Results. We find roughly 10(7) M-circle dot of oxygen in the CGM of simulated galaxies having M-star similar to 10(10) M-circle dot, in fair agreement with the lower limits imposed by observations. The M-oxy is found to correlate with M-star, at odds with current observational trends but in agreement with other numerical results. The estimated profiles of O VI column density reveal a substantial shortage of that ion, whereas Si III, which probes the cool phase, is overpredicted. Nevertheless, the radial dependences of both ions follow the respective observed profiles. The analysis of the relative contributions of both ions from the hot, warm and cool phases suggests that the warm gas (10(5) K < T < 10(6) K) should be more abundant in order to bridge the mismatch with the observations, or alternatively, that more metals should be stored in this gas-phase. These discrepancies provide important information to improve the subgrid physics models. Our findings show clearly the importance of tracking more than one chemical element and the difficulty of simultaneously satisfying the observables that trace the circumgalactic gas at different physical conditions. Additionally, we find that the X-ray coronae around the simulated galaxies have luminosities and temperatures in decent agreement with the available observational estimates.