If you use sp_placeobject, checkstorage generates a 100028 soft fault for the object. This may also be followed by a 100025 fault, but this does not mean that the OAM row count is wrong. It merely indicates that checkstorage could not collect an accurate row count because of the use of sp_placeobject.
When 100028 and 100025 faults occur in pairs, therefore, you do not need to be concerned about the 100025 fault. The 100028 (and corresponding 100025) fault can be removed by using bcp to unload and reload the table.
