Batch System Operation & Interaction with the Grid LCG/EGEE Operations Workshop May 25 th 2005 CERN.ch.

Slides:



Advertisements
Similar presentations
Libra: An Economy driven Job Scheduling System for Clusters Jahanzeb Sherwani 1, Nosheen Ali 1, Nausheen Lotia 1, Zahra Hayat 1, Rajkumar Buyya 2 1. Lahore.
Advertisements

Pricing for Utility-driven Resource Management and Allocation in Clusters Chee Shin Yeo and Rajkumar Buyya Grid Computing and Distributed Systems (GRIDS)
CSF4 Meta-Scheduler Tutorial 1st PRAGMA Institute Zhaohui Ding or
National Institute of Advanced Industrial Science and Technology Advance Reservation-based Grid Co-allocation System Atsuko Takefusa, Hidemoto Nakada,
U.S. Department of the Interior U.S. Geological Survey Data Services Task Team Proposal Discussion at WGISS #25 February, 2008 Lyndon R. Oleson U.S. Geological.
Ch 11 Distributed Scheduling –Resource management component of a system which moves jobs around the processors to balance load and maximize overall performance.
HTCondor and the European Grid Andrew Lahiff STFC Rutherford Appleton Laboratory European HTCondor Site Admins Meeting 2014.
FP7-INFRA Enabling Grids for E-sciencE EGEE Induction Grid training for users, Institute of Physics Belgrade, Serbia Sep. 19, 2008.
4/2/2002HEP Globus Testing Request - Jae Yu x Participating in Globus Test-bed Activity for DØGrid UTA HEP group is playing a leading role in establishing.
CERN LCG Overview & Scaling challenges David Smith For LCG Deployment Group CERN HEPiX 2003, Vancouver.
Scheduling under LCG at RAL UK HEP Sysman, Manchester 11th November 2004 Steve Traylen
CoreGRID Workpackage 5 Virtual Institute on Grid Information and Monitoring Services Authorizing Grid Resource Access and Consumption Erik Elmroth, Michał.
6/2/20071 Grid Computing Sun Grid Engine (SGE) Manoj Katwal.
Enabling Grids for E-sciencE Medical image processing web portal : Requirements analysis. An almost end user point of view … H. Benoit-Cattin,
The SAM-Grid Fabric Services Gabriele Garzoglio (for the SAM-Grid team) Computing Division Fermilab.
Copyright © 2010 Platform Computing Corporation. All Rights Reserved.1 The CERN Cloud Computing Project William Lu, Ph.D. Platform Computing.
HEPiX Catania 19 th April 2002 Alan Silverman HEPiX Large Cluster SIG Report Alan Silverman 19 th April 2002 HEPiX 2002, Catania.
December 8 & 9, 2005, Austin, TX SURA Cyberinfrastructure Workshop Series: Grid Technology: The Rough Guide Configuring Resources for the Grid Jerry Perez.
Workload Management WP Status and next steps Massimo Sgaravatto INFN Padova.
Computing for ILC experiment Computing Research Center, KEK Hiroyuki Matsunaga.
The National Grid Service User Accounting System Katie Weeks Science and Technology Facilities Council.
Tier-1 Batch System Report Andrew Lahiff, Alastair Dewhurst, John Kelly, Ian Collier 5 June 2013, HEP SYSMAN.
Stephen Booth EPCC Stephen Booth GridSafe Overview.
LCG and HEPiX Ian Bird LCG Project - CERN HEPiX - FNAL 25-Oct-2002.
1 BIG FARMS AND THE GRID Job Submission and Monitoring issues ATF Meeting, 20/06/03 Sergio Andreozzi.
Computational grids and grids projects DSS,
Semantic Interoperability Berlin, 25 March 2008 Semantically Enhanced Resource Allocator Marc de Palol Jorge Ejarque, Iñigo Goiri, Ferran Julià, Jordi.
Grid Workload Management & Condor Massimo Sgaravatto INFN Padova.
Grid Workload Management Massimo Sgaravatto INFN Padova.
Ideas for a virtual analysis facility Stefano Bagnasco, INFN Torino CAF & PROOF Workshop CERN Nov 29-30, 2007.
CERN IT Department CH-1211 Genève 23 Switzerland t Tier0 Status - 1 Tier0 Status Tony Cass LCG-LHCC Referees Meeting 18 th November 2008.
Some Title from the Headrer and Footer, 19 April Overview Requirements Current Design Work in Progress.
