Well, I found the reason here:
One major caveat is that in version 0.12, performing a decay-only timestep still results in a transport calculation (even though we know the reaction rates will be all zero). This has been fixed in the developmental version of OpenMC so that no transport solve takes place when the power is zero and will be part of the next release.
Is there a way to force transport calcluations during decay to study criticality of spent fuel along decay? I am doing it by setting a very low power for the moment.