CESM version 2.1.4. New User!.
I've carried out a short run of b1850.basic, and the restart files are now in a local directory. This is:
$SHORT_ARCH/b1850.basics/rest/0001-01-06-00000/
and the restart files have names like
b1850.basics.cam.r.0001-01-06-00000.nc
Following this I cloned to a new case b1850.a_00. I then used xmlchange to define the restart case and date, and the length of the new run, after which it is built and submitted.
cd $CASEROOT/$CASE
./xmlchange RUN_TYPE=branch
./xmlchange RUN_REFCASE=b1850.basics
./xmlchange RUN_REFDATE=0001-01-06-0000
./xmlchange STOP_OPTION=ndays,STOP_N=6
./case.build | tee 04_build.out
./case.submit
On the first run through the loop, I found that case.build was looking for the b1850.basics restart files in directory $DATA_IN/cesm2_init/n1850.basics, adjacent to "b.e20.B1850.f19_g17.release.cesm2_1_0.020". I therefor generated the directory and a link '0001-01-06-0000' to the SHORT_ARCH restart directory with the same name.
On the second trial, case.build completed but soon after the main cesm program started it aborted because it could not find:
b1850.basics.cam.r.0001-01-06-00000-00000.nc
For the third trial I renamed the files in the SHORT_ARCH directory with the extra '-00000' and also changed the names within the pointer files. This time the program started without aborting - until the land ice code complained about the restart time, probably because it had never been time-stepped in the bl1850.basics run.
Anyway, forgetting about land ice, my question is - why the contrast between the two names:
b1850.basics.cam.r.0001-01-06-00000.nc generated by the previous case and
b1850.basics.cam.r.0001-01-06-00000-00000.nc needed for the restart?
I can live with changing the names whenever I want to restart from a different restart time, but I suspect I am missing something.
Thanks for any help.
David.
I've carried out a short run of b1850.basic, and the restart files are now in a local directory. This is:
$SHORT_ARCH/b1850.basics/rest/0001-01-06-00000/
and the restart files have names like
b1850.basics.cam.r.0001-01-06-00000.nc
Following this I cloned to a new case b1850.a_00. I then used xmlchange to define the restart case and date, and the length of the new run, after which it is built and submitted.
cd $CASEROOT/$CASE
./xmlchange RUN_TYPE=branch
./xmlchange RUN_REFCASE=b1850.basics
./xmlchange RUN_REFDATE=0001-01-06-0000
./xmlchange STOP_OPTION=ndays,STOP_N=6
./case.build | tee 04_build.out
./case.submit
On the first run through the loop, I found that case.build was looking for the b1850.basics restart files in directory $DATA_IN/cesm2_init/n1850.basics, adjacent to "b.e20.B1850.f19_g17.release.cesm2_1_0.020". I therefor generated the directory and a link '0001-01-06-0000' to the SHORT_ARCH restart directory with the same name.
On the second trial, case.build completed but soon after the main cesm program started it aborted because it could not find:
b1850.basics.cam.r.0001-01-06-00000-00000.nc
For the third trial I renamed the files in the SHORT_ARCH directory with the extra '-00000' and also changed the names within the pointer files. This time the program started without aborting - until the land ice code complained about the restart time, probably because it had never been time-stepped in the bl1850.basics run.
Anyway, forgetting about land ice, my question is - why the contrast between the two names:
b1850.basics.cam.r.0001-01-06-00000.nc generated by the previous case and
b1850.basics.cam.r.0001-01-06-00000-00000.nc needed for the restart?
I can live with changing the names whenever I want to restart from a different restart time, but I suspect I am missing something.
Thanks for any help.
David.