AstroGrid-D WG 5: Resource Management for Grid Jobs Report by: Rainer Spurzem (ZAH-ARI) spurzem@ari.uni-heidelberg.de and T. Brüsemeister, J. Steinacker
WG5: Resource Management for Grid Jobs Tasks Task V-1: Specification of Requirements and Architecture AIP (8), ARI-ZAH (6), ZIB (6), AEI (2), MPE (2), MPA (1) Start Sep. 05, Deliverable D5.1 Oct. 2006 COMPLETED Task V-2: Development of Grid-Job Management (Feb. 07) ZIB (24), ARI-ZAH (12), MPA (5) Start June 06, Deliverable D5.2 Feb., D5.6 June 2008 5.2 COMPLETED Task V-4 Adaptation of User- and Programmer Interfaces (May 07) AIP (18), ARI-ZAH (12), AEI (5), MPE (4), MPA (1) Start Dec. 06 Deliverable D5.4 May, D5.7 Sep. 2008 IN PROGRESS 5.4 Task V-3 Development Link to Robotic Telescopes, Requests (Feb 07) AIP (17), ZIB (6), Start Sep. 06 Deliverable D5.3 Feb., D5.5 Oct., D5.8 Sep. 2008 COMPLETED 5.3, IN PROGRESS (?) 5.5 (Person-Months) according to initial application (Person-Months) accordingastrogrid-d to initialmeeting application 2
GAC-Grid / Astro-GRID-D WG 5: Resource Management for Grid Jobs May 06 Sep 07 5.1 = D5.1 5.2 = D5.2 5.4a = D5.4 5.2 = D5.6 5.4b = D5.7 5.3a = D5.3 5.3b = D5.5 5.3c = D5.8 Are we here?
Entscheidung für die Job Submission Data Language (JSDL) wird vom open grid forum (OGF) unterstützt WG5: Current status, Job Management GUI JSDL jsdlproc RSL/XM L (GT4.2 wird gerade entwickelt und wird JSDL direkt unterstützen) GT4.0 4
NBODY6++: Another use case enters production phase Self-Adopted submit.sh (Thomas Brüsemeister) including deployment with configure, but using standard GridGateway Job Submission Template for further use cases? AstroGrid-D Meeting 5
NBODY6++ Pre- and Postprocessing by T. Brüsemeister AstroGrid-D Meeting 6
WG5: Current usage status, Scheduler/Broker with NBODY6++ Information System Stellaris Interaction? GT4/GW Resources Too few!! Gridway gwhost Matchmaking Scheduler / Broker Job Status: gwps 7
Grid Gateway / Globus Architecture basic system running, but problem with data staging at some sites accounting gridway / globus how to proceed? Central Server Architecture in principle, but backup servers (hydra.ari.uni-heidelberg.de) WG5: Current status, Scheduler/Broker Client Comple scheduling algorithm, self-adjusted by Gridway So far only GRAM adapter working, need PBS/torque/SGE at least in HD, titan/hydra Grid-Gateway Server hydra AstroGrid-D Meeting 8
NBODY6++ UseCase Status Accomplished: Prototype of deployment via configure inside GGW Job Eecution on some AstroGrid-D Resources Return of File Data (Staging) and Job Log(s) Robotic Telescopes In Immediate Progress within WG5: Parallel Processing (Adapter to PBS/SGE, Gridway Problem!) (is there any SGE implementation?) Encourage NBODY6++/Grid Production, other use cases (AG Meeting?) Ping s to Other Workgroups from WG5: Job Accounting: wrapper script, start/end to Stellaris (WG1, Task Force?) MDS problem, nodes do not report to Gridway (WG1) AstroGrid-D Meeting 9
To be started soon: User Portal, use case specific vs. general portal WG5: Net Steps, Job Submission, Scheduler/Broker AstroGrid-D Meeting (WG7, integration of legacy codes into service architecture!) Use of File Management System in JSDL Jobs (WG3, keep files local, stage files to grid-global file-id, location tracking?) Deployment etract standards from single solutions (modules, soft environment in globus? more parameters from information system, Task Force, WG1, WG5?) Etension of Information Service MDS, Stellaris (WG2, Resource Daten, Module-Info, Special Hardware, robotic telescopes, Job Statistik, WG1, Betriebskonzept) Use of Data Stream Management for nbody visualization? (WG6, so far only Unicore/FZJ implementation and demo) Medium to long-term goals: Interoperability with others EGEE, DEISA, Unicore, Global GRAPE Grid Workshop? Support for Workflows (as an etension to JSDL, Unicore?) Improve scheduling/brokering capabilities (special hardware, robotic telescopes, parallel/distributed jobs, global GRAPE Grid workshop) 10