Wednesday, January 14, 2009

Deploying to Production

I was reading today this interesting post about what happened when someone realized he forgot to put production code on a production machine. As you could have guessed, a small scale catastrophe happened as the service went down.
You all know this situation: you're developing or installing something, testing it, giving its URL (or other method of access) to other developers, and all of a sudden - it's production.
Normally, this is the time the product would be installed on a production-worthy server, with DRP procedures and the rest of the things that are must-have for production. But from time to time such things are missed. I know this, because I was in this situation before, and because I'm in this situation right now. I'm maintaining production code on my workstation, and I cannot turn it off for a few days because of that. Moreover, in order to avoid the need of changing the scheduled task's password (when my password expires) which is involved with this code, I used this scheduling trick. At least I have everything under source control, so when a disaster happens, I'm only half-screwed.
So it might be a good thing that such things happen, as long as they happen to someone else, to it reminds us to put our production stuff on production.

1 comment:

  1. This comment has been removed by a blog administrator.