8 Supporting Scripts:

A number of supporting scripts are included in the CCSM2.0 release. Some of these scripts are found in the scripts/test.a1 directory and others are located in the scripts/tools directory. Scripts in the test.a1 directory are experiment-specific and may need to be changed on an experiment by experiment basis. Scripts in the tools serve as a suite of utilities that are intended to be experiment- and machine-independent.

8.1 Experiment dependent scripts

The $SCRIPTS/ccsm_archive and $SCRIPTS/test.a1.har scripts are supporting scripts to the main CCSM run script . Both of these scripts are run from the main script, $SCRIPTS/test.a1.run. By default, calls to these scripts are currently commented out. These scripts are used mainly for production runs. In addition, there is normally a script call ccsm_joe that is created by the main script, test.a1.run, to store job operating environment (joe) variables and make them available to other supporting scripts.

• $SCRIPTS/ccsm_archive is executed in the main script, test.a1.run, after the CCSM has finished running. The purpose of this script is to move data from the executable directories to a separate directory in order to stage the data for harvesting. In particular, the archiving process moves data from $EXEROOT to $ARCROOT. ccsm_archive gets many of the environment variables from ccsm_joe. Then, on a model-by-model and file-by-file basis, ccsm_archive copies or moves data from the executable directories to the archive directories. Users may need to change the filenames to be archived for the different models depending on the particular CASE setups.

• $SCRIPTS/test.a1.har is the ccsm harvester script and it is submitted from the main script as a single-processor, serial, job. It can also be run interactively. test.a1.har checks to see whether a file has been harvested by first acquiring the file from the local mass store and then differencing it with the local copy in the archive directory. test.a1.har then either removes the local file or copies the local copy of the file to the mass store. It also has the ability to copy files to another location via the Unix copy (cp or scp) commands.

• $SCRIPTS/ccsm_joe. which contains many experiment dependent variables for use by other scripts. is generated by the main script, $SCRIPTS/test.a1.run, every time it's executed. It also serves as a point of reference for checking important environment variables. The ccsm_joe script is generated by executing scripts/tools/ccsm_checkenvs.

8.2 Scripts in the tools directory

Under the scripts directory is a tools directory. There are several generic utility scripts in this directory. The purpose of these scripts is to provide a generic interface to mass storage systems at different sites, to encapsulate a suite of commands into one utility, to provide a hierarchical search capability for files, and to provide machine-specific commands or notes necessary for running the code at other sites. Scripts in the tools directory are meant to be completely case independent. Some of the scripts require case-specific information, which is provided by the ccsm_joe script. As mentioned above, these utility scripts are meant to be generic. However, they have been tested on a limited subset of the current available hardware and therefore some limited modifications may be necessary. In particular, the interface to the local mass storage system may need to be altered in these scripts in order for them to work properly for new hardware.

A brief description of the scripts in the tools directory is provided below. The scripts are also self-commenting to help the user understand and modify them.

• ccsm_checkenvs Echos environment variables associated with a CCSM run. This script is used to create the ccsm_joe file in the main run script. This script has no arguments.

• ccsm_cpdata [machine1:][dir1/]file1 [machine2:][dir2/][file2] Generic utility to copy a file. Can identify whether cp or scp is required based on an optional machine argument. Takes one or two arguments.

• ccsm_getfile [dir1/]file1 [dir2/][file2] Hierarchical search for a file. Search for a file in several sources including locally, on the mass store, and on other machines. Copy that file into dir2/file2. Takes one or two arguments.

• ccsm_getinput [dir1/]file1 [dir2/][file2] Hierarchical search for a ccsm input file. Assumes the file will exist under the specific directory "inputdata". dir1 represents the directory path under inputdata to search for the file. Searches locally, in the local inputdata file system, on the local mass store, then at other sites for the file. Takes one or two arguments.

• ccsm_getrestart Copies restart files from a specific restart archive directory into the working directories of the components. This will eventually be configured to look for restart files on the local mass store as well. No arguments.

• ccsm_msmkdir mssdir Creates a directory on the local mass store. Takes one argument.

• ccsm_msread [mssdir/]file1 [locdir/][file2] Copies a file from the local mass store to the local disk. Takes one or two arguments.

• ccsm_mswrite [locdir/]file1 mssdir/[file2] Copies a file to the local mass store from the local disk. Takes two arguments.

• ccsm_splitdf [-d]/[-f]/[-h] dir/file Returns the directory path or the filename for an argument like dir/file. Can handle very general cases. -d returns the directory. -f returns the file. -h is help. -f is the default return argument. Takes one argument.

• modules.*.* Machine-specific setup scripts to run ccsm. No arguments. Called specifically by the main ccsm script.

• test.a1.mods.* These are notes about how to change the main test script to get it running on other machines.