ARC1/Usage

NOTE: This page is out of date and is kept for purely historical reasons.

This page describes how to quickly test the ARC1 clients. If you are looking for server side configuration follow the instructions: Server install and setup instructions

If a production ARC release (v0.6 or v0.8) has been installed, the system needs to be able to find the Globus shared libraries. Otherwise warnings will be displayed since all other installed pugins are loaded. To avoid these messages you can do something like:

 export LD_LIBRARY_PATH=/opt/globus/lib

A-REX service (job submission and management)

• generate your proxy certificate with grid-proxy-init -rfc or with arcproxy utility (see --help for usage). When using arcproxy the following commands may help to setup the certificate/proxy environment if the command-line options are not preferred:

export X509_USER_CERT=$HOME/.globus/usercert.pem
export X509_USER_KEY=$HOME/.globus/userkey.pem
export X509_USER_PROXY=`mktemp /tmp/x509up.XXXXXX`

• in any case make sure that the X509_USER_PROXY env variable points to the location of your proxy certificate. grid-proxy-init requires the following setting:

 export X509_USER_PROXY=/tmp/x509up.u`id -u`

• create a job description in JSDL form (e.g. http://vls.grid.upjs.sk/testing/job_descriptions/get_hostname.html) and save it as .xml. Please note that the current release requires that the executable (if locally staged) is explicitly listed as an inputfile.
• submit a job to your A-REX service (you have to know your A-REX endpoint) using arcsub command (see man arcsub for more details)

lynx  --dump http://vls.grid.upjs.sk/testing/job_descriptions/get_hostname.html | sed '1,5d' > your_job.xml
arcsub -c ARC1:https://example.org:60000/arex your_job.xml  # (a jobID will be returned)

• monitor the status of your job with arcstat (see man arcstat for more details)

arcstat -a
arcstat jobID

• get the results of finished job using arcget command (see man arcget for more details)

arcget -a 
arcget jobID

• clean your job on a cluster using arcclean command (see man arcclean for more details)

arcclean jobID

• kill your running job using arckill command (see man arckill for more details)

arckill jobID

ECHO service (simple WS test)

For version 0.9.1 things have still been a bit more complicated.

• edit the example config file (echo_client.xml.example) located in $ARC_LOCATION/share/doc/nordugrid-arc1-client-0.9.1 according to your ECHO service configuration
• point the ARC_ECHO_CONFIG env variable to your echo_client.xml file (you do not need to do that step if you will issue arcecho command from directory where your echo_client.xml file is located)
• use arcecho for getting the echo response

arcecho hello

versions since 1/2009 (0.9.3+) can specify the server directly. In this case the client.xml config file needs to be correctly configured

arcecho https://localhost:60000/Echo hallo

HOPI service (simplified HTTP server)

• point your favourite web browser to your HOPI endpoint url (e.g. http://example.org:50000/httpd). You should see the content of /tmp/httpd directory.
• use wget or curl to PUT, GET the files to/from your HOPI service (again use the same endpoint url here)

Storage service

• for this step see detailed description in a Storage service README file

CHARON service (policies)

• for this step see the detailed description in Charon client README and Charon service README
• usage of CHARON service is related to proper configuration of charon_policy.xml which was defined on the CHARON service side
• Notice that to use Charon client a userkey has to be unencrypted