My testing using SIM_cannon_integ indicates that a dry run file does in fact work to drive Monte Carlo.

I set trick.mc_set_dry_run(1), and generated runs using trick.MonteVarRandom.GAUSSIAN.
Then a copied the resulting MONTE_RUN_TEST/monte_runs to "M_cannon_init_angle".
Then I read the monte values from "M_cannon_init_angle" trick.MonteVarFile. Seems to work fine.

Added the following to SIM_cannon_integ/RUN_test/input.py for testing:
########################

Begin MonteCarlo Setup

########################
trick.mc_set_enabled(1)
trick.mc_set_num_runs(20)
#trick.mc_set_dry_run(1)

monte_var_src = 'file'

if monte_var_src == 'generated' :
print "###############################"
print " Generating Monte Carlo Values "
print "###############################"
var0=trick.MonteVarRandom("dyn.cannon.init_angle", trick.MonteVarRandom.GAUSSIAN)
var0.set_seed(1)
var0.set_sigma(0.6667)
var0.set_mu(0.6)
var0.set_min(.3)
var0.set_max(.9)
trick_mc.mc.add_variable(var0)

elif monte_var_src == 'file' :
print "#####################################################"
print " Reading Monte Carlo values from M_cannon_init_angle."
print "#####################################################"
var0 = trick.MonteVarFile("dyn.cannon.init_angle","M_cannon_init_angle", 2)
trick_mc.mc.add_variable(var0)

else :
print "#######################"
print " No Monte Carlo Values."
print "#######################"

######################

End MonteCarlo Setup

######################

from trick.