On a properly designed split-rail PSU, the OCP trip-point is set such that either rail is actually capable of providing more than 1/(number of rails) x (total +12V amperage). This is also why the rails are not strictly additive. For example, on a PSU with a maximum of 65A on +12V, the rails might allow for a max of 22A on each, or perhaps 25A, 25A, and 28A, but the total is still limited to 65A (not 78A). This way, power is not "trapped" on under-utilized rails, but you don't end up with the potential safety issue of a near-short creating a heavy load that is yet within the limit of a single rail. This could keep the PSU energized, even though an open diode somewhere is glowing cherry red and smoking.