NGI France-Grilles: Infrastructure evolution H. Cordier
Operations and infrastructure pole Operations organization Infrastructure: -Accounting -Monitoring -Certification/Deployment -Certification Authority -Operations portal EGI (incl. EGI Operations Dashboard)
Infrastructure Pole Accounting Monitoring Certification/Deployment
Architecture Globale 4 All regions to publish via ActiveMQ Provide regional accounting system Allow regions to continue using central system Collect records from other grids/tools via same interface Source: © Cristina Del Cano Novales Currently – EGI Project
APEL-DB SITE 1 SITE 2 SITE 3 APEL-Publisher AMQ Broker SITE 1 Apel-Parser SITE 2 Apel-Parser SITE 3 Apel-Parser PORTAIL d'ACCOUNTING – NGI FRANCE SITE non EGI SITE 4 Parser Currently–NGI_France
Currently/Forecasts Now : From 3 types to 1 type of sites publishing CC : Direct publication at RAL Sites using CC’s monbox: Auvergrid/LPC/SUBATECH/IPNL/IRES/LPSC Sites using their own monbox Publication to EGI : From GRIF Filtering of relevant sites, Mutualization of glite-APEL /1 GRIF Transition - Data storage and backup Participation in NGI task forces for a reliable accounting
Monitoring Accounting Monitoring Certification/Deployment
National Monitoring since June 2010 MyEGI set-up On line documentation for sites-admins: Tests to turn critical on dec 1rst : bdii.freshness Procedure in progress Cream-CE into availability/Reliability algorithm Algorithm alteration Currently – NGI_FRANCE
Systematic registration of incidents/problems: # – Introduction of arbitrary probes # – Nagios box redundancy # – Close SE and CE Av/Rel computations Liste_des _diff.C3.A9rentes_Requetes Under study : Several Nagios boxes set-up updates, certification, redundancy, security, VO specific, « Supra » VO, National VO
Accounting Monitoring Certification/Deployment
Currently First Meeting on Nov 9th infrastructure: 1 WMS/LB, 1 site BDII, 1 TopBDII mission 1: nodes certification tests on-demand and monitoring mission 2 : deployment /Quattor template 2 infrastructures, 2 roles Roadmap : certification- Provide Test nodes Test m/w components before deployment SupportSupport for deployment at sites completeOrganize a complete & reliable infrastructure : (co-administration, tech & proc checks, communication channels) VM implementation & Quattor Server
Perspectives: From last workshop on National VO Set- up Interface between Operations/Users Set- up working groups roadmaps and specific procedures Set- up metrics