Difference between revisions of "Atlas:Analysis Challenge"

Un article de lcgwiki.
Jump to: navigation, search
(<b> Phase 1 : Site stress test run centrally in a controlled manner (2 days) </b>)
(Phase 1 : Site stress test oraganized by ATLAS and run centrally in a controlled manner (2 days))
 
(95 intermediate revisions by 2 users not shown)
Ligne 1: Ligne 1:
 +
21/01/2009 : ''E.Lançon, F.Chollet <br>
 +
Thanks to Johannes Elmsheuser, Dan van der Ster, Cédric Serfon''
 +
 +
====Information & Contact ====
 +
Mailing list ATLAS-LCG-OP-L@in2p3.fr
 +
 
==== Goals ====
 
==== Goals ====
 
* measure "real" analysis job efficiency and turn around on several sites of a given cloud
 
* measure "real" analysis job efficiency and turn around on several sites of a given cloud
Ligne 5: Ligne 11:
 
* check load balancing between analysis and MC production
 
* check load balancing between analysis and MC production
  
==== First exercise on the FR Cloud (>= December 8th ) ====
+
==== Required services @ T1 and GRIF====
===== Phase 1 : Site stress test run centrally in a controlled manner (2 days) =====
+
* LFC catalog : lfc-prod.in2p3.fr
DA challenges have been performed on IT and DE clouds in october 08.
+
* ATLAS Disk space : ATLASUSERDISK on T1 SE (fail-over for outputs in case of problems with T2 disk storage)
Proposition has been made to extend this cloud-by cloud challenge to the FR Cloud.
+
* GRIF and CC-IN2P3 TOP BDII : should be available as they are used remotely by some sites
See  ATLAS coordination DA challenge meeting (Nov. 20) 
 
* http://indico.cern.ch/conferenceDisplay.py?confId=45718
 
First exercise will help to identify breaking points and bottlenecks. It is limited in time (a few days) and requires careful attention of site administrators during that period,in particular network (internal & external), disk, cpu monitoring.
 
This first try (Stress tests) can be run centrally in a controlled manner. ATLAS coordination (Dan van der Ster and Johannes Elmsheuser) needs to know which sites to be tested and when.
 
* See Results : http://gangarobot.cern.ch/st/
 
* Nov. 28 Submission (Tot 200 jobs sur 12 sites): http://gangarobot.cern.ch/st/test_43/
 
* Dec. 8-9: 1rst round with Tokyo and possibly GRIF 
 
* Dec. 15-16 :
 
**possibly : LAPP, CC-IN2P3-T2(to be confirmed), Tokyo, GRIF,
 
**sites with 1gbps LAN : CPPM, NIPNE, LPC (to be contacted)
 
  
===== <b> Phase 2 :Pathena Analysis Challenge  </b>=====
+
==== Target and metrics ====
* Dec 17 : Restart of MC production
+
* Nb of events : Few hundred up to 1000 jobs/site
* Dec 17-18 : Data Analysis exercice open to physicists with their favorite application and tools
+
* Rate (evt/s) : up to 15 Hz
 +
* Efficiency (success/failure rate) : 80 %
 +
* CPU utilization :  CPUtime / Walltime > 50 %
  
==== Procedure ====
+
==== FR Cloud ST (2009 plans) ====
* Replication of target datasets accross the cloud
+
* See http://lcg.in2p3.fr/wiki/index.php/Atlas:Analysis_ST_2009
* Preparation of job
 
* Generation n jobs per site (Each job processes 1 dataset)
 
* Bulk submission to WMS (1 per site)
 
  
==== Test conditions ====
+
====Test report sent to ATLAS (D.van der Ster and J. Elmsheuser) Jan.2009 ====
 +
[http://lcg.in2p3.fr/wiki/images/FR-ST82-Feedback.pdf http://lcg.in2p3.fr/wiki/skins/common/images/icons/fileicon-pdf.png]
  
The testing framework is ganga-based. It is currently using LCG backend but it will soon be possible to use PANDA backend as well. Both POSIX I/O and "copy mode" may be used allowing performances comparaison of the 2 modes. <br>
+
==== First exercise on the FR Cloud (December 2008) ====
It uses regular AOD analysis in 14.2.20 with mc08*AOD*e*s*r5 DQ2 inputs<br>
 
Input datasets are read from ATLASMCDISK and outputs are stored on ATLASUSERDISK (no special requirements there). Input data access is the main issue. No problem on data output <br>
 
