7 messages in com.mysql.lists.win32Re: mysql and Dr. Watson for Win NT| From | Sent On | Attachments |
|---|---|---|
| Jon Cline | 06 Apr 2000 11:14 | |
| Michael Ellis | 06 Apr 2000 11:41 | |
| sin...@mysql.com | 07 Apr 2000 04:22 | |
| Peter Strand | 27 Apr 2000 14:48 | |
| sin...@mysql.com | 28 Apr 2000 04:54 | |
| Peter Strand | 02 May 2000 04:16 | |
| Michael Ellis | 02 May 2000 08:47 |
| Subject: | Re: mysql and Dr. Watson for Win NT![]() |
|---|---|
| From: | Peter Strand (pet...@femte-gear.dk) |
| Date: | 04/27/2000 02:48:06 PM |
| List: | com.mysql.lists.win32 |
----- Original Message ----- From: <sin...@mysql.com> To: Jon Cline <jo...@clinecorp.com>; <win...@lists.mysql.com> Sent: Friday, April 07, 2000 1:22 PM Subject: Re: mysql and Dr. Watson for Win NT
On Thu, 06 Apr 2000, Jon Cline wrote:
I have had problems lately with MySQL on Windows NT, Service Pack 5 giving the
following error:
*** Dr. Watson for Windows NT
An application error has occurred and an application error log is being
generated.
MySQLd-nt.exe
Exception: access violation (0xc0000005),Address: 0x0045d7e3
***
I have included a copy of my current my.cnf file since I have installed mysql on
my D: drive. Please give >any feedback that may be causing this. Currently it
is happeing every 7 minutes or so.
***
# Example mysql config file. # Copy this file to c:\my.cnf to set global options # # One can use all long options that the program supports. # Run the program with --help to get a list of available options
# This will be passed to all mysql clients [client] #password=my_password port=3306 #socket=MySQL
# Here is entries for some specific programs # The following values assume you have at least 32M ram
# The MySQL server [mysqld] port=3306 #socket=MySQL skip-locking set-variable = record_buffer=1M set-variable = sort_buffer=4M set-variable = table_cache=256 set-variable = key_buffer=64M set-variable = max_allowed_packet=48M set-variable = thread_stack=128K set-variable = flush_time=1800
# Uncomment the following row if you move the MySQL distribution to another # location basedir = d:/mysql/
[mysqldump]
[mysql]
[isamchk]
[client_fltk]
***
jon cline jo...@enthusiastonline.com
Hi!
Regarding your my.cnf, please keep in mind that record buffer is allocated on each thread, as well as thread_stack. sort_buffer is allocated for each SELECT query which has DISTINCT or ORDERY BY or GROUP BY. Next, max_allowed_packet is too large. It can not be greater then 24 Mb.
How many threads did you have when you got that error. Having in mind how have you have allocated buffers, may be you ran out of memory.
Please tune up your MySQL variables, and after that if you can build the test case that would always reproduce the error , we would like to take a look at it.
Please, stay in touch.
Regards,
Sinisa
We have the same bug. Except the exact error message is "access violation
(0xc0000014),Address: 0x0045d7e3"
Our setup is running on a two-processor (Intel P3) windows 2000 adv. server.
MySQL version 3.22.32, using mysqld-nt
We have only seen this occur a couple of times, but there is a possible pattern.
We have a set of applications working on a number of databases. These
applications are all native win32 executeables, accessing mysql via libmysql (
opt ) - they are all done in Delphi. Some run background processing, while
others are CGI applications activated from the web server. We have not seen the
error while doing intensive testing, but after we started running in a realistic
test environment, they suddenly appeared. All of them occured when there was a
background processing application running while there was some activity from CGI
applications. When the exception occurs, the error dialog appears. But only the
most active database is actually unavailable - the mysqld-nt actually still
responds correctly if accessing another database than the one that "crashes". It
seems that the one which is "crashing" is the most active at the moment. The
applications trying to access the dead database just hangs. I think they time
out eventually.
I know this is not much - I will try to reproduce it in our local test
environment on a single processor machine. Any advice on what information I
should concentrate on gathering ?
Thanks, Peter Strand, Femte Gear




