We develop a numerical algorithm for the solution of the Sturm-Liouville differential equation governing the stationary radial oscillations of nonrotating compact stars. Our method is based on the Numerov's method that turns the Sturm-Liouville differential equation in an eigenvalue problem. In our development we provide a strategy to correctly deal with the star boundaries and the interfaces between layers with different mechanical properties. Assuming that the fluctuations obey the same equation of state of the background, we analyze various different stellar models and we precisely determine hundreds of eigenfrequencies and of eigenmodes. If the equation of state does not present an interface discontinuity, the fundamental radial eigenmode becomes unstable exactly at the critical central energy density corresponding to the largest gravitational mass. However, in the presence of an interface discontinuity, there exist stable configurations with a central density exceeding the critical one and with a smaller gravitational mass.

Strange stars are one of the possible compact stellar objects that can be formed after a supernova collapse. We consider a model of strange star having an inner core in the color-flavor locked phase surmounted by a crystalline color superconducting layer. These two phases constitute the {\it quarksphere}, which we assume to be the largest and heaviest part of the strange star. The next layer consists of standard nuclear matter forming a ionic crust, hovering on the top of the quarksphere and prevented from falling by a strong dipolar electric field. The dipolar electric field arises because quark matter is confined in the quarksphere by the strong interaction, but electrons can leak outside forming a few hundreds Fermi thick electron layer separating the ionic crust from the underlying quark matter. The ionic matter and the crystalline color superconducting matter constitute two electromagnetically coupled crust layers. We study the torsional oscillations of these two layers. Remarkably, we find that if a fraction larger than $10^{-4}$ of the energy of a Vela-like glitch is conveyed to a torsional oscillation, the ionic crust will likely break. The reason is that the very rigid and heavy crystalline color superconducting crust layer will absorb only a small fraction of the glitch energy, leading to a large amplitude torsional oscillation of the ionic crust.