"Bus error" with hetergeneous forcing data hdf5 file on NERSC about ats HOT 7 CLOSED

Do you need 0.87? Can you reproduce this on a newer release? We've definitely updated HDF5 versions since then.

from ats.

Note you'll have to update your input file using the script here: https://github.com/amanzi/ats/blob/master/tools/input_converters/xml-0.87-0.88.py

There were no changes that should affect you going from 0.88 to 1.0 or master at this point (I'm fairly sure!)

from ats.

Actually your best bet is to grab the rock creek example for 1.0, change to your mesh, and drop in the forcing file (does that have the physics you need?) Let's see if we can reproduce this in the new code first.

I recognize that if this works, it won't help you if you need chemistry, but our chemistry merge is 99.9% ready to be merged and we can bug @ajkhattak to get it done if this fixes the problem in flow only.

from ats.

Hi Ethan,

Thanks for helping out. I tried it with 1.0 (with the modified version from your rock creek example), but I reproduce the same issue as seen in 0.87. Again, this issue is not reproduced on our local workstation.

I copied the reproducer at /project/projectdirs/m2398/zexuan/coppercreek/hydrology/transient-1.0, with the met focing dataset I used at /project/projectdirs/m2398/zexuan/coppercreek/metforcing/pr_et3_sm0r_10yrs_heter15_64_10days.h5

from ats.

Great, I should be able to take a look there. We may have a better way of doing this anyway, at least in 1.0 (definitely not in 0.87).

from ats.

Just confirming that @jd-moulton is working on this?

from ats.

This issue has been resolved with the dynamic link maser build on NERSC

from ats.