We stop & start ours daily (not because we have so little data, but because we have so much). If you just set the file name using the date, you...

Those look right.

You may not get a shrinked file the first time you run that if the transactions are still operating out at the end of the file. The lazy...

Just wait for the trace to roll over and you can get the data.

We created a new file name each time we started it, using the date of the start...

By the way, I realized that it looks like I'm recommending sp_OA procs. I'm not. They're slow and memory pigs, but an option. Just wanted to clarify that.

Lumigent Entegra.

It's related to Log Explorer, using some of the same technology. It also tracks object changes in the system so you can see if unknown users are changing tables...

I seriously doubt this is your problem, but we ran into the same error message due to two problems on our system, so these might help you. First was the...

I'm not sure how to help you on the automatic kill of any long running procs (never forget that, depending on what a procedure is doing, it can take as...

I believe I got this script somewhere on this site, but it cycles through the indexes and checks fragmentation. Depending on the fragmentation it does nothing, runs a rebuild or...

TSQL all by itself can't write out to a file. From TSQL, to get to a file you'd need to call sp_OA procedures. Otherwise, I'd look at DTS as a...

It is fast. It's not necessarily safe though. Once you've detached the files, in theory, just about anything could happen to them. What's wrong with backup & restore. I've never...

I've never actually found a situation that both required an indexed view yet was simple enough in it's design to be able to implement one, so I understand what you're...

Honestly sounds like you might benefit from an indexed view. Did you look into that instead of loading a table? Ony reason I ask is because the indexed view is...

Swing over to SourceForge.net and search for SqlUnit. It uses the nUnit test driven development approach for TSQL. It would allow for a set up step, an execution step, verification...

Sure. It's not hard. You create the database without the NDF files or the file group, just the basic MDF & LDF file. Then you add the filegroup and files...

All true.

However, the problems we ran into weren't related to I/O or memory, but rather when using the table variables (@ tables) in joins with each other and regular tables,...