PIC Cloud ATLAS coordination meeting

Name: PIC Cloud ATLAS coordination meeting
Start: 2009-06-03T11:00:00+02:00
End: 2009-06-03T12:19:00+02:00
Location: No location set

Wednesday 3 Jun 2009, 11:00 → 12:19 Europe/Zurich

Phone conference (CERN)

Xavier Espinal (PIC/IFAE)

Description

ATLAS coordination meeting for the PIC cloud (PIC, IFAE, IFIC, UAM and LIP) Bi-weekly lightweight format meeting, binded to one hour of duration. All the uncovered topics will be collected and covered offline within the mailing list (project-atlas-sw-t1-t2(at)cern.ch) or added to the next meeting. Agenda topics will be covered only in case relevant info exists. Phone conference details: +41227677000 (ask for the PIC CLOUD meeting with Xavier Espinal)

Support

project-atlas-sw-t1-t2(at)cern.ch

- 11:00 → 11:10
  Actions/issues from previous meetings 10m
  - STEP09 - Disk space watching 5m
    
    Disk space for each ST/site can be monitored in: https://sls-atlas.cern.ch/sls/service.php?id=ATLAS-storage_space Please have a regular look to ensure STEP09 data have enough room at your site: merged-AODs replicas, Functional Tests, Monte Carlo production and User Analysis activities depends on it. PIC, UAM, IFAE and LIPs are fine IFIC: no space left in DATADISK and MCDISK
- 11:10 → 11:40
  
  Round table Tier-1 and Tier-2s incidences/status 30m
- 11:40 → 11:55
  ATLAS activities 15m
  - STEP09 10m
    
    CCRC-like exercise (multi-VO test) User Analysis: - Pilots Role implemented at all Sites - New Pilot Factory at PIC to feed T2s with user analysis pilots. -- Logs at: http://vobox05.pic.es/ES-Factory-Analysis/logs/ -- Compat 32/64 bits problems at PIC (possibly at IFIC too ?): http://savannah.cern.ch/bugs/?51141 Monte Carlo: - G4 simulation started (1.8K jobs running) - http://panda.cern.ch:25980/server/pandamon/query?dash=prod Data transfers: - Throughput challenge started (last 24h as of today: 02-06-09@ 10h30') -- Good rates an efficiencies: - PIC cloud (~200MB/s). -- IFIC: 74MB/S -- IFAE: 35MB/s -- LIBSON: 66MB/s -- UAM: 15MB/s -- Noticed problems at LIP-COIMBRA T1-T2: http://atladcops.cern.ch:8000/drmon/ftmon_tier2s.html T1-T1: http://atladcops.cern.ch:8000/drmon/ftmon_T1-T1.html Wednesday: Start to add stress to some T1 tape systems with merge and reco and continue to monitor the progress of the reprocessing tasks.
    
    DataTransfers_Step009_030609
    
    MC_step09_030609.pdf
  - T2s report on STEP09 User Analysis challenge 10m
    
    As agreed we should coordinate to perform a follow up and report about the UA challenge during STEP09 in our cloud. ATLAS want T2s to report on: - Job efficiencies - Job stats (Event processing rate, Wall/CPU, I/O latencies, diffreence between remote reading and File Stager) - Job errors - Fair shares (AtlasPlain- AtlasPilot) From Jordi: + load in the wn + bandwidth per wn / global (farm) + efficiencies + share of different jobs (prd,pilot, analysis) + wall-time total + DN of people running STEP09 and how many real (outside step09) analysis jobs have run in parallel. Need to collect stats, plots and incidences at all Tier-2s (T2 responsibility) and work out a merged report to be presented to ATLAS (possibly during one of the Thursday's operations meeting)
- 11:55 → 12:00
  
  Review of actions for this meeting 5m
- 12:00 → 12:05
  
  A.O.B 5m