I agree with Colin. The tools might have bugs that don't necessarily update all dates. I know I've seen issues with product versions being off between the 3 or 4...

Schemas allow you to group objects and then assign permissions to users on that schema instead of each object. They replace the "owner" in 2000 with a true schema that...

I don't think they ever support 3. I think they cover the current and one prior SPs for the current and one prior versions.

Run a trace, be sure it's not being called twice.

Haven't seen that, but 2008 is still wet paint and I'd likely point to that as the issue. Course I run 2008 in a VM for this very reason. Don't...

Are you sure they're the same build no? That should work.

good advice above. Search sp_help_revlogin on this site and script out logins. They should come over with a master restore, but if they don't......

Script out everything. Every object, job, alert,...

Every row in the cursor ends up delaying the transaction. Each of those delays can add up, especially if you update lots of stuff and end up escalating from row->page->table...

Keep in mind that a cluster shares the disk, so it shouldn't be IO, however you should be sure that your connections (SCSI or SAN fabrics) are set the same....

Also, how is the data truncated? Are you sure there's enough space in the destination table? When you say available, does that mean it's copied correctly or that it's viewable...

I think there are some things that are overwritten with 2005, so if you do 2005 first, things will break.

As far as issues, one has to be a named instance....

Even small tables should have a clustered index. The optimizers expect to have a clustered index, and I think it's worth doing as a general practice.

A database consists of multiple files whose name and location are stored in the master database. They don't necessarily have anything to do with the database name, although they default...

Can rows be 0 or negative? If not, you could make it status > 1.

Or do two updates with > 1 and < 1 as the clauses, which could run...

The separate server/scale out type approaches make sense, but only if the optimizer can actually respect this. And then how can you determine which two (or more) of a connection...