Client tools for Panda Analysis jobs

Intro

The panda-client package contains following tools to submit/manage analysis jobs on PanDA. The following instructions pertain to installing pandatools and using pathena.

• pathena
  • Athena jobs
• prun
  • ROOT/general jobs
• psequencer
  • Analysis chain
• pbook
  • Bookkeeping

Installation

The panda-client package is distributed in two kinds of format, tarball and rpm. Users can choose one format as they like. For installation on the MSU Tier 3, tarball is recommended.

tarball install

Go to the same directory that the Atlas Software Kit is in (for use on the MSU Tier 3),

cd /msu/data/dzero/ATLAS/programs

then do:

wget https://twiki.cern.ch/twiki/pub/Atlas/PandaTools/panda-client-0.1.39.tar.gz
tar xvfz panda-client-*
cd panda-client-*
python setup.py install --prefix=[install dir]

The most recent version (2009-Aug-7) of the panda-client tarball is 0.1.83. See here for more information.

Setup

First, You need to set PATHENA_GRID_SETUP_SH which configures Grid runtine environment. Ask your site admin if you don't know its location. Note that PATHENA_GRID_SETUP_SH must be .sh even if you are using csh or tcsh. e.g. On the MSU Tier 3, use BNL/acas do:

export PATHENA_GRID_SETUP_SH=/afs/usatlas.bnl.gov/lcg/current/etc/profile.d/grid_env.sh

Next, source the panda setup script according to your installation. The script sets PANDA_CONFIG_ROOT (~/.pathena by default) in addition to some environment variables. PANDA_CONFIG_ROOT is a directory name where user's configuration/data files are contained.

source [install dir]/etc/panda/panda_setup.[c]sh

What to do each time you use pathena

First source the athena setup script by doing

source [path to script]/athena_setup.sh

then source the pandatools setup script by doing

source [install dir]/etc/panda/panda_setup.sh

-- CharlieMueller - 13 Apr 2009