Use of SRMCP at UCSD Tier2

A brief note of how to use srmcp at the UCSD tier2 center from a job script.

I assume in this document that there already is a file in dcache that you want access to. If you want to copy into the dcache system via srmcp make sure your proxy and/or role has write access to the target area. Most areas are readable by any proxy.

From a Compute Node (Accessible via the Batch System)

Create your job script in a shell language. Bash or the Bourne Shell are recommended and are what I use for testing. Make sure you source the $VDT_LOCATION/ in your script. You can then use srmcp in your script as;

srmcp -debug=true srm:// file://localhost//state/data/osgtmp/out-$FNAME.file

In this example we are transferring to the local compute nodes temporary directory.

A few notes

  • You should create a directory in the local nodes temporary space that is unique to your job. For example by adding the current time in seconds since Epoch to the directory name
  • Any files you create in the temporary area should be removed as soon as the job completes.

From the User Analysis Machines (eg.

source /data/vdt/

Then run srmcp

srmcp -debug=true srm:// file://localhost//$PWD/

This should copy a file from the dcache system, via srmcp into the current directory and call the copied file

-- BruceThayre - 2009/06/11

Topic revision: r1 - 2009/06/11 - 00:32:45 - BruceThayre
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback