LCG Deployment Report

1
LCG Deployment Report

• LHCC closed session 26/09/2006
• Sijbrand de Jong also for
• Mario Martinez-Perez, Paul Dauncey, Silvia Dalla
  Torre
• Based on presentations of
• Markus Schulz, Maite Barroso-Lopez, Ruth Pordes
  and Torsten Antoni
• Although deployment and performance issues
  pervaded lots of talks

• gLite 3.0 Deployment
• Operation of EGEE and OSG
• User Support

2
gLite 3.0 deployment

• Merge of gLite 2.7 and gLite 1.5
• Merge of software,
• but also merge of teams (terminology)
• limited development time
• Much integration done by wrappers
• very limited ? test only few bugs uncovered
• deployment on all sites nearly in 1 go
• This was not a success dont repeat
• Many bugs and problems found and fixed
• Communication to sites not optimal
• SUCCESS Now service OK !
• Now clean up software and improve stability

3
EGEE resources
EGEE gt 190 sites, 40 countries 155 sites
certified and in production gt 28,000
processors, 26 PB storage
4
EGEE organisation
Operations Coordination Centre (OCC) Management
and oversight of all activities Regional
Operations Centres (ROC) core of support,
supporting resource centres within its
region Grid Operator on Duty (COD) Resource
centres providing computing, storage, network,
etc. Grid User Support (GGUS) At FZK,
coordinating support, single point of contact for
users
A security incident put organisation to the test
implementation of protocols
5
EGEE performance
6
EGEE monitoring
Site Functionality Test
7
EGEE monitoring
Service Availability Monitoring
8
EGEE monitoring
Service Availability Monitoring Need
more Virtual Organisation Specific Tests
9
OSG resources
Service Availability Monitoring Need
more Virtual Organisation Specific Tests
96 Resources across production integration
infrastructures
gt15,000 CPUs 6 PB MSS 4 PB disk
27 Virtual Organizations including operations and
monitoring groups
10
OSG organisation
11
OSG performance
50K 90K CPU Hours/day
ATLAS
CDF
CMS
12
User Support
Catch more at VO level
13
Summary Conclusion

• Congratulations
• gLite 3.0 deployment chaotic, but successful
• EGEE growing and well used by LHC exps
• OSG project funded, OSG well used by exps
• User Support improving
• Concerns
• Stability and reliability of service now key !
• Involve exps in User Support for EGEE
• And
• EGEE and OSG could handle more LHC jobs !