Required CPUtime :~ 1 day (typical job duration 5 h)<br>
 
LAN saturation observed at ~ 3 Hz in case of 1 Gb network connection between WN and SE. <br>
 
  
* GangaRobot
+
=====  Phase 1 : Site stress test organized by ATLAS and run centrally in a controlled manner (2 days) =====
DA challenge runs similar to existing analysis functional tests : http://gangarobot.cern.ch/<br>
+
DA challenges have been performed on IT and DE clouds in october 08.
See [http://indico.cern.ch/getFile.py/access?contribId=63&sessionId=11&resId=0&materialId=slides&confId=22137 J.Elmsheuser's presentation (Nov. 5 08)]<br>
+
Proposition has been made to extend this cloud-by cloud challenge to the FR Cloud.
<b> Site should pass the GangaRobot test successfully </b> , especially :
+
See  ATLAS coordination DA challenge meeting (Nov. 20) 
* http://gangarobot.cern.ch/20081119_02/index.html
+
* http://indico.cern.ch/conferenceDisplay.py?confId=45718
* http://gangarobot.cern.ch/20081119_03/index.html
+
First exercise will help to identify breaking points and bottlenecks. <b>It is limited in time (a few days) and requires careful attention of site administrators during that period,in particular network (internal & external), disk, cpu monitoring.</b>
 +
This first try (Stress tests) can be run centrally in a controlled manner. The testing framework is ganga-based.
 +
* Any site in the Tiers_of_ATLAS list can participate. 
 +
* ATLAS coordination : Dan van der Ster and Johannes Elmsheuser
 +
* [http://lcg.in2p3.fr/wiki/index.php/Atlas:Analysis_Challenge_ST  Details of Site Stress test] : procedure, test conditions and targets
 +
* GlueCEPolicyMaxCPUTime >= 1440 (1 day , typical duration : 5 hours)
 +
* Jobs run under DN : /O=GermanGrid/OU=LMU/CN=Johannes_Elmsheuser
 +
* <b>Results </b> available here : http://gangarobot.cern.ch/st/
 +
* <b>Test 43 - Nov. 28 </b> http://atlas-ganga-storage.cern.ch/test_43/
 +
* <b>Test 61 - December 8-10</b> http://atlas-ganga-storage.cern.ch/test_61/
 +
Sites: IN2P3-LPC, GRIF-LPNHE, TOKYO-LCG2, IN2P3-CPPM
 +
Max Jobs Per Site: 300
 +
* <b>Test 62 - December 8-10</b> http://atlas-ganga-storage.cern.ch/test_62/
 +
Sites: TOKYO
 +
Max Jobs Per Site: 300
 +
Input DS Patterns: mc08.*Wmunu*.recon.AOD.e*_s*_r5*tid*
 +
                    mc08.*Zprime_mumu*.recon.AOD.e*_s*_r5*tid*
 +
                    mc08.*Zmumu*.recon.AOD.e*_s*_r5*tid*
 +
                    mc08.*T1_McAtNlo*.recon.AOD.e*_s*_r5*tid*
 +
                    mc08.*H*zz4l*.recon.AOD.e*_s*_r5*tid*
 +
                    mc08.*.recon.AOD.e*_s*_r5*tid*
 +
* <b>Test 82 - December 15-17</b> http://atlas-ganga-storage.cern.ch/test_82/
  
<b>Participation required at cloud and site level. Any site in the Tiers_of_ATLAS list can participate.</b>
+
=====[http://lcg.in2p3.fr/wiki/index.php/Atlas:Analysis_Challenge-STsummary  ST 2008 summary]=====
  
  It is possible for sites to limit the number of jobs sent at a time.
+
===== Phase 2 : Pathena Analysis Challenge =====
DA team is ready to take into account site constraints.
 
DA team is open to any metrics
 
  
=== Target and metrics ===
+
* Data Analysis exercice open to physicists with their favorite application
* Nb of events : Few hundred up to 1000 jobs/site
+
* Physicists involved : Julien Donini, Arnaud Lucotte, Bertrand Brelier, Eric Lançon, LAL ?, LPNHE ?
* Rate (evt/s) : up to 15 Hz
 
* Efficiency (success/failure rate) : 80 %
 
* CPU utilization : CPUtime / Walltime > 50 %
 
  
=== Results ===
+
===== Dec. 2008 Planning =====
See
 
* ATLAS Twiki page : https://twiki.cern.ch/twiki/bin/view/Main/GangaSiteTests
 
* Results of analysis challenge performed on [http://indico.cern.ch/getFile.py/access?contribId=128&sessionId=8&resId=0&materialId=slides&confId=22137 IT Cloud]
 
