From: Mathieu Bruneau (mathieu.bruneauargontechnologies.ca)
Date: Mon Jan 21 2008 - 21:21:34 CST

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

Mark Kozikowski a écrit :
>
>
> Hello all,
>
> I have been using MySQL for about 5 years now in a company project.
>
> I store a lot of data, very rapidly into the database.
>
> Presently, I am having a problem where the MySQL server appears to
> be denying a connection when I reach a database size of about
> 10 billion bytes.
>
> I am running a mostly default installation on Fedora core 4.
> We modified the blob size to 1 million for a special case. But for
> the failing operations, the blob size is only about 1.5K.
>
> I am storing about 3 millions records per hour, each averaging
> 1.5K.
>
> The DB is a single table with columns. 4 of which are integers.
> One is a blob.
>
> I am using RHFC 4 with ext3 file system.
>
> After running for about 2.5 hours, MySQL drops the connection and
> refuses to allow any others to the specified database.
>
> Are there any configurations I can adjust or look at that may enable
> me to extend my DB to more data storage?
>
> Mark Kozikowski
>
Hi Mark,
  First if you are using a 32 bits architecture, mysql by default use 32
bits to create is pointer for dynamic table. Thus creating a "false" but
effective limit of 4G of data in a table. Not sure if that's applicable
to your case. You can easily see, if that's your issue by doing a show
table status and look at the data_lenght and max_data_length.

If that's the case, there's a few possible solutions.. First you could
split your table to have your blog outside in but I don't know if that
would be efficient, depends of your application I guess :) Second you
could create the table with generous MAX_ROWS and AVG_ROW_LENGTH to tell
mysql to use bigger pointer. I don't know the real impact on performance
that will have but at least you won't be limited to 4G of data anymore!

Regards,

--
Mathieu Bruneau
aka ROunofF

===
GPG keys available http://rounoff.darktech.org

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql

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