SPI project meeting

Tuesday 5 Feb 2008, 15:00 → 17:00 Europe/Zurich

32-SC-22 (Bldg 32)

Stefan Roiser (CERN)

The weekly meeting of the Software Process and Infrastructure project of the LCG/AA.

================================================== Yves Perrin + Actions done and progress - Savannah: . new projects approved and created . provided Manuel Gallas and David Quarrie with information concerning Savannah functionality related to items reassignment between projects and features assoociated with the use of 'categories'. An informal meeting should take place this week to discuss how to best simplify the current set of Atlas Offline projects and how to structure the support for the data taking phase. . iterated with Joachim Flammer (GLite MDW) to explain the Savannah db structure (they want to extract data for statistics) . completed installation of Savannah on my development system - Hypernews: . new fora created . helped users with lost password and subscription to fora . modify the Atlas HN 'New Membership' page to display a link to the Atlas Twiki HN FAQ page for additional info. . iterated with S. Haywood concerning terminology inconsistencies between the 'New Forum Request' form and the resulting mail sent to project admins thus leading to misunderstandings and confusion. Terry Hung will fix it. + Next actions - organize Savannah Atlas Offline projects with Manuel Gallas and David Quarrie - complete the set of LCGCMT dynamic web pages with Stefan - modify the NB summary to fit on the screen (no horiz scroll) and harmonize it with the LCG config viewer - AOB: - I will be absent this friday (Feb 8), plus the following week and probably the 21st and 22nd. ================================================== Bertrand Bellenot + Actions done - Investigated the problem of ROOT nightly build on Windows The problem has been fixed (thanks to Axel) but not fully understood. The origin of the problem was 'grep'. Running the tests by hand was working, but they were failing when called from the nightly script... ================================================== Piotr Kolet + Actions done - Nightlies maintenance - - Preparing dev3 slot before LCGCMT_54 relase. - - playing with QMTEST for windows (the problem finally fixed by Bertrand) - - fixing dblookup.xml problem on macosx. - - removing dev3 slot, preparing the same slot for LCGCMT_54-patches - - adding GAUDIATLAS as new project for dev3 - - preparing LCGCMT and nightlies for gcc 4.2 - - - linking externals for *_gcc42 - - - trying to rebuild BOOST (no success so far) - - space allocation for new builds (gcc42 and new GAUDI) - Preparing nightlies for LCG_BUILDERS - - devel version already uses pkg_get, pkg_config and pkg_make and pkg_test (only roottest) - one week in Poland on Red Hat Certified Engineer (RHCE) course with exam. - + Next actions - finishing with LCG_BUILDERS and put the changes into production. ================================================== Stefan Roiser + Actions done - Savannah - Installation of an "intermediate Certificate" which is needed since the issuing of the new versions. Changing the http.conf on savannah - LCGCMT - new interface package for mysql-python (thx to ATLAS) - a typo was found in the oracle interface package (fixed for LCG 54) - a test suite for ROOT/CASTOR plugins has been setup in LCGCMT with qmtest this is running since the beginning of the week in the nightly builds. - bug fix in the sip interface package (add python version to dir name) - External software - Boost 1.34.1 had to be rebuilt in order to make boost_python work with py 2.5 This will be part of the first patch configuration (LCG 54a) - /Grid/vdt/globus has been removed, this installation was a duplication of /Grid/globus - fastjet 2.3.0 has been installed for linux platforms - siscone 1.3.0 ( a plugin to fastjet ) has been installed (linux/mac) - valgrind 3.3.0 installed, including patches from CMS - Nightly builds - Atlas has reported several afs glitches while building against the LCG/AA nightlies. We should keep an eye on the evolution here. - Hypernews - There was a problem with the swap space on the w/e. After rebooting the machine everything was back to normal. We need to observe it via lemon. - Macphsft02 - The time between the server and the official CERN time was off by 5 min!! This prevented to receive kerberos tokens (eg for afs). Now we are using the CERN time server instead of the Apple one. - afs increase of q.sw.lcg (project scratch space) requested by 100 G (needed for nightlies). + Next actions - Mysql filling scripts for web pages !!!!!!! ==================================================

15:00 → 15:10 News

15:10 → 15:40 Round table

15:40 → 16:00 Aob