[NTLUG:Discuss] Capabilities of MySQL

[Mike]

Correction: http://alphalinix.org, not visa versa. And of course, all
ridiculously large calculations assume you have equivalent ridiculously
large amounts of storage---Even Sun's Marketing department can't get 8T
on a 6 Gig drive.

Mike wrote:
> 
> Steve Jackson wrote:
> >
> > I am designing a database driven app (on Linux of course), and I was
> > wondering if MySQL is a capable database engine when accessing tables
> > with 500,000 records (or 1,000,000)?  If no, what is the largest table
> > size that MySQL can handle?  Thanks.
> >
> > _______________________________________________
> > http://ntlug.org/mailman/listinfo/discuss
> 
> It would seem that you need to consider more than the number of records
> alone,  specifically the product of the record number and the record
> size. Each table's data is stored in its own file. I believe that to
> access the millionth record in a table, you are effectively calculating
> the offset in the data file by record_size*1E6, which cannot be greater
> than 2 billion, 4 billion, or 8E12, depending on your architecture and
> OS. The MySQL manual sheds a little light on this for several platforms:
> http://www.mysql.com/Manual/manual_toc.html#Table_size
> 
> So if you had a stock Linux i386 (2 Gig max file size), and a table
> which has say, 10 fields which are each 30 characters wide (10*30=300
> bytes), then you can hold
> 
> 2,147,483,648 (bytes/table) / 300 (bytes/rec) = 7,158,277 records
> 
> in this paritcular table.
> 
> Now, if Linux Alpha had a true 8T limit, you could get somewhere in the
> line of:
> 
> 8E12 / 300 = 26.6 billion records
> 
> in this particular table.
> 
> A significant improvement to say the least. However, odds are you are
> more likely going to see a 1T limit for Linux alpha, if that much. You
> may try the linuxalpa.org, but I couldn't seem to find the max file size
> anywhere. I called Red Hat, wondering if they could give me a cheap and
> easy answer, but the best a got was "real big, to be sure," which tends
> to make calculations a little vague.
> 
> I am shooting from the hip, but this should get you in the ballpark
> based on the given information. As long as you are dealing with tables
> which don't contain wildly variable BLOBS, you should be able to
> reasonably determine your limitations.
> 
> So far, I have really enjoyed MySQL. I have used it in Apache/mod_perl,
> as well as other applications. The Windows ODBC drivers work like a
> charm, so people on the LAN who have MS Access can connect to the
> central MySQL database and do reports. However, I have not dealt in
> great detail with Access 2000. I think that when I did, I had some ODBC
> problems, but perhaps I just hadn't upgraded the ODBC drivers. Best of
> luck.
> 
> --
>           _  _    _  __  ___
>          / \/ \  | |/ / / O \
>         /_/\/\_\o|_|\_\o\___/o
>         M I C H A E L O W E N S
> 
> _______________________________________________
> http://ntlug.org/mailman/listinfo/discuss

-- 
	  _  _    _  __  ___    
	 / \/ \  | |/ / / O \   
	/_/\/\_\o|_|\_\o\___/o  
	M I C H A E L O W E N S