The low temperature Monte Carlo dynamics of an ensemble of linear harmonic oscillators shows some entropic barriers related to the difficulty of finding the directions in configurational space which decrease the energy. This mechanism is enough to observe some typical non-equilibrium features of glassy systems like activated-type behavior and aging in the correlation function and in the response function. Due to the absence of interactions the model only displays a one-step relaxation process.Slow relaxation processes are widespread in condensed matter physics. These include magnetic relaxation in spin glasses, transport processes in structural glasses, pinning effects in superconductors among others. A large class of these systems show what is commonly referred as aging, i.e. dependence of the response of the system on the time in which it is perturbed. Aging effects [1] are a signature that the system is far from thermal equilibrium and consequently the fluctuation-dissipation theorem is not valid [2]. It has been realized quite recently that aging is indeed a solution of the off-equilibrium dynamics in some exactly solvable models [3,4]. Aging appears if relaxation to the equilibrium is slow due to the presence of energy barriers in a rugged free energy landscape as well as in systems with entropy barriers with a quite simple landscape [5,6]. In this last case, as the system relaxes towards the equilibrium, the number of directions in phase space where the system can move decreases progressively. This means that the system needs more time to decorrelate or to forget the previous configuration. This effect is usually encoded in the two time correlation function where the C(t, t ′ ) depends on both time indices [7].From previous considerations it is clear that aging can also be present in extremely simple relaxing systems without any interaction, the only condition being the progressive reduction of available phase space where the energy decreases. This was an essential ingredient in the Backgammon model recently proposed to explain glassy behavior in the absence of energy barriers [5]. Here we consider a simpler example and analyze the Brownian oscillator. The Brownian oscillator is usually studied in the Langevin approach. It is described in any textbook of stochastic theory [8]. It is possible to show that in this case there are no slow processes involved. In fact, the relaxation turns out to be exponential as expected for the dynamics of a particle in a single parabolic potential well. Here we consider the Monte Carlo approach and choose a dynamics based on the Metropolis algorithm [9]. This Monte Carlo approach was already studied in a disordered model with long-range interactions which turns out to be non trivial, at least in the zero temperature limit [10]. The simplest case of an harmonic oscillator is solvable and we analyze the dynamics here.In [10] we checked that, after a suitable rescaling of time, the equilibrium Langevin and Monte Carlo dynamics are equivalent. Also, we showed how the Langevin dy...