110 likes | 195 Views
LcgCAF:CDF submission portal to LCG. Federica Fanzago for CDF-Italian Computing Group Gabriele Compostella, Francesco Delli Paoli, Donatella Lucchesi, Daniel Jeans, Subir Sarkar and Igor Sfiligoi. Head Node. CafExe MyJob.sh MyJob.exe. CDF Data Handling Model.
E N D
LcgCAF:CDF submission portal to LCG Federica Fanzago for CDF-Italian Computing Group Gabriele Compostella, Francesco Delli Paoli, Donatella Lucchesi, Daniel Jeans, Subir Sarkar and Igor Sfiligoi
Head Node CafExe MyJob.sh MyJob.exe CDF Data Handling Model • Fermilab: host data and a farm used mainly to • analyze data, the CAF Central Analysis Facility • Remote sites: have a decentralized CAF (dCAF) where Monte Carlo are produced and later copied to FNAL, some sites (ex. CNAF) has data • CAF and dCAF are CDF dedicated farms structured: CAF User Desktop Monitor Worker nodes Submitter Developed by: H. Shih-Chieh, E. Lipeles, M. Neubauer, I. Sfiligoi, F. Wuerthwein Mailer CDF Italian Computing Group
Moving to a GRID Computing Model • Reasons to move to a GRID computing model • Need to expand resources, luminosity expect to increase by factor 8 until CDF ends • Resource were in dedicated pools, limited expansion and need to be maintained by CDF persons • GRID resources can be fully exploited before LHC experiments will start analyze data • CDF currently access in Europe more than 5000 KSpecInt2000! No hope of having so many resources without GRID CDF Italian Computing Group
User Desktop job job … output output LcgCAF: General Architecture GRID site CE and worker nodes CDF-UI CDF user must have access to CAF clients and Kerberos V ticket CDF-SE Robotic tape Storage CDF Italian Computing Group
CDF-UI Job Submission and Execution WN Monitor … Job WMS WN Submitter Mailer • Monitor: provide job information to user • Submitter: create JDL. Retrieve user tarball and make it available to WN • Mailer: send mail to user when job is over CDF Italian Computing Group
Interactive Job Monitor: CafMon kill CafMon tail CafMon dir CafMon log CafMon top CafMon ps Job Monitor CDF-UI Job information WMS Information System WN information WN WN: job wrapper collects information like load, dir, ps, top WMS : keeps track of job information CDF Italian Computing Group
CDF Code Distribution Lcg site: CNAF Proxy Cache … … /home/cdfsoft /home/cdfsoft Lcg site: Lyon Proxy Cache • No CDF code or Run Condition Database need to be available on WN • On demand Parrot translates local warning in HTT requests towards Fermilab code server and DB • To be scalable when a site has many equal requests the information are cached near WN using squid CDF Italian Computing Group
… … CDF Output storage gridftp Job output Worker Nodes CDF-SE KRB5 rcp CDF-Storage • From CDF-SE or CDF storage output copied using gridftp to FNAL fileservers. • An automatic procedure copies files from fileservers to tape and declare them to the catalogue CDF Italian Computing Group
#segments succeeded # segments submitted e = Performances CDF B Monte Carlo production GRID failures: site miss-configuration, temporary authentication problems, WMS overload LcgCAF failures: Parrot/Squid cache stacked Retry mechanism: necessary to achieve ~100% efficiency WMS retry for GRID failures and LcgCAF “home-made” retry based on logs parsing CDF Italian Computing Group
Future developments An interface between SAM (CDF catalogue) and GRID • services is under development. This will allow: • File declaration into the catalogue directly from LcgCAF • Automatic transfer of files from LCG sites to FNAL tape • Possibility of data analysis on LCG sites by single user CDF Italian Computing Group
Conclusions • CDF has developed a portal to access the LCG • resources • Currently CDF access the following sites: • CDF has produced and is producing Monte Carlo data exploiting these resources demonstrating that GRID can serve also a running experiment CDF Italian Computing Group