I recently had some "troubles" in my restart simulations. I set the NTIMES and DT to run a one-year simulation starting from 2006/1/1, and I force the model to stop at 2007/1/1 00:00:48.
Due to HPC walltime limit, the job was killed at some point, and I restarted it using NRREC = -1 and pointing initial to the restart file. I found that the restart run will pass my desired end time.
Fortunately I had saved the restart records at 2006/12/28, and then I used this file to do some tests. First, I comment one line out in the ocean.in and saw the below error as expected:
MULTIFILE_INFO_S2D - Error while checking input Lateral Boundary file:
../../files/forcings/HYCOM/bdy/year2006/merged_coawst_bdy_2006.nc
last data time record available is for day: 2006-12-31 21:00:00.00
but data is needed to finish run until day: 2007-01-01 00:00:48.00
The error made me confident that I did force the model to stop at that time. Then, I added this file back in the ocean.in, the model run again. But still, it passed 2007/1/1 00:00:48.
I have another restart run having the same problem. Until now, if i run a simulation as a coldstart (NRREC=0), the model stops at the desired end time, just not for the restart simulation.
Is it a possible bug? This is not a big issue, but a little annoying since I did not know why it happened.