r2 - 28 Feb 2012 - 16:34:39 - ShuweiYeYou are here: TWiki >  AtlasSoftware Web > GlobusUsecases

Use Cases for Globus Online operation withing the PanDA Framework


This page is intended for documentation of actual use cases for the integrated Globus Online operation within the PanDA framework. The detailed technical aspects of Globus/Panda interaction are documented in a separate document.

Globus Online is a Web service and a set of Grid tools that provides management, automation and monitoring tools for data transfers in the grid environment. In the context of PanDA, it has the potential to facilitate work of smaller groups of researchers (PanDA users) who would like to be flexible in managing transfer of data (including to and from their personal desktops and laptops).

Currently, the functionality of Globus Online is utilized in PanDA in the following manner:

  • there are Panda queues which handle globus-aware PanDA Pilots
  • users utilize existing tools (such as prun) to define destination point for the data
  • the pilot will take care of transferring the data

Use Case 1: job submission to T3 via prun

This case demonstrates how to submit a job via prun to a site with the integrated Globus Online operation within the PanDA framework, and how to access to the globus online endpoint of the job output.

At BNL, a testbed site "ANALY_BNL_T3" has been created for the integrated Globus Online operation within the PanDA framework. 1. First set up the environment of grid and panda, then submit your job via prun

    source /opt/usatlas/bin/panda_setup.sh=
2. Then submit your job using prun
   prun --exec "echo %IN | sed 's/,/\n/g' > input.txt; ./runPanda-BNL.sh" \
    --excludeFile=".svn" --noBuild \
    --useChirpServer MWT2 --nFiles=5 --outputs="myOutput.root" --site=ANALY_BNL_T3 \
    --inDS=data11_7TeV.00180149.physics_Egamma.merge.NTUP_TOP.f368_m806_p530_p577_tid367196_00 \
3. After the job is done, you can look up the physical location of your output files in pilot.txt.
source endpoint 1s7t8j 
201 Created {u'globus_connect_setup_key': u'9d1846cd-fe38-4848-a0fc-3ac88317dc3a', u'resource': u'/endpoint', u'DATA_TYPE': u'endpoint_create_result', u'canonical_name': u'wdeng#1s7t8j', u'code': u'Created', u'request_id': u'mgtxM8Zlq', u'message': u'Endpoint created successfully'}
Activating 9d1846cd-fe38-4848-a0fc-3ac88317dc3a
Activation result Cleaning up files from version .10
Configuration directory: /usatlas/grid/usatlas1/.globusonline/lta
Contacting relay.globusonline.org:2223
Installing certificate and key
Creating /usatlas/grid/usatlas1/.globusonline/lta/gridmap
1s7t8j MWT2_DCACHE /data/panda/tmp/Panda_Pilot_845_1330373939/PandaJob_1436291794_1330373942/user.yesw.000112._00001.myOutput.root /pnfs/uchicago.edu/atlasuserdisk/user/wdeng/d986441f-f9fa-42f6-9726-5e558d2e978e/user.wdeng.d986441f-f9fa-42f6-9726-5e558d2e978e.HIST.120216162935_sub052228798/user.yesw.000112._00001.myOutput.root
task No. 61ef7edc-6182-11e1-81e6-1231381bd061
task: ACTIVE
task: ACTIVE
task: ACTIVE
Globus Connect is currently running and connected to Globus Online
4. You need set up grid env and create a grid-certificate first.
    source /afs/usatlas.bnl.gov/lcg/current/etc/profile.d/grid_env.sh
    voms-proxy-init -voms atlas
5. Now you can list the files at Globus Online endpoint "wdeng#MWT2_DCACHE".
    gsissh cli.globusonline.org
    $ help
    $ endpoint-list -v wdeng#MWT2_DCACHE
    $ endpoint-activate -g wdeng#MWT2_DCACHE
    $ ls -l wdeng#MWT2_DCACHE/
    $ ls -l wdeng#MWT2_DCACHE/pnfs/uchicago.edu/atlasuserdisk/user/wdeng/d986441f-f9fa-42f6-9726-5e558d2e978e/user.wdeng.d986441f-f9fa-42f6-9726-5e558d2e978e.HIST.120216162935_sub052228798/user.yesw.000112._00001.myOutput.root
    $ ls -l wdeng#MWT2_DCACHE:/pnfs/uchicago.edu/atlasuserdisk/user/wdeng/d986441f-f9fa-42f6-9726-5e558d2e978e/user.wdeng.d986441f-f9fa-42f6-9726-5e558d2e978e.HIST.120216162935_sub052228798/user.yesw.000112._00001.myOutput.root
6. You should be able to copy the above file onto any Globus endpoint at your desktop, or laptop, on which you just run Globus Connect, and input the setup key generated via gsissh command endpoint-add
$ endpoint-add yesw2000#desktop --gc
Created the Globus Connect endpoint 'yesw2000#desktop'.
Use this setup key when installing Globus Connect: 
7. Now can you can copy your output files on the source Globus Endpoint "wdeng#MWT2_DCACHE" to your own destination Globus Endpoint "yesw2000#desktop".
$ scp wdeng#MWT2_DCACHE:/pnfs/uchicago.edu/atlasuserdisk/user/wdeng/d986441f-f9fa-42f6-9726-5e558d2e978e/user.wdeng.d986441f-f9fa-42f6-9726-5e558d2e978e.HIST.120216162935_sub052228798/user.yesw.000112._00001.myOutput.root yesw2000#desktop:/tmp/yesw/
Please be aware that scp only accepts the format source_point:source_dir, while command "ls" also accepts the format source_point/source_dir.

Use Case 2

Use Case 3

About This Site

Please note that this site is a content mirror of the BNL US ATLAS TWiki. To edit the content of this page, click the Edit this page button at the top of the page and log in with your US ATLAS computing account name and password.


Powered by TWiki
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