Solving a quadratic problem using Lagrange multipliers

Hi all,

I would like to solve the following quadratic optimization problem:
b1.JPG
Solving this model with CONOPT works very well. However, I would also like to solve it using Lagrange multipliers.

So I formulate the Lagrange functions and set the derivatives to zero:
b2.JPG
b3.JPG
I had thought that I could solve these equations to obtain the same optimal solution as before. However, the problem is unfeasible, apparently because of the last derivative. Do you have an idea where I might be wrong?