Soliton equation and non-killing potential vector field

I am searching for a non-Killing vector field $\zeta \in\frak X\rm (M)$ where $(M,g)$ is a Riemannian manifold such that
$$\frac12 \frak L_\zeta \rm g+Ric=\lambda g$$ $$\frak L_\zeta \rm Ric=\lambda \frak L_\zeta \rm g$$ The first equation is called soliton equation and $\zeta$ ia called a potential vector field. It is clear that if a Killing vector field is a solution of the first equation, then $(M,g)$ is Einstein with factor $\lambda$ and the second equation holds. I want to consider a non-Killing vector field. I checked 8 examples with non-constant ricci curvature. In each case I got a system of difficult PDEs so I considered many special cases. All solutions are Killing vector fields.
My question
Is there an example with a non-Killing vector field satisfying these equations?
Or any hind to a non-existence result.
I don't know about general dimensions, but, for surfaces (i.e., when $M$ has dimension $2$), it is not hard to show, using the Cartan structure equations, that there is no solution to this system for which $\zeta$ is not a Killing field. It is an overdetermined system, and the compatibility conditions rapidly eliminate any possibility of there being a non-trivial solution in this case.
• @Holonomia I think it should be as follows. The first part is: If $\zeta$ is Killing, then $M$ is Einstein with factor $\lambda$. Its equivalent is: If $M$ is not Einstein with factor $\lambda$, then $\zeta$ is not Killing. – Semsem Jan 9 '16 at 19:29