June 24-25, 2008 Regional Grid Training, University of Belgrade, Serbia Introduction to gLite gLite Basic Services Antun Balaž SCL, Institute of Physics.
Scheduling in HPC Resource Management System: Queuing vs. Planning Matthias Hovestadt, Odej Kao, Alex Keller, and Achim Streit 2003 Job Scheduling Strategies.
LCG Pilot Jobs + glexec John Gordon, STFC-RAL GDB 7 November 2007.
HEPiX FNAL ‘02 25 th Oct 2002 Alan Silverman HEPiX Large Cluster SIG Report Alan Silverman 25 th October 2002 HEPiX 2002, FNAL.
Enabling Grids for E-sciencE SGE J. Lopez, A. Simon, E. Freire, G. Borges, K. M. Sephton All Hands Meeting Dublin, Ireland 12 Dec 2007 Batch system support.
CASTOR evolution Presentation to HEPiX 2003, Vancouver 20/10/2003 Jean-Damien Durand, CERN-IT.
APST Internals Sathish Vadhiyar. apstd daemon should be started on the local resource Opens a port to listen for apst client requests Runs on the host.
US LHC OSG Technology Roadmap May 4-5th, 2005 Welcome. Thank you to Deirdre for the arrangements.
Conference name Company name INFSOM-RI Speaker name The ETICS Job management architecture EGEE ‘08 Istanbul, September 25 th 2008 Valerio Venturi.
Proposal for a IS schema Massimo Sgaravatto INFN Padova.
Trusted Virtual Machine Images a step towards Cloud Computing for HEP? Tony Cass on behalf of the HEPiX Virtualisation Working Group October 19 th 2010.
Adrian Jackson, Stephen Booth EPCC Resource Usage Monitoring and Accounting.
Cracow Grid Workshop, October 15-17, 2007 Polish Grid Polish NGI Contribution to EGI Resource Provisioning Function Automatized Direct Communication Tomasz.
23 January 2007WLCG workshop, CERN System Management Working Group Alessandra Forti WLCG workshop CERN, 23 January 2007.
16-Nov-01D.P.Kelsey, HTASC report1 HTASC - Report to HEP-CCC David Kelsey, RAL rl.ac.uk 16 November 2001, CERN ( )
EGEE is a project funded by the European Union under contract INFSO-RI Grid accounting with GridICE Sergio Fantinel, INFN LNL/PD LCG Workshop November.
Accounting in LCG Dave Kant CCLRC, e-Science Centre.
Latest Improvements in the PROOF system Bleeding Edge Physics with Bleeding Edge Computing Fons Rademakers, Gerri Ganis, Jan Iwaszkiewicz CERN.
EGEE-II INFSO-RI Enabling Grids for E-sciencE EGEE and gLite are registered trademarks ROCs Top 5 Middleware Issues Daniele Cesini,
LCG Issues from GDB John Gordon, STFC WLCG MB meeting September 28 th 2010.
8 August 2006MB Report on Status and Progress of SC4 activities 1 MB (Snapshot) Report on Status and Progress of SC4 activities A weekly report is gathered.
WMS baseline issues in Atlas Miguel Branco Alessandro De Salvo Outline  The Atlas Production System  WMS baseline issues in Atlas.
Campus Grid Technology Derek Weitzel University of Nebraska – Lincoln Holland Computing Center (HCC) Home of the 2012 OSG AHM!
INFN/IGI contributions Federated Clouds Task Force F2F meeting November 24, 2011, Amsterdam.
DGAS Distributed Grid Accounting System INFN Workshop /05/1009, Palau Giuseppe Patania Andrea Guarise 6/18/20161.
Trusted Virtual Machine Images the HEPiX Point of View Tony Cass October 21 st 2011.
Enabling Grids for E-sciencE Claudio Cherubino INFN DGAS (Distributed Grid Accounting System)
DIRAC: Workload Management System Garonne Vincent, Tsaregorodtsev Andrei, Centre de Physique des Particules de Marseille Stockes-rees Ian, University of.
Jean-Philippe Baud, IT-GD, CERN November 2007
Connecting LRMS to GRMS
EGEE Middleware Activities Overview
How to connect your DG to EDGeS? Zoltán Farkas, MTA SZTAKI
EGEE VO Management.
CREAM-CE/HTCondor site
Artem Trunov and EKP team EPK – Uni Karlsruhe
Developments in Batch and the Grid
WLCG Collaboration Workshop;
Ian Bird LCG Project - CERN HEPiX - FNAL 25-Oct-2002
Presentation transcript:

