Hi, I am using CESM 2.1.5, compset FW2010climo with grid f09_f09_mg17. I am trying to continue a run from a successfully completed run, but the code just does a startup run ignoring that it has already been run, and does not continue from the last completed time stamp.
I have set xmlchange CONTINUE_RUN=TRUE, but all that did was just rerun the simulation from the initial start date, not continue the run from the last successful time stamp.
I confirmed that my rpointer files contain the correct restart files.
I noticed with this compset that the ref start date is 0003-01-01 yet the initial run date is 0001-01-01, so not sure if that is causing an issue either. After seeing the two different initial dates, I tried to set the initial run date to 0003-01-01, so that the initial dates would be common. But then when I tried to restart, I got an error about overrunning the time stamp.
I am not using case.submit due to how CESM is ported to our system. I am instead using the mpiexec command line from the preview_run into a manual PBS job submission. There are no errors from using this means of submitting a job that I can tell, as I can run a case initially.
Do I need to do any additional steps other than change CONTINUE_RUN=TRUE? Am I missing steps that case.submit interprets for a continued run by manually running a job on our system cluster?
Thanks
I have set xmlchange CONTINUE_RUN=TRUE, but all that did was just rerun the simulation from the initial start date, not continue the run from the last successful time stamp.
I confirmed that my rpointer files contain the correct restart files.
I noticed with this compset that the ref start date is 0003-01-01 yet the initial run date is 0001-01-01, so not sure if that is causing an issue either. After seeing the two different initial dates, I tried to set the initial run date to 0003-01-01, so that the initial dates would be common. But then when I tried to restart, I got an error about overrunning the time stamp.
I am not using case.submit due to how CESM is ported to our system. I am instead using the mpiexec command line from the preview_run into a manual PBS job submission. There are no errors from using this means of submitting a job that I can tell, as I can run a case initially.
Do I need to do any additional steps other than change CONTINUE_RUN=TRUE? Am I missing steps that case.submit interprets for a continued run by manually running a job on our system cluster?
Thanks