From: Dittmar, Daniel (daniel.dittmarsap.com)
Date: Fri Mar 18 2005 - 08:22:24 CST

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

>I'm currently developing a multithreaded python application
>which uses MaxDB
>as database. It's been a very pleasant experience using MaxDB
>for the task.
>However when running stress-tests on the system the
>application exits with a
>segfault.
>

Do you happen to have MS VC++ installed? I could send you a debug
version so that I could see where the segfault occurs.

>I'm using a connection pool to handle the connections, and I'm
>sure that no
>threads are sharing the connections at the same time.

Could you configure your connection pool such that you allocate all
sessions at the beginning? Just recently, I had this problem on Unix
that connects would interfere with existing connections. I looked at the
equivalent Windows code and it looked OK, but it would be great if we
could definitely exclude this as a possible cause of the error.

> For testing purposes
>I've even wrapped the calls to the 'cursor.execute' with thread locks.

cursor.execute isn't enough as the fetch* methods also access the layer
written in C.

Could you send me
- the structure of the result sets (column types including legth)
- the number of threads you're using
- if your program is running on a multi cpu machine

I'll try to reproduce the problem locally. If I don't see anything, I'll
have to send you a new driver that can write traces.

Daniel Dittmar

--
Daniel Dittmar
SAP Labs Berlin
daniel.dittmarsap.com

--
MaxDB Discussion Mailing List
For list archives: http://lists.mysql.com/maxdb

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]