BEIJING-LCG2 Tire2 Grid. Outline Introduction about IHEP Grid Fabric infrastructure of BeiJing-Lcg2 Middleware grid-service and supported VO’s.

BEIJING-LCG2 Tire2 Grid

Outline Introduction about IHEP Grid

Fabric infrastructure of BeiJing-Lcg2

Middleware grid-service and supported VO’s.

LCG2 Grid-operations.

Introduction of IHEP Grid There are three grid system in IHEP1.BEIJing-LCG2(production)2.Gilda Grid3.BESIII test Grid

LCG Deployment in China

UI

CE_torque

SE_classicl

MON WN1

UI

CE_torque

SRM_dcache

BDII

PX

RB

LFC_mysql

WN2

SDU VO: ATLASSDU VO: ATLAS

PC farm

UI

CE_torque

SE_classicl

WN1

WN2

PC farm

UI

WN1

WN2

PKU VO: CMSPKU VO: CMS

CSTNET\CERNET

poolnode1

poolnode2

CNICCNICIHEPIHEP

PC farm

Disk pool

Fabric infrastructure of BeiJing-Lcg2

Fabric infrastructure of BeiJing-Lcg2 Core Server(SRM-dcache, CE, TOP-BDII, RB, LFC, Myproxy, MON, VOBOX )

IBM-X3650 (SRM-dcache, TOP-BDII,RB, LFC, Myproxy, MON, VOBOX,UI) (totally 9)

CPU intel 5130 X 2

2X2GB RAM

2X73G SAS 10k Disk Raid1

1Gb network.

HP DL380 G5 ( CE-Torque, DPM-SE, DPM-Disk,DPM-database) (totally 10)

CPU intel 5420 X 2

4X4GB RAM

2X146G SAS 10k Disk Raid1

1Gb world network. 10Gb local network

Worknodes(50 WNs)(The cpus(cores) is 408 now) Dell M1000e Blasd server CPU intel 5430X2 4X4GB RAM 1GB network

Fabric infrastructure of BeiJing-Lcg2 Storage Element(dcache) 3 head node + 2 poolnode + 4 disk array total amount of space is 40TB Storage Element(DPM) 2 head node + 8 poolnode + 16 disk array total amount of space is 160TB

Computing Element Ths cpus(cores) is 408 now.

Grid-service of Beijing-lcg2

middleware Almost all the core service han been update to

SLC47 glite-3.1. Now CE, SE_dcache, WN, UI , BDII, MON, LFC

running on slc47 and glite-3.1. RB,running on slc308 and glite-3.0, These

service will be update to new version ASAP.

Dcache Storage Element system We established a new central Dcache Storage

Element which replace the old atlasse01 and cmsse01.

New SE running on slc45 and glite-3.1. Dcache version is dcache-server-1.8.0, it

support new version of SRM v2.2. New SE have 3 headnode some poolnode. Now the total amount of space is 20TB.

Dcache Storage Element system

SE-dcache architecture IHEP

dcache-poolsGridftpGsidcap

dcap

Poolnodes

lmDomain poolManager adminDoor httpDomain

gplazmaService

seadmin.ihep.ac.cn

pnfsManager dirDomain

sepnfs.ihep.ac.cn

SRMutilityDomaininfoProvider

GridftpGsidcap

dcapXroot

Srm.ihep.ac.cn

Disk array box:1X12TB Raid 5 + Hot spare, xfs file

system.4GB FC HBA card

connected to poolnode

SE-dcache architecture

SE-dcache monitoring

SE-dcache disk usage

SE-dcache statics

PhEDEx data transfer from or to Beijing

The totale average speed is up to 15.6MB/s

PhEDEx data transfer from IN2P3 to Beijing

Main Grid-Service CA center : https://gridca.ihep.ac.cn /

Our CA has signed 241user certificate and 132 host 236 certificate

Main Grid-Service bes voms Central LFC : bes Local LFC: atlas cms biomed

Gang Chen/CC/IHEP 23/4/18 - 20

VOMS for BES experiment

Main Grid-Service Computing service and supported VOs CE : lcg002.ihep.ac.cn Support queue: - CE: lcg002.ihep.ac.cn:2119/jobmanager-lcgpbs-argo - CE: lcg002.ihep.ac.cn:2119/jobmanager-lcgpbs-atlas - CE: lcg002.ihep.ac.cn:2119/jobmanager-lcgpbs-bes - CE: lcg002.ihep.ac.cn:2119/jobmanager-lcgpbs-biomed - CE: lcg002.ihep.ac.cn:2119/jobmanager-lcgpbs-cms - CE: lcg002.ihep.ac.cn:2119/jobmanager-lcgpbs-dteam - CE: lcg002.ihep.ac.cn:2119/jobmanager-lcgpbs-euchina - CE: lcg002.ihep.ac.cn:2119/jobmanager-lcgpbs-ops - CE: lcg002.ihep.ac.cn:2119/jobmanager-lcgpbs-esr