Should I Bother Running a Database Test?

matttrax

Should I even bother running the suggested database test in the upgrade manual on a 75GB database. I get the feeling that even if I ran it on our Quad proc - quad core box it would sit there churning for a week

Thoughts are appreciated. Thanks.

DaveT

Hi Matt,

This is a tough one without knowing the background.

Basically it come down to how confident you are of the upgrade. The test database is throw out if mistake have been made (not that you would ) in the upgrade. It could save you a whole lot of time after go-live if an issue is discovered and also the pressure will be on to resolve it quickly then.

My $0.02 is run leave it run on a spare machine if only its to see how big the output file ends up as. A quick review might just save you bacon on the go-live.

Good luck with the upgrade

girish.joshi

In general, I think its worth it. Upgrades are one of the few times in an implementation that you get to address data consistency issues - and there are always a lot of those.

Dave T makes a good suggestion. I would take it a little further, though, and consider how you intend on evaluating the file that test database puts out. There will be a massive amount of information, a large portion of it redundant. It's doable, I just suggest you think about it in advance.

Personally, I've been meaning to solve both problems (the time it takes for test db to run and the evaluation of its results) by slapping a front on it, and having it output to the database directly. That way, you could run the test db in parts, with internal commits.... one could get to work on correcting some of the errors while it continued to run... anyway... just another NAV dream for now...