PASTE meeting

Tuesday 21 Aug 2007, 10:00 → 12:00 Europe/Zurich

Andrei Tsaregorodtsev

CERN, bat 2 R-030 To join the telephone conference, click here: https://audioconf.cern.ch/call/0122882 To dial in to the conference: a. Dial +41 22 7676000 b. Enter access code 0122882

Production and PASTE Meeting, 21st August 2007, CERN Chair: A.Tsaregorodtsev Minutes: S. Paterson CERN: M.Bargiotti, N.Brook, G.Castellani, P.Charpentier, A.Smith, S.Paterson, R.Santinelli Phone: R.Bernet, R.Graciani, U.Marconi, R.Nandakumar, M.A.S.Miguelez, A.Tsaregorodtsev Apologies: J.Closier 1. Production Status Report Minutes from the production meeting are available here. Raja reported that a new record of 1200 concurrent reconstruction jobs has been achieved (CERN / RAL). During the ramp-up of the DC06 restart an increasing number of stalled jobs have been observed at RAL (~600). This triggered some discussion on the merits of downloading files local to the WN. Philippe conveyed that the advice from Castor people is not to bring files local. Each Castor server can support ~300 paralllel connections and this should be sufficient for our jobs. The dCache settings should be investigated at RAL to understand where the limitations exist and what they are. ( ACTION: Raja ) Since the recent problems are also relevant for GridKa and NIKHEF it would be advantageous to involve site representatives once the RAL case is understood. IN2P3 should be back from maintenance this evening. The dCache upgrade fixes the CERN CA incompatibility with the gsidcap door. Testing is required before reentering into the Analysis mask. ( ACTION: Stuart ) Philippe contacted the physics working groups to determine which of the b-inclusive samples were eligible for reconstruction and Boole v12r1 data are OK. Nick suggested creating a new reconstruction production for this data so as not to interfere with the currently running productions. Raja agreed to define the new production and add a new transformation for this. ( ACTION: Raja ) The bookkeeping issue is now resolved, Marianne updated the views by hand and mentioned that a new version of the AMGA client has been installed. Unfortunately the files still do not appear in the views so Marianne is waiting for some advice from Carmine. ( ACTION: Marianne ) 2. Production Operations Issues - DIRAC v2r15p4 release DIRAC v2r15p4 contains some fixes from Andrei for the current transfer logic and will be deployed this week. Andrei raised the issue of using lcg-cp in favour of lhcb-lcg-cp on the Worker Nodes. Ricardo mentioned that this still should be kept on VO-boxes since the short timeout is not present in the standard lcg-cp. This lhcb-lcg-cp feature currently protects the VO-boxes against broken connections during transfer requests between two SRM endpoints. Roberto suggested a feature request should be made to the LCG Utilities developers. ( ACTION: Roberto ) Philippe suggested the use of well-controlled versions of all the Grid middleware clients but until this is in place we can add a temporary workaround for the VO-boxes which will create an alias to lhcb-lcg-cp for now. To cope with the current lcg-cp release, edg-gridftp-mkdir should be used to create any directories in advance of calling lcg-cp. This release can be rolled out and should work for both SLC3 and SLC4. This should be top priority as there is no longer SLC3 at PIC and GridKa. ( ACTION: Andrei ) - gLite 3.1 WMS usage Gianluca agreed to begin tests with production and user jobs prior to CHEP. ( ACTION: Gianluca ) Stuart raised the issue of VOMS proxy testing for the production managers. Roberto suggested publishing results to the SAM DB from production manager test jobs to the Tier-1s. The idea of allowing the production managers should be discussed with the SAM developers. ( ACTION: Roberto, Stuart ) - SRM2.2 tests progress Andrew gave a summary of progress with the GFAL python binding and indicated that there had not been much change from last week. The Tier-1's are still actively being debugged but the show-stoppers for LHCb will be compiled and conveyed on a Twiki page for the GFAL developers in the mean-time. ( ACTION: Andrew ) Philippe mentioned that Castor has no concept of a lifetime for a staged file however it is possible to flag a priority for garbage collection. The implementation of pinning for dCache is yet to be understood. 3. CHEP Contributions Discussion The CHEP contributions page is available here. The contributions discussion was postponed until after lunch. 4. AOB Andrei received a response from the CHEP07 organisers that the DIRAC posters had been arranged together.

10:00 → 10:15 Production status report

- Report from the Production meeting:

10:15 → 10:45 Production operations issues

- DIRAC v2r15p4 release - gLite 3.1 WMS usage - SRM2.2 tests progress

11:05 → 11:35 CHEP presentations discussion

CHEP abstracts page http://indico.cern.ch/conferenceDisplay.py?confId=12665

11:35 → 11:45 AOB