I am wondering if, for instance, 2-torus can be considered symplectic manifold. On one hand, it seems easy to construct a `symplectic 2-form', which is given by \begin{equation} \omega=d\theta\wedge d\varphi \end{equation} where $(\theta,\varphi)$ parametrizes the circles $\mathbb T=S^1\times S^1$. However, because it is a compact manifold, we also know that it does not have a global symplectic potential i.e. $\omega=d\phi$ for some 1-form $\phi$, otherwise it will contradict Stokes' theorem.
I originally suspected it may not admit symplectic structure because 'visually', I could draw a line from the centre (origin) of the torus such that it is tangential to the torus. A vector $p$ along this line would be 'tangential' to the tangent space at the point where the line touches the torus and hence it leads to degenerate 2-form. But I think there may be a hole in this argument.