If $F:C \to D$ is a Cartesian fibration of $\infty$categories, I would like to show that $F$ preserves pullbacks. This seems intuitively clear, but I haven't found it in HTT (but perhaps I missed it).
This is false already for ordinary fibrations of 1categories. It is a theorem that if $D$ has pullbacks and $F:C\to D$ is a fibration, then $C$ has and $F$ preserves pullbacks if and only if each fiber of $F$ has pullbacks and the reindexing functors preserve them. Thus, the Grothendieck construction of any functor $D^{\mathrm{op}}\to \mathrm{Cat}$ which does not factor through $\mathrm{Cat}_{\mathrm{pullbacks}}$ will give a counterexample. 

