LHCb Core Software Meeting

Date and Location

10:30 - 11:30
CERN (2-R-030)


Ben Couturier, Gerhard Raven (Vidyo), Illya Shapoval, Joel Closier, Liang Sun, Marco Cattaneo, Marco Clemencic (minutes), Victoria Santiago Alcala



  • LIM
    • Marco Cl. suggested a separation of LCG, Generators and Middleware builds (according to the deployment cycles)

Round Table

Marco Cl.

  • Stripping cache deployment
    • Chris managed to make it work. I'll review the changes and we will commission it.
  • Brunel functors cache
    • to be prepared
  • Gaudi gitlab workflow
    • created (with help from Ben) a nightly build slot to be run on demand to test merge requests. For the moment it does not really test the merge, but only the incoming branch (to test the merge I need to commission the new nightly build scripts)
  • Issue with gcc versions in production (bug of CMT)
    • it was discussed to have environment isolation between LHCbDirac and the applications, but it can be very tricky to implement it properly.
    • for the moment the best approach is to add a workaround in SetupProject

Marco Ca.

  • Unfortunate convergence of changes that caused severe problems in production, just before most of the team went in vacations
    • we should be more conservative and careful also for transaprent changes


  • Progress on the new type of scheduling for GaudiHive (to be presented at the Concurrecy Forum)


  • "do0" platform does not work with CMT
    • not sure if it is worth fixing it
  • EOS_HOME not set properly in LbLogin
  • The problem between gcc 48/49 highlighted that we are not testing AVX
    • discussed with Marco Cl. to split the SSE and AVX tests (run only when possible) and have proper report of skipped tests in the dashboard.

Marco Cl.: we also need to update CMake (>=3.0) to properly report the skipped tests.


  • lxbuild119 will be retired on the 15th of October
    • we should prepare for the migration using aliases for the services we need to split
    • a lot of documentation refer only to buildlhcb: to be updated
  • There was an issue with the tag collector login
    • the JS code was looking for a Shibboleth cookie that is not visible anymore (work around in place)
  • gfal2 Python module is linked against Boost (which is causing troubles)
    • we need to sit with Alejandro to converge on a portable build


  • Applied some global tag
  • Clean up some mess in Oracle due to a glitch.


  • Fixed the AVX lock problem in gcc 4.9 (work-around). Probably there's a but in the gcc runtime, but didn't manage to reproduce it in a simple test case.

Marco Ca.: we agreed in July to drop gcc 4.8, but there are still issues with 4.9. We need a proper validation cycle.

-- MarcoClemencic - 2015-08-26

Edit | Attach | Watch | Print version | History: r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r2 - 2015-08-26 - MarcoClemencic
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LHCb All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2021 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback