Unable to produce history files from CAM3.1

[covey1@llnl_gov]

I'm starting to run CAM3.1 on mcr.llnl.gov (a Linux cluster with the MPI Intel Fortran compiler). Despite setting nhtfrq(1) = -24 in my namelist input and running for 5 simulated days to normal-looking completion -- including restart file production -- I don't get any netCDF history files produced. One possible clue: toward the end of the log file there appears the message, "HTAPES_WRAPUP: history tape 1 : no open file to close." This seems to be the only mention of history files in the log file, apart from "history restart" files.

Where am I going wrong here?

- Curt Covey (covey1@llnl.gov)

 

[covey1@llnl_gov]

Following suggestions by Phillip Cameron-Smith, Art Mirin and Phil Rasch, I solved this problem by resetting a couple of run-time parameters in namelist input:

MSS_IRT = 0, which directs the program to not try archiving output on the NCAR Mass Storage System

MFILT(1) = 1 (number of time samples per history file)

NSREST = 0, which specifies an initial-condition run (later did NSREST = 1, specifying continuation of the earlier run, but Phil advised that switching the parameters above in "mid-stream" messes things up)