A little summary of ora 3113

[复制链接]
查看11 | 回复0 | 2005-10-30 17:05:33 | 显示全部楼层 |阅读模式
First I am sorry I have to use English becasue I am not good at Chinese input.
------------------------
Ora 3113 is a common problem. It does not tell you anything useful except that the connection is broken. Most likely
there will be a trace file or core file generated with 3113 error, look into this file for some useful hint.
I think this is the first thing (and most helpful) we should do.The following
summary is a list of some possible reasons. You can use it as a check list and I hope you can add more.
1. When installing Oracle
1.1 Process in the init.ora is set too large and the semaphore setting is small. (UNIX only).
Action: decrease process or increase semaphore
1.2 Memory problem
Action: Check swap area to see if it is too small and make sure share memory is not too small (UNIX)
1.3 DB_BLOCK_SIZE is invalid
Action: Set it to 2K, 4K, 8K, 16K or 32K
2. When running a batch or long-time import
2.1 You session is killed?
2.2 Network problem
2.3 expire_time is set >0 in sqlnet.ora?
Action: set it to 0.
3. Others
3.1 Oracle executables need relinking
3.2 Shared pool size too small
3.3 Check and remove SQLNET.AUTHENTICATION_SERVICES= (NTS) in sqlnet.ora (Windows 2000 platform)
3.4 Check if other client has the same problem to judge it is server problem or problem with this client
3.5 Check disk usage, make sure there is free space
3.6 Check oracle in the ?/bin directory is set to 6751 permission
3.7 There may be a corrupted record in a table and you hit it
4. There is quite a few bugs that will report 3113 error. Check it on metalink.
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

主题

0

回帖

4882万

积分

论坛元老

Rank: 8Rank: 8

积分
48824836
热门排行