DB2崩溃了?!? [英] DB2 crash?!?

查看:110
本文介绍了DB2崩溃了?!?的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

嘿所有,


在我工作的公司我们昨天遇到了一个奇怪的情况。

我们的DB2数据库被锁定了,或者后来似乎,DBM,或者一些

连接管理器。


我们无法打开与数据库的新连接,但是现有的

连接仍然存在。所以实例工作正常。


另外,我无法使用IPC连接,所以我无法关闭数据库

方式。


DB2Diag没有提供有用的信息,也没有OS日志(linux 2.6)

仍不确定它是否是由操作系统或数据库造成的。


有没有人知道导致它的原因?

-R-


运行DB2 8.2 FP8 SuSE EE 9.1

2GB Ram,1 AMD Opteron CPU,80GB的微小测试突袭。

解决方案

" Jurgen Haan" ; <菊**** @ fake.dom>在消息中写道

news:42 ********************* @ news.xs4all.nl ...

嘿所有,

在我工作的公司,我们昨天遇到了一个奇怪的情况。
我们的DB2数据库被锁定了,或者后来看起来像是DBM,或者是一些
连接管理器。

我们无法打开与数据库的新连接,但现有连接仍然存在。所以实例工作正常。

另外,我无法使用IPC连接,所以我无法以干净的方式关闭数据库。

仍然不确定它是否是由操作系统或数据库引起的。

有没有人知道是什么导致的它?-R-
在SuSE EE 9.1上运行的DB2 8.2 FP8
2GB Ram,1个AMD Opteron CPU,80GB的微小测试raid。



这绝对不是好事,但不是崩溃。


如果db2diag.log没有提供任何信息,你可能需要增加

诊断水平(dbm parm)。但它可能与

达到允许的最大连接数有关,无论是通过配置还是通过

许可证限制。下次我会尝试取消(强制)部分或全部

个别应用程序句柄,看看我是否可以在

之后建立新的连接。

我不相信在SuSE EE 9.1发行版中曾经支持过DB2,

但我可能错了。但它确实让我对这个许可证感到疑惑。


Mark A写道:


这绝对不是好事,但它是没有崩溃。

如果db2diag.log没有提供任何信息,您可能需要提高
诊断级别(dbm parm)。但它可能与
达到允许的最大连接有关,无论是通过配置还是通过许可证限制。下次我会尝试取消(强制)部分或全部
个别应用程序句柄,看看我是否可以在
之后建立新的连接。

我不相信在SuSE EE 9.1发行版中曾经支持DB2,
但我可能错了。但它确实让我对这个许可证感到疑惑。




在锁定期间,我尝试了一些东西。

首先,我关闭了大部分客户连接。但是这些结果仍然存在(非常奇怪,没有等待或等待任何事情)。

其次我刚刚关闭了ETH界面。

仍然..连接仍然存在(在关闭网络接口上)。


我似乎连接管理器处于无限循环中,除了

CPU使用率,是正常的。


没有一个进程会对信号作出反应,甚至没有一个kill -9

我在本地开始的所有事情(db2 cli,db2stop)都会产生在一个

挂起过程中,所以我无法杀死cli,也不能杀死db2stop。

经过几次尝试,我用我剩下的1个控制台重新启动<机器和交叉机器没有发生腐败(它没有)。


我们拥有的许可证是每CPU(所以没有用户限制)。

设置MAX USER = MAX代理等..是200

在问题期间打开连接,少于20个。


I将提高DIAG的详细程度。

B.我希望这是一次性的问题。


这个数据库实例运行了4个月,唯一改变的是b / b $ b更改的事情是我最近添加了一张包含6M +记录的表格。


仍然不知道发生了什么。


-R-


Jurgen Haan写道:

Mark A写道:


这绝对不好,但它不是崩溃。

如果db2diag.log没有提供任何信息,您可能需要提高诊断级别(dbm parm)。但它可能与通过
配置或许可证限制达到允许的最大连接数有关。下次我会尝试取消
(强制)一些或所有单独的应用程序句柄,看看我是否可以在此之后建立新的连接。

我不相信在SuSE EE 9.1
发行版中曾经支持DB2,但我可能错了。但它确实让我对这个许可证感到疑惑。



在锁定期间,我尝试了一些事情。
首先,我关闭了大多数连接的客户端。但是这些内容仍然存在(非常奇怪,没有关闭等待或其他任何事情)。
其次我关闭了ETH界面。
仍然......连接仍然存在(在关闭网络接口上)。 br />
我似乎连接管理器处于无限循环中,除了CPU使用率,这是正常的。

没有任何进程会对信号作出反应,甚至没有杀死-9
我在本地开始的一切(db2 cli,db2stop)会导致挂起过程,所以我不能杀死cli,也不能杀死db2stop。
之后几次尝试,我用我剩余的1个控制台重新启动机器并交叉我的手指没有发生腐败(它没有)。

我们拥有的许可证是每CPU(所以没有用户限制。
设置MAX USER = MAX Agent等..是200
问题期间打开连接,少于20个。

我会提高的详细程度DIAG。
但希望这是一次性问题。

这个数据库实例是wa运行了4个月,唯一改变的是我最近添加了一张带有6M +记录的表格。

仍然不知道发生了什么。

-R -



什么显示内存可视化工具?

