February WLC GDB Short summaries

Slides:



Advertisements
Similar presentations
CERN IT Department CH-1211 Geneva 23 Switzerland t T0 report WLCG operations Workshop Barcelona, 07/07/2014 Maite Barroso, CERN IT.
Advertisements

Database Services for Physics at CERN with Oracle 10g RAC HEPiX - April 4th 2006, Rome Luca Canali, CERN.
Introduction to CVMFS A way to distribute HEP software on cloud Tian Yan (IHEP Computing Center, BESIIICGEM Cloud Computing Summer School.
LCG Introduction John Gordon, STFC-RAL GDB September 9 th, 2008.
WebFTS File Transfer Web Interface for FTS3 Andrea Manzi On behalf of the FTS team Workshop on Cloud Services for File Synchronisation and Sharing.
Data Transfer Service Challenge Infrastructure Ian Bird GDB 12 th January 2005.
Feedback from CMS Andrew Lahiff STFC Rutherford Appleton Laboratory Contributions from Christoph Wissing, Bockjoo Kim, Alessandro Degano CernVM Users Workshop.
Testing Infrastructure Wahid Bhimji Sam Skipsey Intro: what to test Existing testing frameworks A proposal.
Development of a Tier-1 computing cluster at National Research Centre 'Kurchatov Institute' Igor Tkachenko on behalf of the NRC-KI Tier-1 team National.
CernVM-FS – Best Practice to Consolidate Global Software Distribution Catalin CONDURACHE, Ian COLLIER STFC RAL Tier-1 ISGC15, Taipei, March 2015.
Considerations on Using CernVM-FS for Datasets Sharing Within Various Research Communities Catalin Condurache STFC RAL UK ISGC, Taipei, 18 March 2016.
The StratusLab Distribution and Its Evolution 4ème Journée Cloud (Bordeaux, France) 30 November 2012.
CVMFS Alessandro De Salvo Outline  CVMFS architecture  CVMFS usage in the.
HEPiX Fall 2011 Summary Report QMUL 10 November 2011 Martin Bly, STFC-RAL.
LCG Introduction John Gordon, STFC GDB December 7 th 2010.
Scientific Data Processing Portal and Heterogeneous Computing Resources at NRC “Kurchatov Institute” V. Aulov, D. Drizhuk, A. Klimentov, R. Mashinistov,
AFS Phase Out experience R. De Maria. AFS and LSF Phase Out (1) I have learned about relevant changes in IT services, only while iterating with IT experts.
ATLAS – statements of interest (1) A degree of hierarchy between the different computing facilities, with distinct roles at each level –Event filter Online.
CERN IT-Storage Strategy Outlook Alberto Pace, Luca Mascetti, Julien Leduc
CernVM-FS vs Dataset Sharing
Dynamic Extension of the INFN Tier-1 on external resources
WLCG IPv6 deployment strategy
Review of the WLCG experiments compute plans
WLCG Workshop 2017 [Manchester] Operations Session Summary
DPM at ATLAS sites and testbeds in Italy
Status: ATLAS Grid Computing
Use of HLT farm and Clouds in ALICE
WLCG Network Discussion
Ian Bird WLCG Workshop San Francisco, 8th October 2016
Ian Bird, CERN & WLCG CNAF, 19th November 2015
Helge Meinhard, CERN-IT Grid Deployment Board 04-Nov-2015
Sviluppi in ambito WLCG Highlights
Report from WLCG Workshop 2017: WLCG Network Requirements GDB - CERN 12th of July 2017
Data Analytics and CERN IT Hadoop Service
Hadoop and Analytics at CERN IT
Virtualisation for NA49/NA61
Dag Toppe Larsen UiB/CERN CERN,
Dag Toppe Larsen UiB/CERN CERN,
Diskpool and cloud storage benchmarks used in IT-DSS
The “Understanding Performance!” team in CERN IT
Database Services at CERN Status Update
CC - IN2P3 Site Report Hepix Spring meeting 2011 Darmstadt May 3rd
Storage Interfaces and Access: Introduction
CREAM Status and Plans Massimo Sgaravatto – INFN Padova
WLCG experiments FedCloud through VAC/VCycle in the EGI
Investigation of the improved performance on Haswell processors
Introduction to CVMFS A way to distribute HEP software on cloud
Update on Plan for KISTI-GSDC
Status and Prospects of The LHC Experiments Computing
Virtualisation for NA49/NA61
Storage for Science at CERN
CERN Lustre Evaluation and Storage Outlook
Savannah to Jira Migration
Passive benchmarking of ATLAS Tier-0 CPUs
HPEiX Spring RAL Site Report
Dagmar Adamova (NPI AS CR Prague/Rez) and Maarten Litmaath (CERN)
Update from the HEPiX IPv6 WG
Project Status Report Computing Resource Review Board Ian Bird
John Gordon, STFC GDB October 12th 2011
John Gordon, STFC GDB April 6th 2011
CernVM Status Report Predrag Buncic (CERN/PH-SFT).
Bernd Panzer-Steindel CERN/IT
CTA: CERN Tape Archive Overview and architecture
WLCG Collaboration Workshop;
Haiyan Meng and Douglas Thain
Using an Object Oriented Database to Store BaBar's Terabytes
CHIPP - CSCS F2F meeting CSCS, Lugano January 25th , 2018.
This work is supported by projects Research infrastructure CERN (CERN-CZ, LM ) and OP RDE CERN Computing (CZ /0.0/0.0/1 6013/ ) from.
IPv6 update Duncan Rand Imperial College London
Building a minimum viable Security Operations Centre
Presentation transcript:

