It was Diovan vs. Norvasc in June of 2004. It came up being essentially equal for its chief end point (cardiac arrest events I believe). I saw a post on the NVS board and remember it but did not find the article to post.
In any event Diovan went on to be the best selling drug in that space and NVS' best selling drug so the results of the study were not detrimental. At the time it seemed like more of a positive for Norvasc.
You're right, there have not been many things that NVS has botched.