Batch System Operation & Interaction with the Grid LCG/EGEE Operations Workshop May 25 th 2005 CERN.ch

CERN.ch 2 Why a Batch Workshop at HEPiX? Proposed after the last Operations Workshop. Remember the complaints then? –ETT doesnt work –ETT is meaningless when fairsharing is in place –The solution of a queue per VO while easy to implement now but is not a good or long term solution. –The [ETT] algorithm was questioned and other proposals were given. Idea was to bring together site managers, grid & local scheduler developers.

CERN.ch 3 Workshop Aims Understand how different batch scheduling systems are used at HEP sites –Are there any commonalities? How do sites see the Grid interface? How would sites like to see the Grid interface? What is the impact of the current interface? How do developers of local and Grid level schedulers see the future? How/can HEP site managers influence future developments? Well attended (70-80) –Definite interest in this area from site managers See

CERN.ch 4 Agenda Local Scheduler usage –SLAC, RAL, LeSC, JLab, IN2P3, FNAL, DESY, CERN, BNL –LSF, PBS, Torque/Maui, SGE (N1GE6), BQS, Condor Impact of Grid on sites –Jeff Templon overview (c.f. previous talk), Local scheduler view –LSF, PBS, LoadLeveler, Condor, BQS Grid Developments –EGEE/BLAHP, GLUE Common batch environment –See earlier.

CERN.ch 5 Site Presentations --- I Site reports covered –Brief overview of the available computing resources, showing (in)homogeneity of resources –Queue configuration---what and why –How do users select queues---cpu time alone or specifying other resources (e.g. memory, local disk space availability) –Need for, and use of, "special" queues---for "production managers", sudden high priority work, other reasons. »Question from LHCC referee: If there is some urgent analysis, how can [gLite] send this to a special queue? –Level of resource utilisation

CERN.ch 6 Site Presentations --- II Overall, configurations and concerns were broadly equivalent across sites. Concerns were around –Scheduling –Security –Interface Scalability Cover these issues in next few slides.

Scheduling Issues

CERN.ch 8 Local Load Scheduling: summary Batch schedulers at local sites enable fine-grained control over heterogeneous systems and are used to enforce local policies on resource allocation and provide SLA for users (turnround time). –Large sites have subdivision of user groups Scheduling is by CPU time, some need to request –minimum CPU capacity for server –memory requirement –available disk work space (/pool, /scratch, /tmp) Sites want Grid interface to use existing queue(s) –NOT to create a queue per VO. –EMPHATICALLY NOT to replicate queue structure per VO

CERN.ch 9 Grid/Local interface problems Jeffs presentation! In short –Not enough information passed from the site to the Grid –No information passed from the Grid to the site Result: –Queues at sites whilst others sit empty –Confused/frustrated site managers –Inefficient behaviour as people work the system »Tragedy of the commons

CERN.ch 10 Should sites (be able to) enforce policies? Sites are funded for particular tasks and need to show funding agencies and users that they are fulfilling their mission. This is a Grid. Why does it matter if you are running jobs for X not Y? Y may be happily running jobs at another site. My view: –Sites need to understand and feel comfortable with the way they accept jobs from the Grid. –If they are comfortable, account may be taken of global activity when setting local priorities. –Lets walk before we try to run…

CERN.ch 11 Can/Should we fix this? … or should we wait to see some general standard emerge? Strong support from commercial people (especially Platform and Sun) for HEP to work out solutions to this problem. –They are interested in what we do. Standards bodies (GGF,…) wont come up with any common solution soon. –But this doesnt mean HEP shouldnt participate »Raise profile of problems of interest to us »Give practical input based on realworld experience.

