FCPPL computing meeting

Europe/Paris
Description
Meeting to follow up projects initiated between IHEP Computing Center and CC-IN2P3 To connect to the VISIOCONFERENCE: ------------- CONFERENCE ----------------- -- Titre : FCPPL_VISIO_II_visio_III_69613 -- Date : 10 fevrier 2012 -- Heure : 10:00:00 heure locale -- Duree : 03:00:00 -- Numero a appeler : ---- IP : 193.48.95.82 ---- ISDN : +33 (0)4 26 68 73 02 ---- TEL : +33 (0)4 26 68 73 02 -- Identifiant numerique : 69613 (terminer par #) -- Code PIN : 3234 (terminer par #) ----------- WEB COMPAGNON --------------- -- streaming, document, chat, etc. ---- http://rms.in2p3.fr/compagnon.php?idConf=69613&identConf=3234_FCPPL_VISIO_II_visio_III_69613 ------------- STREAMING ----------------- -- Pour suivre cette conference en streaming (RealPlayer ou QuickTime) : ---- rtsp://193.48.95.82/conf_h263_g711u_384000_3234_FCPPL_VISIO_II_visio_III_69613 (384 Kbts/s) ---- rtsp://193.48.95.82/conf_h263_g711u_768000_3234_FCPPL_VISIO_II_visio_III_69613 (768 Kbts/s) -- Code PIN : 3234 --------------- DOCUMENTS -------------- -- http://rms.in2p3.fr/spaceConfLogin.php?conference=69613 PASSWORD : 3234

Throughput between CC and IHEP

- Many progresses done so far concerning the investigations of the low throughput rate from IHEP to CC-IN2P3. Problem area has been restricted the CSTNet internal network.
- A new RENATER 10Gb/s link dedicated to LHCONE and GRIF has been installed on the 26th of january. It should prevent bottleneck and saturation observed sometimes at this level.
-In the process of investigating low throughput between far sites and CC-IN2P3, Jérôme found a bottleneck at the level of a router. The modification of the configuration of the router has shown a clear improvement of the throughput.
-All these change result in much better performance from IHEP to CC but for the moment, preformances from CC to IHEP did not improve
-Measurements of the throughputs with ATLAS data have confirmed improvements in the rates from Beijing to CC. It shows also that they are lower on both sides to those between Tokyo and CC. 
-Perfsonar tests: Scheduled bandwidth tests between CC and IHEP are running and results are finally coming in the CC perfsonar MA (measurement archive) after a small manual configuration change (increasing a timeout with a default value which is too short).

 

Actions

  • send the slide 9 of Fabio's presentation to Xavier Jeanin to ask Renater and GEANT to investigate  the differences between european sites.

  •  see with Jerome if we can do the same kind of tests IHEP did: iperf tests between CC and different nodes on the route to IHEP
  •  Check the tuning of the FTS channels to see if they don't create a bottleneck.
  •   Modify perfsonar at IHEP to prevent timeout.
 

France-Asia VO

Actions

- Yonny will contact André Tilquin to make some tests.

 

AOB

  • Présentation at the computing session of FCPPL: Fabio to present a joint report?
  • Meeting at IHEP in April:-- will be on the 8th of April (2 days?): 
                 Need to setup the agenda 
             France-Asia VO: need to setup a working session with Yonny and the persons involved at IHEP.
                Prepare a short meeting before the visit to IHEP to finalize the agenda.
  • Dirac : An instance of DIRAC is being installed at CC-IN2P3. At IHEP, it is being installed for BESS experiment. Tutorials are regularly held at CC-IN2P3 either for experts of for experiments usage.
      

Il y a un compte-rendu associé à cet événement. Les afficher.