110 likes | 344 Views
Grid. NIKHEF Jamboree 22 december 2005. Jeff Templon PDP Group, NIKHEF. Throbbing jobs. Googled Grid. The PDP spear points. Operations Installing, operating, maintaining NIKHEF grid systems Feed improvements back into system Worldwide Deployment Coordination (Bos: GDB Chair) Software
E N D
Grid NIKHEF Jamboree 22 december 2005 Jeff Templon PDP Group, NIKHEF Throbbing jobs Googled Grid
The PDP spear points • Operations • Installing, operating, maintaining NIKHEF grid systems • Feed improvements back into system • Worldwide Deployment Coordination (Bos: GDB Chair) • Software • Security (Groep, Venenkamp, Koeroo, Steenbakkers) • How to organize, identify, authorize, manage user groups • Big win: important aspect with few experts (several @ NIKHEF) • Integration / certification / packaging for VL-E (van Dok, Keijser) • Worldwide Grid “trust management” coordination (Groep: IGTF Chair) • Applications (Klous, Templon, Groep) • Helping people / groups become effective grid users • Evangelism • ATLAS Trigger (on Grid)
PDP operations • Ronald Starink, David Groep, JT, Paul Kuipers, Ton Damen • Some statistics for last six weeks • 27 problem reports • 21 Willem van Leeuwen • 2 Philips • 2 LCG operations team • 1 SARA • 1 ATLAS operations • Almost all concern job submission problems
PDP ops 2 • From our system actions log: • 45 restarts of batch system components • 16 broken nodes removed from system for repairs • 14 repaired nodes returned to service • 13 restarts of ‘nscd’ service • 5 restarts of resource broker • Handful of node reboots (‘stuck’) • One major upgrade (batch system torque 1.2 -> 2.0)
Common Question to Grid Guys Why are physicists at a HEP lab spending timeworrying about computing problems inbioinformatics and earth sciences? BIG GRID
What is BIG GRID? • Infrastructure based on supporting science cases • Biology (food informatics, drug discovery, proteomics) • Linguistic and cultural studies • Astronomy (LOFAR) • High-Energy Physics (LHC) • Industrial Research (GEANT4 sims by Philips!!!) • Total budget 30 M€ (almost all either stuff or operating costs)
What’s in BIG GRID? • THE WORKS for HEP: full Tier-1 for ATLAS / LHCb / ALICE • Distributed / central computing infra for biomed • LOFAR data center • Cluster for in silico industrial research (hosted by Philips) • “other” -- common pool for disciplines who don’t have good estimates • All tied together with grid tools & infrastructure • Operating funds for four years (cooling / electricity) + small operational manpower budget
Writing Team: Kors Bos, David Groep, Frank Linde, Arjen van Rijn, JTBob Hertzberger (UvA), Gert Vriend (Nijmegen), Peter Michielse (NCF) Proposal Pre-Rejection: Gerard v/d Steenhoven, Stan Bentvelsen,Marcel Merk Mr. Telephone: Arjen van Rijn Presentation at NWO Interview:Frank Linde
What’s Up for 2006 • BIG GRID organization • EGEE-2 starts (approved last week!) • Ramp Up (CPU factor 3, disk factor 10) • Calm down LCG management (deploying empty capacity) • Guidance of new BIG / VL-E user communities • Continued improvements in facilities & software • Experiments: please test ANALYSIS models!!! • Experiment Computing Operations • Assumed coming from T1 physics groups • Do we want this??
Progress at SARA LCG Service Challenge II “The Dutch Contribution”