#4 xios_config/*xml not copied in the run directory leading to early crashes of the model

head -n 3 ~/modeles/nemo/3.6/CONFIG/ORCA1_LIM3_REF/EXP00/EXP_REF.o1495821

[PROLOGUE] Starting prologue on node node0293 at date Fri Oct 27 10:21:58 CEST 2017
cp: cannot stat `xios_config/*xml': No such file or directory

mkdir -p "${run_dir:?}"$
cd "${start_dir}"
cp -u xios_config/*xml "${run_dir}"
cp -u namelists/* "${run_dir}"
cd "${run_dir}"
cp -u "${nem_exe_file:?}" .
cp -u "${xio_exe_file:?}" .

On Zenobe, when executing `coral submit`, the job passes the waiting list, but fails early afterwards, during the running process. No output data are produced. The logging file shows: ``` head -n 3 ~/modeles/nemo/3.6/CONFIG/ORCA1_LIM3_REF/EXP00/EXP_REF.o1495821 ``` ``` [PROLOGUE] Starting prologue on node node0293 at date Fri Oct 27 10:21:58 CEST 2017 cp: cannot stat `xios_config/*xml': No such file or directory ``` The run directory doesn't contain xml files, indicating that `xios` config is not copied by coral. The copy command is contained in [Coral/templates/scripts/skeleton.sh, line 53](http://climate.be:3000/TECLIM/Coral/src/master/templates/scripts/skeleton.sh#L53). The `cp` is run in the wrong directory. Adding a `cd` to the start directory after its creation seems to solve the problem: ``` mkdir -p "${run_dir:?}"$ cd "${start_dir}" cp -u xios_config/*xml "${run_dir}" cp -u namelists/* "${run_dir}" cd "${run_dir}" cp -u "${nem_exe_file:?}" . cp -u "${xio_exe_file:?}" . ```

Bug not seen in the current version of coral. Closing the ticket.