最好的办法是提高diaglog级别。

i有一些崩溃我做了之后非常好的sql语句

正常一切正常。

有时一个ipclean命令有助于开发人员制作一些可怕的

命令比这个经常帮助。

查看你系统上的数据堆栈可能你没有正确的

参数。
http://www-128.ibm.com/developerwork .. .ex.html#N108DA

只是内存设置的链接。


请发贴db cfg和你的dbm cfg这可以帮助我们也是。


cu,


indika


Hey all,

at the company where I work we had a strange situation yesterday.
Our DB2 database locked up, or as it later seemed, the DBM, or some
connection manager.

We couldn''t open new connections to the database, but existing
connections remained. So the instances were working fine.

Also, I couldn''t connect using IPC, so I was unable to shut the database
down in a clean way.

DB2Diag provided no usable information, nor the OS logs (linux 2.6)
Still not sure if it was caused by the OS or the DB.

Does anyone have a clue on what caused it?
-R-

DB2 8.2 FP8 running on SuSE EE 9.1
2GB Ram, 1 AMD Opteron CPU, Tiny test raid of 80GB.

解决方案

"Jurgen Haan" <ju****@fake.dom> wrote in message
news:42*********************@news.xs4all.nl...

Hey all,

at the company where I work we had a strange situation yesterday.
Our DB2 database locked up, or as it later seemed, the DBM, or some
connection manager.

We couldn''t open new connections to the database, but existing connections
remained. So the instances were working fine.

Also, I couldn''t connect using IPC, so I was unable to shut the database
down in a clean way.

DB2Diag provided no usable information, nor the OS logs (linux 2.6)
Still not sure if it was caused by the OS or the DB.

Does anyone have a clue on what caused it?
-R-

DB2 8.2 FP8 running on SuSE EE 9.1
2GB Ram, 1 AMD Opteron CPU, Tiny test raid of 80GB.



That is definitely not good, but it is not a crash.

If the db2diag.log does not provide any info, you might need to increase the
diagnosis level (dbm parm). But it probably had something to do with
reaching the maximum connections allowed, either by configuration or by
license limits. Next time I would try to cancel (force) some or all
individual application handles to see if I could make new connections after
that.

I don''t believe that DB2 was ever supported under SuSE EE 9.1 distributions,
but I could be wrong. But it does make me wonder about the license.


Mark A wrote:


That is definitely not good, but it is not a crash.

If the db2diag.log does not provide any info, you might need to increase the
diagnosis level (dbm parm). But it probably had something to do with
reaching the maximum connections allowed, either by configuration or by
license limits. Next time I would try to cancel (force) some or all
individual application handles to see if I could make new connections after
that.

I don''t believe that DB2 was ever supported under SuSE EE 9.1 distributions,
but I could be wrong. But it does make me wonder about the license.



During the lockup, I tried a few things.
First, I closed most of the clients connected. But the conntections
remained (very strange, not closed waiting or anything).
Secondly I just closed the ETH interface.
Still... Connections remained (on a shutdown network interface).

I seemed that a connection manager was in a endless loop, except for the
CPU usage, which was normal.

None of the processes would react on signals, not even a kill -9
everything I started locally (db2 cli, db2stop) would result in a
hanging process, so I couldn''t kill the cli, nor the db2stop.
After a few attempts, I used my 1 remaining console to reboot the
machine and cross my fingers no corruption had occured (it didn''t).

The license we have is Per CPU (so no user limit).
The set MAX USER = MAX Agent etc.. is 200
Open connections during problem, less than 20.

I will raise the verboseness of the DIAG.
But hope it was a one time problem.

This instance of the database was running for 4 months, and the only
thing that changed was that I recently added a table with 6M+ records.

Still no clue on what happened.

-R-


Jurgen Haan wrote:

Mark A wrote:


That is definitely not good, but it is not a crash.

If the db2diag.log does not provide any info, you might need to
increase the diagnosis level (dbm parm). But it probably had something
to do with reaching the maximum connections allowed, either by
configuration or by license limits. Next time I would try to cancel
(force) some or all individual application handles to see if I could
make new connections after that.

I don''t believe that DB2 was ever supported under SuSE EE 9.1
distributions, but I could be wrong. But it does make me wonder about
the license.


During the lockup, I tried a few things.
First, I closed most of the clients connected. But the conntections
remained (very strange, not closed waiting or anything).
Secondly I just closed the ETH interface.
Still... Connections remained (on a shutdown network interface).

I seemed that a connection manager was in a endless loop, except for the
CPU usage, which was normal.

None of the processes would react on signals, not even a kill -9
everything I started locally (db2 cli, db2stop) would result in a
hanging process, so I couldn''t kill the cli, nor the db2stop.
After a few attempts, I used my 1 remaining console to reboot the
machine and cross my fingers no corruption had occured (it didn''t).

The license we have is Per CPU (so no user limit).
The set MAX USER = MAX Agent etc.. is 200
Open connections during problem, less than 20.

I will raise the verboseness of the DIAG.
But hope it was a one time problem.

This instance of the database was running for 4 months, and the only
thing that changed was that I recently added a table with 6M+ records.

Still no clue on what happened.

-R-


what shows the memory visualizer?
best thing is to upper the diaglog level.
i had some crashes on really good sql statements after i made them
proper everything works fine.
sometimes an ipclean command helps if developers make some terrible
commands than this helps often.
look at your data stack on your system maybe you have not the right
parameters.
http://www-128.ibm.com/developerwork...ex.html#N108DA
just a link for memory settings.

please post you db cfg and your dbm cfg this could help us, too.

cu,

indika


这篇关于DB2崩溃了?!?的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!

查看全文
登录 关闭
扫码关注1秒登录
发送“验证码”获取 | 15天全站免登陆