Yes, avoid the primary key on all 3 columns, especially a clustered PK. Not everyone is aware that whatever columns you use in a clustered index are ALSO added...

To change the role of a standby server when the primary server fails or is taken offline, disable the database restoration job on that server, execute the role change stored...

He's using AWE by specifying the /PAE switch in the boot.ini file, isn't he?

David - what errors are you getting, exactly? Why do you think SQLServer should be using...

Interesting - we're having EXACTLY the same problem right now.

In our case, we believe that we've traced the problem to some failing disks on our SAN array (or the SAN...

Sounds like your problem is exactly what the error says. The record number 15 in your data file is too large for whatever your column definition is in your...

One final, final postscript Greg - when you upgrade to 2005, I strongly recommend you replace the partitioned views with partitioned tables. Microsoft specifically recommend this course of action,...

No, I don't understand that either, and I suspect that you're underestimating just how fast it would be with no indexes at all!

If you have an index (any index) on...

Depends on your usage of those variables. If you're planning to put large volumes of data in them (by which I mean several thousand rows or more), then I'd...

Performance will depend on what degree of precision is used for your float. Single-precision uses 4 bytes, double-precision requires 8 bytes. This is an overhead when it comes...

Greg,

If you write a query like this:

declare @var int

set @var=2

select * from vw_ul where mon = @var

then yes, the optimizer will ASSUME that each table will contribute equally to the...

Why do you say that 2Gb should be sufficient? With a database that size, and a cache hit ratio that low, I'd say that it clearly isn't sufficient....

Yes - your Cache Hit Ratio is way too low. This should be 95% or better, especially if you're only asking it to service 2 transactions per second. ...

Greg,

When I run this, I see what you're seeing - 33% for each table.

HOWEVER - I think this is simply because of the very low volumes of...

I'm afraid I'm sticking by my original reply. I've just run a simple test script that inserts and updates data in a table, once without intermeditate checkpoints and then...

Interesting. I must confess, I've never actually tried this. My earlier reply was a purely theoretical one, but after Colin's follow-up I think I'd better investigate further!