The storage of long bunches for long time intervals needs flattened stationary buckets with a large bucket height. The longitudinal motion of the initially mismatched beam has been studied for both the single and dual harmonic RF systems. The RF amplitude is determined to be r.m.s wise matched. The bucket height of the single harmonic system is too small even for shorter bunch with only 20% increased energy spread. The Halo formation and even debunching can be seen after a few synchrotron periods for single particles with large amplitude. In the case of small energy spread for a cooled beam, Coulomb interaction cannot be ignored. The external voltage has to be increased to keep the r.m.s bunch length unchanged. The new voltage ratio R(N) simplifies physics for the emittance-dominated bunches with modest particle number N. For the single harmonic system, substantial amount of debunching occurs without increasing the external voltage, but very little if the RF amplitude is doubled. Results from the ORBIT tracking code are presented for the 1 GeV bunch in the HESR synchrotron, part of the GSI FAIR project.