Yeah slap head moment on my part as of course if the replication does have an issue, the agent will just keep retrying until it sends those transactions! Duh!

I'm sure...

The database mainly. The log is bigger than I'd expect to (5 gig) but the database is steadily rising and is now at 15 gig. I will see how this...

Yeah you'll need to get all those connections to close. That includes the window you are running the sp_who2 in.

Just an Update FYI -

The distributer on Server A was set to keep transactions no more than 35000 (or therebouts) hours! Not sure why as ALL 3 servers were set...

Question - In SQL2K you would set the replicationdboption "Sync with backup" to True in order to get the commands to wait in the log until a backup of them...

Is it just me or was some of the advice in this topic very very risky?

Will ALWAYS ruin the data???

COMMIT had been commented out! :crazy:

Thanks

Nightly wouldn't be out of the question. I didn't mean I sit and watch each one by the way.

Initially we are catching up on 2 years of archiving. That's one...

It's a once a week thing. I guess if I stick a commit on it then there should be no harm done actioning each task in one hit.

Well I can tell you that the PNo is passed to an SP, and that this SP then analyses that PNo, and depending on the type etc it will copy...

It's SQL 2K but 65 mode so no Table Variables for now 🙁

I can't get the alternatives to run faster than the Cursor so thats annoying, but I'll keep plugging on. Just tried a ROWCOUNT method but that took double the time.

I'm such a fool. I've used a Counter variable before. Nice one.

How does that ROWNUM get populated? EDIT: Identity of course! Ignore me.

Great help guys,. Thanks.

The same idea using cursors took 4 minutes compared to 16 for the code above, so clearly I'm not doing it properly!