LHCb Bookkeeping Working Group Meeting

Date and Location

27 April 2007
14:30 - 15:30

Attendees

Carmine, Markus, Michael, Olivier C., Olivier D., Philippe (chair, notes)

Apologies

Andrew, Niko

Subjects

Summary of the recommendations made so far

A presentation is planned for the coming SW week (9-11 May). Philippe will prepare slide that someone else could possibly present in case he cannot be present.

  • The warehouse schema seems adquate for real data, if the "job" producing the RAW files corresponds to a DAQ run.
  • The BKDB would be logically split in two parts: real data and MC. The warehouse schema is shared, but view contents can be different as the queries are not the same.
  • A table would contain a list of "periods" that are a set of runs. Within each period, processed data can be mixed and any chronological order might be lost. A similar concept could possibly be useful for MC data, where a "period" would be a set of productions certified for analysing a certain real time period.
  • Processing number: a dedicated table should contain the definition of a "processing pass", that consists of a set of pairs (application version, CondDB tag). It is recommended that although in the CondDB there could be separate tags for the base geometry and the alignement/calibration only the latter be mentionned as it refers to a single geometry tag and thus this information should be contained in it.
  • The BKDB "configuration" consists of the partition and the run type. For the browser, there is the request to not present the two as a single entity but as a tree in order to simplify the choice.
  • For queries, a tree browsing facility is recommended. However, users should be able to configure the tree presentation (e.g. by processing or by application) as well as the order of selection (although the default one should be the main case). This implies "preference-like" menus for the browser. Users request the possibility to ask for a given number of events rather than files.
  • Datasets: a consistent view of datasets in BK, DIRAC and Ganga is a must. Action: PhC to organise a meeting with the 3 parties.
  • The views should allow a variable unlimited number of steps in the history of files (currently limited to the last 3 steps). Carmine is working on it currently and the technical problem is not simple...
  • Browser: a single browser is recommended. Due to technical limitations, the WG proposes that support for the web browser be dropped in favor of a pyQt browser usable as standalone or within Ganga. The web facility should be kept for file search still and history browsing, access to log files etc... as well as for providing statistics.

Next meeting:

Next week if Philippe can make it (Friday 4 May @ 14:30)

-- Main.phicharp - 30 Apr 2007

Edit | Attach | Watch | Print version | History: r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r1 - 2007-04-30 - PhilippeCharpentierSecondary
 
    • 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