Insert/delete makes sense rather than trying to track column level changes. Writing your own replication is a lot of work! Overall the plan seems ok. Is dumping the front end...

Not sure if you have a problem. Maybe its just keeping data cached? You could try dumping the data cache to see. Or set a max mem usage for MSDE?

Andy

Great solution!

Andy

I think you "have" to build the string outside don't you? Shows how often I used exec()!

Andy

Honestly I dont think you want to stop logging. Turn off trace flags to reduce the volume maybe, but its a valuable resource - I'd encourage you to leave it...

Not sure I would tackle it via triggers. Why not just import the data first, then apply a DTS package (or code as you prefer) to do the transform all...

Interesting question. Not directly that I know of. You could use xp_cmdshell to run ipconfig /all and then parse the output.

Andy

Is the DTC running? You're using SQL login and not NT authentication? I'd double check the data in MSsubscription_properties to make sure everything look good.

Andy

Maybe they already had a uniqueidentifier? Or you were handling inserts differently?

Andy

I run a job daily that calls sp_cycle_errorlog - that way I have one log per day to keep the size manageable. More of an issue for easy browsing than...

For the attach, try wrapping the file name in brackets, like this: "[c:\test data.mdf]". Not sure about MSDE - if nothing else you should be able to alter it via...

I know its the not the easiest thing, but maybe makes sense to just back up a step and try setting up the one problem server again. If you want...

Man thats big. I've seen a few that were 800-900 lines with lots of calls to other procs - would take minutes for the query plan to draw in QA....

I typically use 1204.

Andy

Definitely you can find tools like WinDiff to compare the files - the fun is in generating a script that will correct the differences!

Andy