Boundary condition on $mathbf{B}$ to describe resonant cavities from waveguides

In Jackson- Classical Electrodynamics when resonant cavities are discussed (8.6, page 252) (but also at page 7 here or at page 19 here) the explanation is made by saying that the solution is the same of a rectangular waveguide, which we can suppose that is for $E$

$$mathbf{E}(x,y,z,t)=mathbf{E_0}(x,y) e^{i(alpha z-omega t )}$$
$$mathbf{B}(x,y,z,t)=mathbf{B_0}(x,y) e^{i(alpha z-omega t )}$$

but instead of $e^{i(alpha z )}$ there is a factor $H cos(wz)+J sin(wz)$, so
$$mathbf{E}(x,y,z,t)=mathbf{E_0}(x,y) [H cos(wz)+J sin(wz)]e^{-i omega t }$$
$$mathbf{B}(x,y,z,t)=mathbf{B_0}(x,y) [H’ cos(wz)+J’ sin(wz)]e^{-i omega t }$$
And the boundary conditions imposed are teh following (supposing the lenght of cavity to be $d$ in $z$ direction)
$$B_z(z=0)=B(z=d)=0 ,,,,,,,,,,, forall x,y tag{1}$$
$$E_x(z=0)=E_x(z=d)=E_y(z=0)=E_y(z=d)=0 ,,,,,,,,,,, forall x,y tag{2}$$

While $(2)$ is clear because it simply requires the electric field to be normal to the surface (which is the necessary condition to avoid dissipation), I do not see why $(1)$ is required.

It would be explained by the fact that $E$ is perpendicular to $B$ but that’s a conclusion that one should get from the solution, not an assumption to impose the boundary conditions. Unfortunately I did not find an explanation for $(1)$ neither on Jackson neither on the references so why is $(1)$ a necessary boundary conditions to impose?