search

uw-cmg / MAST

MAterials Simulation Toolkit for use with pymatgen
17 stars 8 forks source link

mast must handle systems where there is a queue limit #447

Closed cmgtam closed 7 years ago

cmgtam commented 8 years ago

Currently, on systems where there is a queue limit, MAST tries to submit all jobs on MAST/CONTROL/submitlist and will therefore error out, not clearing submitlist (which causes problems of potential resubmission for the jobs MAST did manage to submit). MAST should be able to submit those jobs which it can, clear the jobs from the submitlist, acknowledge the error message and possibly log it, and be set for the next MAST run without crashing.

cmgtam commented 7 years ago

tam branch commits are defunct. Use:

a9f9ee312d91408048508f4c23dbcb61ef3e87af f734e191941fbced71cde62eafda1373d0fe85bc

cmgtam commented 7 years ago

Have to make a provision for the mast monitor, which may in some cases be running on the same queue...either that or doc note that it should run on a different queue (which usually it should).

cmgtam commented 7 years ago

Addressed. Should work on Stampede as long as two separate queues are used, which is indicated in the stampede_knl submit profile folder's mastmon_submit.sh and submit_template.sh files.