LHCb Core Software Meeting

Date and Location

January 30, 2008
10:30 - 11:45
CERN (2-R-030)

Attendees

Gloria, Hubert, Marco Ca., Marco Cl. (minutes), Markus, Olivier, Patrick (EVO), Philippe

Apologies

Thomas

Subjects

News

  • LCG 54 released
    • ROOT 5.28/00 has got a bug preventing to use Python on OSX

Software Releases

Gaudi (Hubert, Marco Cl.)

Marco Ca. reported a problem of compatibility between Gaudi v19r6 and Gaudi v19r7: ETC files produced with one version cannot be read with the other. It might be a problem in ROOT or in Gaudi. If it is a Gaudi problem, we have to re-tag Gaudi v19r7 (unfortunately already announced, but not used by anybody). Marco Cl. and Markus will investigate the problem.

Marco Cl. fixed a problem in GaudiPython: the options for the MessageSvc were not passed to the instance (so no way of fine tuning of output levels). He also added a lock to the configurables so that they cannot be modifies after the instantiation of GaudiPython.AppMgr (i.e. when it is too late to do anything useful).

Hubert reported a problem in the LCG tarballs. The PyQT tarball is incorrect and there are inconsistencies between the version directory name and the tag *_native_version. Already agreed with Stefan about how to sort out the inconsistencies without impacts on LHCb scripts.

LHCb, Boole, Brunel (Marco Ca.)

The release is waiting for more information about the Gaudi problem.

We need to stop the DC06 support in new versions of Boole. Gloria reminds that the same apply to Panoramix.
We may also have problem with XmlVis. For the moment we try to make it work with both old and new geometries. If/when it will not be possible we will use the same trick used in XmlDDDB (different entry points in the XML hierarchy).

Gauss (Gloria)

A DC06-compatible release is ready to be deployed. There is a problem with a special generator library (ipgen) that is taken from the new GENSER tree, but it is not in the LCG (53f) tarball. Since it is needed only for special cases, we can address the problem only when it occurs.

G4 people ask for a round table with experiments to understand the requests for new versions and support for old ones. The basic question is: how long do we want that a version is supported? It has to be discussed with The Physics Planning Group.

HLT/DaVinci (Patrick)

DaVinci is essentially ready.

Few things to be checked in HLT.

Online (Markus)

Some patches on Gaudi are needed for Online to work (another reason why to re-tag Gaudi).

Panoramix (Marco Cl. for Thomas)

Guy published OpenScientist v16r2 (using Python 2.5), used to build Panoramix v15r10 in DEV on top of latest Gaudi.

There is a problem with Panoramix_main.exe. It seems that ROOT is started in batch mode, so RootSvc fails to initialize. In Python there is no problem (we have more control on the order of loading).
We can drop the support for the binary executable (since we are going to move away from binary executables for all applications).

Guy will be at CERN next week for a few days. He might give us a demo of the new features of OpenScientist.

Round Table

Marco Cl.
  • Andrew Maier reported a problem with the new default of CMTCONFIG. 64 bits Operating Systems are not supported on the GRID, so, if a user connect to lxplus, compile his algorithms and submit jobs on the grid, they will never run. Andrew's suggestion is to switch back the default to 32 bits.
    After some discussion, the default will not be changed. It is believed that Ganga should do something (e.g. issue a warning).
  • Added to SQLDDDB the settings needed to be able to use the Oracle back-end instead of SQLite. It will be enough to get a Grid proxy certificate and append SQLDDDB-Oracle.opts to the options. (Marco Ca. cannot use it because of problems with the registration in the VO)
Hubert
  • Will prepare a pattern for Python binary modules
  • A new version of CMT is foreseen. To be thoroughly tested before using it.
Patrick
Gerard sent a mail to ask for automatic CVS mail notifications.
This is currently not possible because the scripts provided by CERN are unusable for us (one cannot register on single packages: either everything or nothing).

-- MarcoClemencic - 01 Feb 2008

Edit | Attach | Watch | Print version | History: r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r1 - 2008-02-01 - 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 & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback