Grid Production Experience in the ATLAS Experiment

1
Grid Production Experience in the ATLAS Experiment

• Kaushik De
• University of Texas at Arlington
• BNL Technology Meeting
• March 29, 2004

2
ATLAS Data Challenges

• Original Goals (Nov 15, 2001)
• Test computing model, its software, its data
  model, and to ensure the correctness of the
  technical choices to be made
• Data Challenges should be executed at the
  prototype Tier centres
• Data challenges will be used as input for a
  Computing Technical Design Report due by the end
  of 2003 (?) and for preparing a MoU
• Current Status
• Goals are evolving as we gain experience
• Computing TDR end of 2004
• DCs are yearly sequence of increasing scale
  complexity
• DC0 and DC1 (completed)
• DC2 (2004), DC3, and DC4 planned
• Grid deployment and testing is major part of DCs

3
ATLAS DC1 July 2002-April 2003Goals Produce
the data needed for the HLT TDR Get
as many ATLAS institutes involved as
possibleWorldwide collaborative
activityParticipation 56 Institutes

• Australia
• Austria
• Canada
• CERN
• China
• Czech Republic
• Denmark
• France
• Germany
• Greece
• Israel

• Italy
• Japan
• Norway
• Poland
• Russia
• Spain
• Sweden
• Taiwan
• UK
• USA
• using Grid

4
DC1 Statistics (G. Poulard, July 2003)
5
U.S. ATLAS DC1 Data Production

• Year long process, Summer 2002-2003
• Played 2nd largest role in ATLAS DC1
• Exercised both farm and grid based production
• 10 U.S. sites participating
• Tier 1 BNL, Tier 2 prototypes BU, IU/UC, Grid
  Testbed sites ANL, LBNL, UM, OU, SMU, UTA (UNM
  UTPA will join for DC2)
• Generated 2 million fully simulated, piled-up
  and reconstructed events
• U.S. was largest grid-based DC1 data producer in
  ATLAS
• Data used for HLT TDR, Athens physics workshop,
  reconstruction software tests...

6
U.S. ATLAS Grid Testbed

• BNL - U.S. Tier 1, 2000 nodes, 5 for ATLAS, 10
  TB, HPSS through Magda
• LBNL - pdsf cluster, 400 nodes, 5 for ATLAS
  (more if idle 10-15 used), 1TB
• Boston U. - prototype Tier 2, 64 nodes
• Indiana U. - prototype Tier 2, 64 nodes
• UT Arlington - new 200 cpus, 50 TB
• Oklahoma U. - OSCER facility
• U. Michigan - test nodes
• ANL - test nodes, JAZZ cluster
• SMU - 6 production nodes
• UNM - Los Lobos cluster
• U. Chicago - test nodes

7
U.S. Production Summary

• Exercised both farm and grid based production
• Valuable large scale grid based production
  experience

Total 30 CPU YEARS delivered to DC1 from
U.S. Total produced file size 20TB on HPSS
tape system, 10TB on disk. Black - majority
grid produced, Blue - majority farm produced
8
Grid Production Statistics
These are examples of some datasets produced on
the Grid. Many other large samples were
produced, especially at BNL using batch.
9
DC1 Production Systems

• Local batch systems - bulk of production
• GRAT - grid scripts, generated 50k files
  produced in U.S.
• NorduGrid - grid system, 10k files in Nordic
  countries
• AtCom - GUI, 10k files at CERN (mostly batch)
• GCE - Chimera based, 1k files produced
• GRAPPA - interactive GUI for individual user
• EDG/LCG - test files only
• systems I forgot
• More systems coming for DC2
• Windmill
• GANGA
• DIAL

10
GRAT Software

• GRid Applications Toolkit
• developed by KD, Horst Severini, Mark Sosebee,
  and students
• Based on Globus, Magda MySQL
• Shell Python scripts, modular design
• Rapid development platform
• Quickly develop packages as needed by DC
• Physics simulation (GEANT/ATLSIM)
• Pileup production data management
• Reconstruction
• Test grid middleware, test grid performance
• Modules can be easily enhanced or replaced, e.g.
  EDG resource broker, Chimera, replica catalogue
  (in progress)

11
GRAT Execution Model
1. Resource Discovery 2. Partition
Selection 3. Job Creation 4. Pre-staging 5.
Batch Submission 6. Job Parameterization
7. Simulation 8. Post-staging 9.
Cataloging 10. Monitoring
12
Databases used in GRAT

• Production database
• define logical job parameters filenames
• track job status, updated periodically by scripts
• Data management (Magda)
• file registration/catalogue
• grid based file transfers
• Virtual Data Catalogue
• simulation job definition
• job parameters, random numbers
• Metadata catalogue (AMI)
• post-production summary information
• data provenance

13
U.S. Middleware Evolution
Globus
Used for 95 of DC1 production
Condor-G
Used successfully for simulation
Used successfully for simulation (complex pile-up
workflow not yet)
DAGMan
Tested for simulation, used for all grid-based
reconstruction
Chimera
LCG
14
DC1 Production Experience

• Grid paradigm works, using Globus
• Opportunistic use of existing resources, run
  anywhere, from anywhere, by anyone...
• Successfully exercised grid middleware with
  increasingly complex tasks
• Simulation create physics data from pre-defined
  parameters and input files, CPU intensive
• Pile-up mix 2500 min-bias data files into
  physics simulation files, data intensive
• Reconstruction data intensive, multiple passes
• Data tracking multiple steps, one -gt many -gt
  many more mappings

15
Grid Quality of Service

• Anything that can go wrong, WILL go wrong
• During a 18 day run, every system died at least
  once
• Local experts were not always be accessible
• Examples scheduling machines died 5 times
  (thrice power failure, twice system hung),
  Network outages multiple times, Gatekeeper died
  at every site at least 2-3 times
• All three databases died at least once!
• Scheduled maintenance - HPSS, Magda server, LBNL
  hardware...
• Poor cleanup, lack of fault tolerance in Globus
• These outages should be expected on the grid -
  software design must be robust
• We managed gt 100 files/day (80 efficiency) in
  spite of these problems!

16
Software Issues

• ATLAS software distribution worked well for DC1
  farm production, but not well suited for grid
  production
• No integration of databases - caused many
  problems
• Magda AMI very useful - but we are missing data
  management tool for truly distributed production
• Required a lot of people to run production in the
  U.S., especially with so many sites on both grid
  and farm
• Startup of grid production slow - but learned
  useful lessons
• Software releases were often late - leading to
  chaotic last minute rush to finish production

17
New Production System for DC2

• Goals
• Automated data production system for all ATLAS
  facilities
• Common database for all production - Oracle
  currently
• Common supervisor run by all facilities/managers
  - Windmill
• Common data management system - Don Quichote
• Executors developed by middleware experts
  (Capone, LCG, NorduGrid, batch systems,
  CanadaGrid...)
• Final verification of data done by supervisor

18
Windmill - Supervisor

• Supervisor development/U.S. DC production team
• UTA Kaushik De, Mark Sosebee, Nurcan Ozturk
  students
• BNL Wensheng Deng, Rich Baker
• OU Horst Severini
• ANL Ed May
• Windmill web page
• http//www-hep.uta.edu/windmill
• Windmill status
• version 0.5 released February 23
• includes complete library of xml messages between
  agents
• includes sample executors for local, pbs and web
  services
• can run on any Linux machine with Python 2.2
• development continuing - Oracle production DB,
  DMS, new schema

19
Windmill Messaging

• All messaging is XML based
• Agents communicate using Jabber (open chat)
  protocol
• Agents have same command line interface - GUI in
  future
• Agents web server can run at same or different
  locations
• Executor accesses grid directly and/or thru web
  services

20
Intelligent Agents

• Supervisor/executor are intelligent communication
  agents
• uses Jabber open source instant messaging
  framework
• Jabber server routes XMPP messages - acts as XML
  data switch
• reliable p2p asynchronous message delivery
  through firewalls
• built in support for dynamic directory,
  discovery, presence
• extensible - we can add monitoring, debugging
  agents easily
• provides chat capability for free -
  collaboration among operators
• Jabber grid proxy under development (LBNL -
  Agarwal)

21
Core XML Messages

• numJobsWanted
• supervisor-executor negotiation of number of jobs
  to process
• executeJobs
• supervisor sends XML based job definitions
• getExecutorData
• job acceptance, handle exchange (supports
  stateless executors)
• getStatus
• polling of job status
• fixJobs
• post-reprocessing and cleanup
• killJob
• forced job abort

22
Core Windmill Libraries

• interact.py - command line interface library
• agents.py - common intelligent agent library
• xmlkit.py - xml creation (generic) and parsing
  library
• messages.py - xml message creation (specific)
• proddb.py - production database methods for
  oracle, mysql, local, dummy, and possibly other
  options
• supervise.py - supervisor methods to drive
  production
• execute.py - executor methods to run facilities

23
Capone Executor

• Various executors are being developed
• Capone - U.S. VDT executor by U. of Chicago and
  Argonne
• Lexor - LCG executor mostly by Italian groups
• NorduGrid, batch (Munich), Canadian,
  Australian(?)
• Capone is based on GCE (Grid Computing
  Environment)
• (VDT Client/Server, Chimera, Pegasus, Condor,
  Globus)
• Status
• Python module
• Process thread for each job
• Archive of managed jobs
• Job management
• Grid monitoring
• Aware of key parameters (e.g. available CPUs,
  jobs running)

24
Capone Architecture

• Message interface
• Web Service
• Jabber
• Translation level
• Windmill
• CPE (Capone Process Engine)
• Processes
• Grid
• Stub
• DonQuixote
• from Marco Mambelli

25
Capone Processing
executeJob
received

• Supervisor requests
• Jobs
• Different job statuses
• 'received','translate','DAXgen','RLSreg','scheduli
  ng','cDAGgen','submission','running',
  'checking','stageOut','cleaning','end','killing
• completed/failed (each step)
• Completion/failure (job)

recovery

stageOut
stageOut
recovery
end
fixJob
26
Windmill Screenshots
27
(No Transcript)
28
Web Services Example
29
Conclusion

• Data Challenges are important for ATLAS software
  and computing infrastructure readiness
• Grids will be the default testbed for DC2
• U.S. playing a major role in DC2 planning
  production
• 12 U.S. sites ready to participate in DC2
• Major U.S. role in production software
  development
• Test of new grid production system imminent
• Physics analysis will be emphasis of DC2 - new
  experience
• Stay tuned