This page lists a number of known bugs or issues in released versions of MESA. Where possible we will also list work arounds, but for some bugs the only option will be to update to a newer version of MESA. Note this list is NOT comprehensive, users should check this first if they have an issue but it may not be complete.
ZAMS Model Central Composition
create_pre_main_sequence_model = .false. and
load_saved_model = .false., we fall back
to loading a ZAMS model based on interpolating from a grid of pre-computed ZAMS models found in
data/star_data/zams_models. The default file included in that directory is meant to start from
a composition of X = 0.70 and Z = 0.02, but one of the models in the grid (1.26 Msun) has partially
proceeded through hydrogen burning already so that its central H abundance is X = 0.58. Interpolation
in this grid of models will impact the central H abundance for initial masses between 1.0 and 1.58 Msun.
This bug affects versions r15140 through r23.05.1, and will be fixed in the next release. For current MESA releases impacted by this bug, the following steps provide a workaround with a patched ZAMS file:
Download this updated ZAMS model file:
Copy the file into
Use the following setting in the
&controlssection of your inlists for models where you want to use the patched ZAMS file:
zams_filename = 'zams_z2m2_y28_patched.data'
There has been a bug present in the rate
r_c12_to_he4_he4_he4 in r22.05.1 and r22.11.1.
This causes an excessive amount of C12 to be burnt during core helium burning.
We strongly recommend that users update to the latest MESA.
There is a bug in the rate selection code that certain endothermic weak reactions are not added to the nuclear network. These are r_be10_wk-minus_b10, r_ni66_wk-minus_cu66, and r_h3_wk-minus_he3. Other weak reactions with heavier parents may also be affected.
A separate issue also meant we are missing the rate r_he4_ap_li7 as the reverse rate of r_li7_pa_he4.
Both issues will effect previous versions of MESA as well.
Both issues have been fixed in the git main branch.
A bug has existed since shortly after r15140 where RTI mixing will be effectively zero in a model even with the
This has now been fixed in the git main.
Convective premixing (CPM) has not worked properly since release r15140. CPM was broken by the
removal of the
lnPgas_flag, which caused some of the necessary EOS updates to be missed after
CPM updates the abundances in mixed cells. CPM does not need
lnPgas_flag, but it does require
EOS updates at constant pressure. This will be fixed in future releases.
Invalid location for overshoot boundary
Sometimes MESA will crash with an error similar to this:
Invalid location for overshoot boundary: cz_bdy_dq, dq= -0.13040604669743103 1.4532774141478022E-003
0 terminate reason: nonzero_ierr
This bug effects many previous versions of MESA as well. This has been fixed in gh-400 . The solution is to update to a newer MESA version.
After running a model with the al26 isomers in your net, when you run the model again, it may immediately crashes and prints a backtrace containing:
create initial model
create rate data for r_al26-1_to_al26-2
create rate data for r_al26-2_to_al26-1
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
#0 0x7fd69c02931f in ???
#1 0x99a523 in __interp_1d_misc_MOD_do_interp_values
The solution for now is to remove all files in
each MESA run, you may also find that changing the number of OMP threads also fixes the problem.
Atmosphere in pulse data
add_atmosphere_to_pulse_data does not work properly with an Eddington atmosphere (the default), and also crashes if
atm_T_tau_opacity = 'varying' is set.
Colors: bad filter name
If you get an error:
bad filter name:
First check that the name matches in your history_coloumns.list file and your color file. Next check for non-printing characters history_coloumns.list in the filter name. This can bee checked with:
cat -A history_columns.list | grep "abs_mag"
Finally, there is a bug if you name any column with
/ in it (for instance
[Fe/H]). The solution is to rename the column to remove the forward slash.
An experimental RSP solver feature was turned on by default, leading to convergence issues in nonlinear model integration. Users should include RSP_do_check_omega = .true. in the &controls section of their inlists to get rid of this issue.
Free Electron Density on FreeEOS
The free electron density (
lnfree_e) reported by FreeEOS was off by a factor of
ln(10) due to tabulations needing to list the log base 10 value of this quantity rather than natural log. For historical reasons related to OPAL tables, the EOS tables report the log base 10 value, which is later converted to natural log before being reported as
lnfree_e in MESA.