If $\Omega=\{a,b,c\}$ give an example that $E(E(X|\mathcal F_1)|\mathcal F_2)\neq E(E(X|\mathcal F_2)|\mathcal F_1)$
If I choose the $\sigma$-algebras disjoint (or better said: one is not fully contained in the other) is the inequality always satisfied ?
Let $$\mathcal F_1=\{\emptyset,\{a\},\{b,c\},\{a,b,c\}\}$$ and $$\mathcal F_2=\{\emptyset,\{b\},\{a,c\},\{a,b,c\}\}$$ and let $$P(\{a\})=P(\{b\})=P(\{c\})=\frac13.$$ Finally, let
$$X(\omega)=\begin{cases}1,&\text{ if } \,\omega\in\{a,b\}\\0,&\text{ if }\,\omega=c.\end{cases}$$
So
$$E[E[X\mid\mathcal F_1]\mid\mathcal F_2]\not=E[E[X\mid\mathcal F_2]\mid\mathcal F_1].$$