SEE-GRID-2 WP3 Internal Meeting

14 Apr 2008, 08:00 → 16 Apr 2008, 18:00 Europe/Zurich

1/1-025 (CERN)

WP3 internal meeting trip report During my visit to the Institute of Physics Belgrade, the most of the time I spent with WP3 leader (Antun Balaz) preparing WP3 material for SEE-GRID-2 final review, which included: - Analysis of the SEE-GRID-2 infrastructure availability - using the data from BBmSAM portal, we calculated availabilities of all CEs and SEs in the regional infrastructure. These data were organized then as an excel file and presented graphically. Also, we produced graphs that compare current site availabilities with previous availability data. The main conclusion is that the monitoring and enforcement of conformance of sites to the SLA criteria improved substantially the availability of Grid services. This is evident by the increase of SEE- GRID-2 sites that fully conform to the SLA criteria (close to 70% of sites for both CEs and SEs), as well as by the steady decrease in the number of sites that are still below the standards set by the SLA. - Analysis of assessment of the SEE-GRID-2 infrastructure usage per VO, per scientific discipline and per country - using the data from SEE-GRID-2 accounting portal we calculated accounting usage per site, per country and per scientific discipline, as well as the total infrastructure usage. These data were organized as an excel file and presented graphically. In summary it resulted in more than 6.2 million CPU hours of accounting to all VOs supported in the SEE-GRID-2 infrastructure. - Analysis of the operational support SEE-GRID-2 infrastructure - using the data from SEE-GRID-2 helpdesk we calculated statistics on the number of trouble tickets per SEE country and per support group according to the resolution time, as well as according to the ticket impact. These data were organized as an excel file and presented graphically. In summary more than 1/3 of all assigned tickets is resolved within 3 days, while during two weeks more than 2/3 of all tickets are resolved. There are some tickets that require longer period of time for their resolution, due to problems related to hardware upgrades, and network connectivity problems. - Analysis of deployed SEE-GRID-2 infrastructure - using the data from GStat and HGSM portals we calculated time evolution of the number of sites, CPUs and TBs during SEE-GRID-2 project. The data were organized as an excel file and presented graphically. In summary the number of sites has grown in each country, and currently almost all institutions participating in the project (partners, 3rd parties) provide Grid site for the SEE-GRID-2 infrastructure. The only exception is Turkey, where initially large number of SEE-GRID-2 sites was decreased in the first year of the project, as they were migrated to EGEE. However, this did not result in decreasing the available resources to SEE-GRID-2 infrastructure, since all committed resources were provided by the sites that remained in the regional infrastructure. - Analysis of the usage and availability of SEE-GRID-2 core services (VOMS, BDII, RB, WMS, R-GMA, LFC, MyProxy and FTS). The data were organized as an excel file and presented graphically. Information was provided by sites that are hosting services. Detailed results of the analysis will be presented at SEE-GRID-2 final review (May, 2008 Athens) as a part of WP3 presentation.