Re: Large Databases

From: mkavis <mkavis_at_cftnet.com>
Date: 1996/04/01
Message-ID: <315FEDDD.2A3D_at_cftnet.com>#1/1


Kevin Konynenbelt wrote:
>
> I am looking for a reference site which has implemented SQL Server 6.x
> for a fairly large database(s) - at least 10 Gig, and hopefully up to
> about 20 - 40 GB. The question being pondered is whether SQL Server
> will be up to the task when dealing with this volume of data ( as
> opposed to Oracle).
>
> Any leads or assistance would be appreciated.
>
> Kevin Konynenbelt
>
> kevink_at_minerva.ca
>
> Minerva Technology Inc.
> 600, 777 - 8th Ave. SW
> Calgary, Alberta, Canada T2P 3R5
> (403) 263-7533
>
> http://www.minerva.caWe have a database that has grown to 60 Gig. We are working on moving off of SQL
server as we speak. Some of the things I don't like about SQL Server:

  1. Slow....we bench marked against several other databases and it got beat badly (30 gig database was used).
  2. Commit logic...If the first transaction after a commit does not finish or takes along perios of time to finish, all preceding transactions get queued up in the transaction log until the transaction is cleared. We import UPC level data which equates to over a million rows a day. If the first transaction gets botched up we blow the transaction log. Other databases like oracle will commit the preceding transactions.
  3. We seem to get corrupt tables from time to time. When a 30 million row table becomes corrupt, it takes some time to recover from it.

I could go on but you get the picture. For a "small" database 5 Gig or less I think SQL server does fine. But when you get to the Data Warehouse size databases, I wouldn't recommend it unless you like working long days and weekends. Received on Mon Apr 01 1996 - 00:00:00 CEST

Original text of this message