A Cristofori1, E Fattibene1, L Gaido2, P Veronesi1 - INFN-CNAF Bologna (Italy)1, INFN-Torino Torino (Italy)2 - hlrmon@lists.infn.it- https://dgas.cnaf.infn.it/hlrmon/ Distributed Grid Accounting System HLRmon accounting portal The accounting layout Architecture DGAS (Distributed Grid Accounting System) sensors collect accounting information at site level Site data are sent to site or multi-site HLRs (Home Location Registers) A top level HLR receives accounting data from all the Grid HLRs HLRmon retrieves data from the top level HLR and presents them through a web interface Aggregated data are also sent to the central EGI Accounting Portal HLRmon is successfully deployed as accounting portal for IGI and other Grid projects/NGIs (National Grid Initiatives) using DGAS as accounting system Collectors retrieve:  aggregated CPU accounting information from a desired HLR  storage accounting data by probes running on the SRM instances at WLCG sites  values of some GLUE attributes about the computing power capacity of the WLCG sites An aggregation layer is responsible for the data manipulation to produce charts and tabular reports visualized by the graphical interface Data aggregated per Grid user are directly retrieved from the HLR database at run-time and are just accessible by authorized users General purpose CPU reports Different metrics to analyze: • CPU and Wall Clock Time, expressed in days and normalized on the mean power capacity of the site • Number of executed jobs • Job efficiency in terms of (CPU Time / Wall Time) ratio An advanced selection menu permits to choose a desired subgroup of values for each aggregation key A specific view with data grouped per discipline is provided: the mapping information between VOs and disciplines is currently retrieved from the EGI Operations Portal Different keys of data aggregation: • Site • Virtual Organization • VOMS group and role • Certification Authority and Registration Authority • Job type (Grid or local) • Grid user DN WLCG reports Storage reports The charts allow the analysis of CPU and storage resources usage at the WLCG Tier1 and Tier2 by the LHC experiments, over the last month and the last year The resource usage and the amount of pledged resources are compared Accounting data are grouped by site and by VO Storage data are taken by a script running on the sites and sent to HLRmon via email Non-standard schema is used to generate records (no standard schema exists) The information on the used space is collected from the BDII once a day Storage URs (Usage Records) are stored on the HLR server as the computing URs HLRmon is able to retrieve storage information directly from the desired HLR The portal shows the daily disk consumption in terms of used space Data are aggregated per site, SE and Storage Area (SA) Used space will be compared to the total available space (at SA, SE and site level) Used space per site Used space per site per day Used space per SA Used per SA per day