CERN.ch 12 How to fix? Improve information available to Grid scheduler –VO information added in GLUE schema (v1.2) »Need volunteer per batch system to maintain dynamic plug-ins and the job manager. u CERN will do this for LSF. Need other volunteers! –but still assumption of homogeneous resources at a site. –There is a plan to start work on GLUE v2 in November »No requirement for backwards compatibility. »Discussion should start NOW! u But need to assess impact of v1.2 changes before rushing into anything. Grid scheduler should pass job resource requirements to the local resource manager. –Not yet. When? How? –Needs normalisation… Does this need to be per VO?

Security

CERN.ch 14 Security Issues Sites are still VERY concerned about traceability of users. Mechanisms seem to be in place to allow this, but sites have little practical experience. –c.f. delays for CERN to block user systematically crashing worker nodes. –Security group have doubts that sites are fulfilling obligations in terms of log retention. –Security Challenges mooted; these may help increase confidence… Whatever, it does NOT seem to be a good idea to have a portal handling user job requests and passing these on with a common certificate…

Interface Scalability

CERN.ch 16 Interface Scalability IN2P3 example: GridJobManager asks job status once per minute (even for 15-hour jobs). –5000 queued jobs running jobs = 100 queries/s Being solved by egee BLAHP –Caches query response But… –further example need for discussion between sites & developers (IN2P3 fixing this issue independently) –are there other similar issues out there? »c.f. LSF targets: u Scalability: 5K hosts, 500K active jobs, 100 concurrent users, 1M completed jobs per day u Performance: >90% slot utilistion, 5s max command response time, 4kB memory/job, master failover in <5mins »What are targets for the CE? RB?

Some other Topics

CERN.ch 18 End-to-End Guarantees The Condor talk raised many interesting points. One in particular was the (in)ability of the overall system to offer end-to-end execution guarantees to the users. Condor glide-in: pilot job submitted via the Grid which takes a job from a condor queue. Fair enough [modulo security…] for system managers PROVIDED pilot job expresses same resource requests as it advertises in a class-ad when it starts. –Shouldnt claim to be maximum possible length then run short job. –Class ads and GLUE schema not so different: Both are ways of saying what a node/site can do in a way that can be used to express (and then match) requirements.

CERN.ch 19 Pre-emption & Virtualisation Strong message from batch system developers that pre-emption is A GOOD THING. With pre-emption schedulers can maximise throughput/resource usage by –suspending many jobs to allow parallel job to run –suspending long running jobs to provide quick turnround for priority jobs. Interest in virtualisation as method to ease this –Also discussed at last operations workshop as a way to ease handling of multiple (conflicting) requirements for OS versions. –Something to watch. How would (pre-empted) users like this? –No guarantee of time to completion once job starts…

CERN.ch 20 Push vs Pull A false dichotomy –Sites can manipulate pull model to create a local queue Real issue is early vs. late allocation of task to resource –Early: site resource utilisation maximised: a free cpu resource can be filled immediately with a job from the local queue –Late: user doesnt see job sent to site A just before a cpu becomes free at site B. Questions: –Long term, will most cpu resources be full? –What do people want to maximise? Throughput or ? »Efficient scheduling important anyway… transparency of grid/local interface will be key. –Pre-emption, anyone?

Conclusion

Summary

CERN.ch 23 Workshop Summary Useful workshop. [IMHO…] Good that there has been progress since the November workshop at CERN (GLUE schema update), but much is still to be done.

The Service is the Challenge

CERN.ch 25 Workshop Summary Useful workshop. Good that there has been progress since the November workshop at CERN (GLUE schema update), but much is still to be done. [Still] Need to increase dialogue between site managers and Grid [scheduler] developers –Site managers know a lot about running services. –Unfortunate that a meeting change created a clash and reduced scope for egee developers to participate in Kaelsruhe discussions. –A smaller session is pencilled in for HEPiX in SLAC, October 10 th – 14 th. More dialogue then? Not too early to start thinking about GLUE v2!