Page History

...

Firstly, you need to obtain a membership in the ilc Virtual Organization (VO), which is sponsored by DESY. This is a VO on the LCG grid. Alternately, you can become a member of the calice VO, but this will only work if your institution is actually a member of this collaboration.

Follow the workflow at https://grid-voms.desy.de:8443/vovoms/ilc/register/vomrsstart.action which will require several rounds of confirmations and emails.

...

Setup

...

The following commands will download the VOMS file and add the information to the user config file in the home directory.

cd ~
mkdir .glite # only needed if .glite does not exist already-p ~/.glite/vomses
cd ~/.glite/vomses
wget http://grid.desy.de/etc/vomses/ilc-grid-voms.desy.de # fetch the VOMS file
touch vomses # make sure this file exists
cat ilc-grid-voms.desy.de >> vomses # add the ILC VOMS information

Now, the voms-proxy-init command should work correctly.

...

lfc-ls -l /grid/ilc/mc/ILC500/SM250fb-1/generated/

Fetching a Grid LCIO File to SLAC from DESY mass storage

Here is a more advanced script showing how to fetch a file to SLAC, assuming that you have an lxplus account at CERN.

# ssh to cern
ssh lxplus

# setup grid tools
source /afs/cern.ch/project/gd/LCG-share/current/etc/profile.d/grid_env.sh

# start a grid session
voms-proxy-init -verify -debug -voms ilc

# copy the file to temp location on lxplus (it's fast)
lcg-cp -v --vo ilc lfn:/grid/ilc/mc-2008/simulated/LDC01_05Sc/ZPole/M-6-5p2_ZPoleuds_LDC01_05Sc_LCP_Test_ZPole_0001.slcio file:/tmp//M-6-5p2_ZPoleuds_LDC01_05Sc_LCP_Test_ZPole_0001.slcio

# copy file to slac (slow...go get some coffee)
scp M-6-5p2_ZPoleuds_LDC01_05Sc_LCP_Test_ZPole_0001.slcio jeremym@iris01.slac.stanford.edu:/afs/slac/g/lcd/public_data/ILC/test/mokka/

The user name jeremym should be replaced above with your real SLAC account name.

Test Commands

The GLITE UI will be used to run test jobs. It uses commands starting with "edg-" for job submission, monitoring, etc.

...

edg-job-status -i slic.jdljid

Retrieve the output. By, default the job output will go into a directory in /tmp so we specify the current directory instead.

edg-job-get-output --dir `pwd` -i slic.jdljid

Links

gLite User Guide
Virtual Data Toolkit (VDT)
CERN AFS UI Setup - setup instructions on lxplus@cern