Closed vanroekel closed 5 years ago
@vanroekel, thanks for pointing this out. I can reproduce the issue. If you access the anaconda environment the way that they prefer (see below) this doesn't happen so I guess the issue is that loading the environment as a module doesn't set required environment variables. The following worked for me:
source /global/project/projectdirs/acme/software/anaconda_envs/${NERSC_HOST}/base/etc/profile.d/conda.sh
conda activate e3sm_unified_1.2.3_py2.7_nox
I'll look into whether there's a feasible way to set the required environment variables as part of the module, but I suspect this may be a problem with more packages than just basemap
.
@vanroekel, I'm trying to figure out how to simplify things a bit. Could you try the following (in a fresh shell on Edison) when you get a chance?
source /global/project/projectdirs/acme/software/anaconda_envs/load_latest_e3sm_unified.sh
run_mpas_analysis --list
This should tell us whether the little script I made can load the environment correctly. Hopefully sourcing that script rather than loading a module won't end up being too onerous. I don't know what else to do because modules seem simply to be an infeasible option at this point.
I tried, but it looks like I don't have permissions to run the script.
Sorry about that. Please try again when you have a chance.
It works for me. Thanks for the quick fix! I will say I have no problem with the source activate method. This seems like a very reasonable way to load python, at least to me.
Yep, this just does that under the hood so users who don't want to think about it can just source a script and that script will know what the latest version is.
I'll make these helper scripts on all the machines tomorrow and then update the instructions and send out a follow-up email explaining moving away from modules (again) and apologizing for the moving target.
Addressed by requesting that users activate the environment directly with the load_latest_e3sm_unified*
script.
I tried using e3sm-unified for MPAS-Analysis on NERSC this morning and am getting an error, see below
e3sm-unified/1.2.2 works fine. If this issue is more appropriate for the MPAS-analysis repo, let me know.