I am running a unsteady compressible problem using coupled implicit solver mode. The problem is a thin film with the lower wall stationary and the upper wall oscillating at high frequency. Each side exits are set as Pressure outlet at 0 gauge. The problem is, it keeps reporting reverse flow on both of those outlet, even if I change them to Pressure inlets. Is is an error message?

I would need more information to help. For example, what is your initial guess? It's possible you get reverse flow during the solution depending on initial guess. Also, why set both to 0 gauge or 1 atm? Do you wish to have migrating flow across each boundary? If you have a steady periodic problem don't allow flow passage. That would surely solve the problem.

How can I disallow flow passage? The problem is initally time dependent, but should result to steady state after some cycles. I am not sure if there would be any flow through the exits, but because they are open to ambient I thought they chould be set to ambient conditions.

If I understand the problem correctly you will have oscilating flows at the boundary. In other words you will have a flow which pumps (of course other factors are involved such as viscosity, etc.), but more than likely you do have pumping occuring. If this is true I would set up the problem in this fashion. On both sides model ambient air as a big chamber. i.e. create a model which includes an area far away from the exits. Then make those periodic boundaries (or walls for example) whereby no flow passes. You should still see the pumping action which appears to be of interest to you. Others may approach the problem differently and have varying opinions, but that's what I would do. As to how within Fluent I am unsure specifically, but sure the program allows this type of BC. All CFD's do. I suggest pull the manual and look under boundaries. Should tell you in there.