Project

General

Profile

Wiki » Documentation » 1 Getting started »

Downloading the input files

Preamble

First of all, you need to figure out:
  • which input file version is available and needed for the specific ICON-HAM version you intend to use
  • if anyone has ever installed input files for ICON-HAM on your system (see platform-specific pages and check if your system's setup is already documented there). Installing input files is meant to be done by one person only, for each machine on which ICON-HAM is run. This person is named 'local input file administrator' thereafter. All other users of this system will then use the input files downloaded by the local input file administrator.

Three possible cases may happen:

  • there is no local input file administrator for your system: --> you may follow the instructions below in order to download the necessary input files on your system
  • there is a local input file administrator for your system, but the specific version you need is not installed yet (or not complete, e.g. missing resolution): --> you may ask your local input file administrator to get/update the necessary input file version
  • the specific version you need is already installed and complete: --> there is nothing to be done

Installing the input distribution toolkit

The input distribution toolkit is a set of a few scripts that ensure proper input files installation and update. In order to facilitate access to the new hammoz data server at TROPOS, it is strongly recommended to update a pre-existing input distribution toolkit to its latest version 2.1.
It can be downloaded via svn with:

svn checkout https://svn.iac.ethz.ch/external/echam-hammoz/input_distrib_toolkit/tags/input_distrib_toolkit-2.1 [name as you wish]

In order to use the toolkit scripts, you must add in your environment (in a ~/.bashrc or ~/.bash_profile, or via a module file):

declare -x PATH=/path/to/input_distrib_toolkit/bin:$PATH
declare -x MANPATH=/path/to/input_distrib_toolkit/share/man:$MANPATH

where you replace /path/to/input_distrib_toolkit by the proper path on your system.

Using the input distribution toolkit

Please read the manual in /path/to/input_distrib_toolkit/share/doc/pdf/users_guide_input_distrib_TK.pdf, as well as the man pages of:

  • get_input_distrib.sh <-- main command. It allows to reproduce input file distribution versions on a local machine as from a file repository server;
  • list_data_on_server.sh <-- lists the file hierarchy on a file repository server;
  • copy_file_hierarchy.sh <-- uses rsync to reproduce a full file hierarchy into a new folder;
  • compare_input_distribs.sh <-- shows the differences between two input file distributions (local to your machine).

For the commands involving the HAMMOZ input file remote server, the requested password is: getinput

Use the option -m icon for the commands to easily access the remote server at TROPOS (available from version 2.1).

This tells the toolkit to use the file /path/to/input_distrib_toolkit/access/icon_access.sh instead of /path/to/input_distrib_toolkit/access/echam_access.sh, which is the default.

Whenever invoking one of the above scripts (like list_data_on_server.sh or get_input_distrib.sh) use the option -m icon to access the icon related input file distribution.

Running the input distribution toolkit via a proxy

If your server is behind a firewall and accesses the internet via a proxy, you may have to set the environment variable RSYNC_PROXY to a relevant value for your server (ask your sys admin):

declare -x RSYNC_PROXY="someproxyaddress:someport"

Warning!! You should drop any potential 'http://' prefix from someproxyaddress.

Getting further help

  • If you have a question concerning the input distribution toolkit, please contact Sylvaine Ferrachat
  • If you have a question concerning any of the input files, please visit the input data pages first. If you can't find any answer to your question, or if you have a special request concerning potential customized input files, you may contact HAMMOZ input file requests. In this latter case, please do not forget to be as precise as possible in your request (resolution, specific years, etc...), and keep in mind that:
  • not all requests may be achievable
  • running ICON-HAM with customized input files is at the expense of user's own risks. The model may be untuned and not suited for performing valid scientific experiments.
  • do not forget to be as precise as possible in your request (resolution, specific years, etc...)

Next step: Starting a run (or platform-specific documentation)