中要长期关闭该特性需要认真测试；
2.2 更新数据库最新的PSU，当通过禁用绑定变量后数据库不再报错，可以应用每个数据库最新的PSU。或者当禁用绑定变量的时候也继续报错，也可以考虑应用PSU进行测试；
以下是相关的PSU信息：
PSU 11.1.0.7.8 Patch 12419384 includes Patch:9703463.
For 11.1.0.7, Patch 9243912 can be applied on top of 11.1.0.7.8 ( no conflict).
Patch:9703463 can also be applied individually but requires PSU 11.1.0.7.6
Patch:8625762 may also be applicable to databases version 11.1.0.7
2.3 更新数据库到11.2.0.3这个版本
在ORACLE 11.2.0.3这个版本中可以解决所有的ORA-3137的错误；
2.4 联系ORACLE SUPPORT解决这个问题 经过以上操作，你都踩到这个地雷了，说明楼主平时烧香少，需要联系ORALCE的大神么出手；
【3】总结：经过验证发现运行alter system set "_optim_peek_user_binds"=false后报错就消失了；