* Results of [http://indico.cern.ch/getFile.py/access?contribId=129&sessionId=8&resId=0&materialId=slides&confId=22137 DE Cloud]
 
  
=== 2009 plans ===
+
* Dec 8 : stop of MC production
 +
* Dec. 8-9: <b>1rst round </b>with Tokyo, CPPM, LPC (LAN limited to 1Gbps), GRIF-LPNHE
 +
* Dec 17 : restart of MC production
 +
* Dec 14 : stop of MC production
 +
* Dec. 15-16 : <b>2nd round </b>with LAPP, CC-IN2P3-T2 (to be confirmed), Tokyo, CPPM, LPC, possibly GRIF (SACLAY, IRFU, LPNHE), RO-07 and RO-02
 +
* Dec 17 : <b> restart of MC production</b>
 +
* Dec 17 : <b> Beginning of Analysis Challenge (Phase 2) </b>

Latest revision as of 17:06, 12 mars 2009

21/01/2009 : E.Lançon, F.Chollet
Thanks to Johannes Elmsheuser, Dan van der Ster, Cédric Serfon

Information & Contact

Mailing list ATLAS-LCG-OP-L@in2p3.fr

Goals

  • measure "real" analysis job efficiency and turn around on several sites of a given cloud
  • measure data access performance
  • check load balancing between different users and different analysis tools (Ganga vs pAthena)
  • check load balancing between analysis and MC production

Required services @ T1 and GRIF

  • LFC catalog : lfc-prod.in2p3.fr
  • ATLAS Disk space : ATLASUSERDISK on T1 SE (fail-over for outputs in case of problems with T2 disk storage)
  • GRIF and CC-IN2P3 TOP BDII : should be available as they are used remotely by some sites

Target and metrics

  • Nb of events : Few hundred up to 1000 jobs/site
  • Rate (evt/s) : up to 15 Hz
  • Efficiency (success/failure rate) : 80 %
  • CPU utilization : CPUtime / Walltime > 50 %

FR Cloud ST (2009 plans)

Test report sent to ATLAS (D.van der Ster and J. Elmsheuser) Jan.2009

http://lcg.in2p3.fr/wiki/skins/common/images/icons/fileicon-pdf.png

First exercise on the FR Cloud (December 2008)

Phase 1 : Site stress test organized by ATLAS and run centrally in a controlled manner (2 days)

DA challenges have been performed on IT and DE clouds in october 08. Proposition has been made to extend this cloud-by cloud challenge to the FR Cloud. See ATLAS coordination DA challenge meeting (Nov. 20)

First exercise will help to identify breaking points and bottlenecks. It is limited in time (a few days) and requires careful attention of site administrators during that period,in particular network (internal & external), disk, cpu monitoring. This first try (Stress tests) can be run centrally in a controlled manner. The testing framework is ganga-based.

Sites: IN2P3-LPC, GRIF-LPNHE, TOKYO-LCG2, IN2P3-CPPM 
Max Jobs Per Site: 300
Sites: TOKYO
Max Jobs Per Site: 300
Input DS Patterns: mc08.*Wmunu*.recon.AOD.e*_s*_r5*tid* 
                   mc08.*Zprime_mumu*.recon.AOD.e*_s*_r5*tid* 
                   mc08.*Zmumu*.recon.AOD.e*_s*_r5*tid* 
                   mc08.*T1_McAtNlo*.recon.AOD.e*_s*_r5*tid* 
                   mc08.*H*zz4l*.recon.AOD.e*_s*_r5*tid* 
                   mc08.*.recon.AOD.e*_s*_r5*tid*
ST 2008 summary
Phase 2 : Pathena Analysis Challenge
  • Data Analysis exercice open to physicists with their favorite application
  • Physicists involved : Julien Donini, Arnaud Lucotte, Bertrand Brelier, Eric Lançon, LAL ?, LPNHE ?
Dec. 2008 Planning
  • Dec 8 : stop of MC production
  • Dec. 8-9: 1rst round with Tokyo, CPPM, LPC (LAN limited to 1Gbps), GRIF-LPNHE
  • Dec 17 : restart of MC production
  • Dec 14 : stop of MC production
  • Dec. 15-16 : 2nd round with LAPP, CC-IN2P3-T2 (to be confirmed), Tokyo, CPPM, LPC, possibly GRIF (SACLAY, IRFU, LPNHE), RO-07 and RO-02
  • Dec 17 : restart of MC production
  • Dec 17 : Beginning of Analysis Challenge (Phase 2)