bgc exact restart fails during icepack testing about icepack HOT 6 CLOSED

@mattdturner
This case is failing because the restart test is looking for restart files in a restart/ directory, which is empty; there are restart files in restart_baseline/. I don't understand why this case (ISPOL) fails but others do not. Other cases have restart files in both places. Can you explain what is going on here, or how it is supposed to work? Thx

from icepack.

I suspect that the problem with this case has to do with the start and end times of the ISPOL data. Since it's from an actual field campaign, the start date is not Jan 1 and it doesn't run for a full year.

   year_init       = 2004
   istep0          = 4008
   npt             = 4776

The first restart file written is

   ./restart/iced.2004-07-01-00000

and the original test is supposed to only run for 4776 hours (199 days). The restart test is looking for

 ./restart/iced.2015-03-01-00000

so it looks like the calendar calculation in the restart testing script is not working correctly.

from icepack.

This case is failing because the restart test is looking for restart files in a restart/ directory, which is empty; there are restart files in restart_baseline/. I don't understand why this case (ISPOL) fails but others do not. Other cases have restart files in both places. Can you explain what is going on here, or how it is supposed to work? Thx

The way the restart tests work is that the 'run-initial' simulation outputs to the restart/ directory. The test_restart.script file then renames the restart/ directory to restart_baseline/, copies a restart file to restart/, and runs the 'run-restart' case.

It appears that the restart file date-time group is currently hard-coded to be 2015-03-01. I'll start looking at a better way to determine a restart date.

from icepack.

@njeffery @chrisknewman
I found the bug preventing exact restarts for both ISPOL and NICE bgc cases: the sst data was read in but not used. I am running a test suite now to confirm that this fixes the problem and does not create others.

from icepack.

from icepack.

Fixed with PR #60

from icepack.