SQL Server Performance Forum – Threads Archive
A 2000 database on 2005
Hi, I need to test out some new indexes for a SQL Server 2000 production database. It couldn’t be done in prod, so it was restored onto my new development SQL Server 2005 server. Should I request a SQL Server 2000 server (took a whole month to get this one) to try the changes or is there NOT much change in 2005 in the way the database engine (specfically the query optimizer) executes queries. The only difference I could see is bookmark lookups have been ditched in 2005. I need some good answers to go the server commissioning team with! Help much appreciated jfor better testing, you should try this with SQL Server 2000. as there are performance enhancement in SQL Server 2005, you will not get proper results —————————————-
Contributing Editor, Writer & Forums Moderator
http://www.SQL-Server-Performance.Com Visit my Blog at
http://dineshasanka.spaces.live.com/
I would suggest to take a route of UPGRADE ADVISOR from SQL 2005 in this case that will give more information for your database. Satya SKJ
Microsoft SQL Server MVP
Contributing Editor & Forums Moderator
http://www.SQL-Server-Performance.Com
This posting is provided AS IS with no rights for the sake of knowledge sharing.
There are a lot of new features in 2005 wrt to "Database Engine Scalability and Performance Enhancements " (looking 2005 BOL), but I couldn’t see any behind-the-scenes changes that would improve adhoc queries. Downloading is locked down here, so no Upgrade Advisor.
I think u shoule test it with SQL 2000.
In my opinion ,testing enviroment shoule be as much the same as possible to the prod .
Life is simple~~~
Why you cannot download a recommended product that can give more help in this regard, as you’ve seen in the BOL they are many enhancements as compared to SQL 2000 in this case. As suggested without testing any of them you wouldn’t get to the point whether to say it is useful or not. Satya SKJ
Microsoft SQL Server MVP
Contributing Editor & Forums Moderator
http://www.SQL-Server-Performance.Com
This posting is provided AS IS with no rights for the sake of knowledge sharing.
I agree with Satya. You can get a trial of 2000 and put it on a box. Even if it is not the primary instance, it would be better to benchmark against. There is a difference between the two performance wise. Michael B
Sr. DBA "The fear of the Lord is the beginning of knowledge,
but fools despise wisdom and instruction." Proverbs 1:7
]]>