February WLC GDB Short summaries Jeremy Coles

Introduction March GDB co-located with ISGC April pre-GDB – collaborating with other communities https://indico.cern.ch/event/578969/ WISE 27th-29th March; LHCOPN-LHCONE 4-5 April; 24th-28th April HEPiX Budapest…

AFS phaseout @ CERN Slow demise of upstream project Deadline ‘soft’ <2yrs Easy: Software-> CVMFS; websites->EOSWEB; FUSE. https://its.cern.ch/jira/browse/NOAFS 2017 for “harder stuff”. Project migrations; remove software. Moving /work and /user. Non-grid experiment use of AFS (home dir, T0 activity,..). Occasional GGUS tickets. “Misconfigured” sites. External AFS disconnection test. 2017-02-15 09:00 CET 24hts. ITSSB entry. Goal to flush out unknown AFS dependencies. Example /afs/cern.ch user home directories. Trying to setup CERN only CVMFS for compiler type software.

Pre-GDB on Benchmarking Mainly HEPiX WG context Mandate: Investigate scaling issues (HS06 vs HEP workloads); Evaluate fast benchmarks; study the next generation of long-running benchmark. Fast benchmarks: converged to DB12 & Atlas KV (looked at job durations, approach to running benchmarks; highlight implications on resource accounting). Linearity of KV and DB12 vs job duration has been demonstrated. But reco, analysis, merge and skim have non-neg I/O component and will not scale well. Cloud environment ‘whole node’ performance. Scaling factor for HS06, KV and DB12 within expectations. Not everything understood vs bare metal. Passive Benchmarks – use real jobs. Approaches to running benchmarks: benchmark in the pilot (not poss on HPC….) Two options provided by LHCb. (DB12 GitHUB). Run in job or at boot. Cloud Benchmark Suite: Toolkit (wrap KV, DB12). Internal to CERN. Adopted by others. Open questions: KV (not discussed so much. DB12 easier to install). KV based on ATLAS Athena (large code base. License issues). But KV can highlight second order effects (interplay CPU speed and mem access). Magic boost with Haswell of 45% over Sandy Bridge. Major contributor is the Cpython interpreter. Turns out boost is from Branch Prediction. ALICE and LHCb happy with DB12. ATLAS still evaluating.

EOS workshop Overview of workshop EOS releases (gem stone naming) Tags to track changes. Currently at CERN 150PB raw capacity Storage node + storage array = Block Filesystem access – FUSE Python Notebook Integration (SWAN) Collaborative editing via MS Office. Australian – distributed setup. Earth Observation data processing. CMS CERN Tier-2 EOS Namespace – using redis. Infrastructure aware scheduling. Putting /eos in production. Gradual role out and checking of performance. IHEP instance. Russian – Federated Storage usecase The EOS workflow engine

Cloud services for Synchronisation and Sharing (CS3) Sessions – Applications; Technology; Storage tech; Services; Industrial; Projects and collaborations and New site services. One recurrent issue under services – scalability

CERN Tape Archive (CTA) Evolution of CASTOR EOS plus CTA is a “drop in” replacement for CASTOR. EOS de facto disk storage for LHC physics data Natural evolution Ready for friendly small expts in mid 2018. Ready for LHC expts end 2018. Could use ENSTORE. Why build from scratch? CASTOR had the tape storage software already. Just redoing the metadata stuff

Baseline for WLCG Stratum 1 Operations Stratum 1 Network. Backbone for CernVM-FS HTTP content distribution. Current WLCG Stratum 1s: loosely coupled set of web services at 5 sites. Maintenance of S0-S1 replication – client configuraton. Typical information needs of experiments – e.g. which stratum-1 has my repo. Suggested baseline storage 20TB (50% growth/year); sync every 15 mins; latest software within 2 months; ports 80 and 80000 etc.