Project planning vs ‘Just do it’

So, we have all had a project dumped on our laps.
Guy at the top thinks its a 5 minute job etc etc.

I currently have a project where I have two choices :

1. Plan to the nth degree, engage practically the whole of IT in order to get the project completed, and hope that we have covered every possible angle. However this will take 6 months and potentialy delay a critical storage refresh.

2. Take what we have now and deploy; let users know that there may be a small impact to them over the first few days and just take the hit at the support end of things. All business-critical applications will be reviewed and fixed prior to deployement.

I see the pros/cons of both methods; I’d like to canvass the opinions of the wider community.