# Minimisation of grand free potential at equilibrium?

I have a couple of questions regarding the grand potential at equilibrium.

I am well aware that in the grand canonical ensemble, particles and energy can be exchanged. I also understand why the grand potential can be expressed as

$$\text{d}\Phi_G=-S\text{d}T-p\text{d}V-N\text{d}\mu.$$

First question: I read that when connecting a system to a large reservoir, the differentials can be written as

$$\text{d}(\Phi_{G1}+\Phi_{G2})=-(S_1-S_2)\text{d}T-(p_1-p_2)\text{d}V-(N_1-N_2)\text{d}\mu.$$

Where does this expression come from, and why do we subtract, not add, the conjugate variables? The source justifies it as $$\text{d}V_1=-\text{d}V_2$$ and so on but I don't see why.

Second question: Knowing what can be exchanged in the GCE, I understand that $$T_1=T_2$$ and $$\mu_1=\mu_2$$, so $$\text{d}T=0$$ and $$\text{d}\mu=0$$, and that for minimisation, $$\text{d}\Phi_G=0$$. This means either $$p_1=p_2$$ or $$\text{d}V$$. Which is it? This feels rather silly to ask, but feel like either one is a possible consequence of equilibrium.