AOB

Speaker : All ( All )


Talk time : 12:55, Duration : 00:05
Pending action items on S&C:
  • Generation of picoDST (Grigory Nigmatkulov on behalf of bulkcorr PWG)
    Status: production_dAu2008 is on disk, 80319 files (or 98%). Did not trace te failure yet. New library assembled for this purpose (and other purposes)
                                 MuDST size    #files
    production_dAu2008/P08ie    30 TB         213797
    cuProductionMinbias/P06ib    6 TB         81938
  • As noted above, new library assembled and being tested (Sl17d)
    Status: contains all code for starting the evaluation, nightly tests were submitted yesterday. See Lidia's fresh announcement here.
     
  • picoDST addiiton: FMS (Maowu / Prithwish / ...) and BEMC (Bingchu / ...) + Ayeh.
     
  • MuDST to picoDST generation - PWGC looking into the datasets that can (should) be converted.
    Current factor x7 space saving.
    See also this post. Small note: the DataManagement is not ready for performing conversion but ready to perform swap (restoring from HPSS). The Generation of picoDST itself is in trial test with Grigory's request as guinea pig.
     
  • production at NERSC/Cori still pending:-(  - Dubna production kaput!
    Status 1: finally, issue with network data transfer found by the EsNET team.
    Status 2: Dubna transfer failing - investigating.
     
  • Pending - d+Au.
    Was presented on 2017-04-12 ( VF study in d+Au ) and action item was to check and update on the use of beamLine 3D
     
  • Near complete - move of directories off the Web server's local storage.
    Identified issues so far (all responsible people informd)
    - Directories with lots of small files - this has prevented backup to finish in any reasonable time (and in fact, was disabled). We need to either stop the propagation of the same scheme generaying small icons, snipets graphs x number of parameters ... or segregate those (and no backup).
    - Saved root files in the monitosirng directory seem redundant
    - Compressed/GZIP root files do not save space (4% a most)
    - Monitoring area should not be a storage space for log files ... (will send instructions on backup into HPSS later in the rts list)
     
  • Web server link issue - Formally not aSTAR issue per-se, ITD has their httpd redirect rules for http->https wrongly formed.
    Status: No acknowledgement of the issue so far.
     
  • Fastoffline statistics (entertainement)
    Currently submitted 3% of files and 2% of events
    The number of files and events highly depends on the resources allocated for this process. Currently, we only use ~ 150 jobs (not too much) over ~ 15,000 possible CPU slots that could be claimed (so only 1%).
     
  • Production status: Latest plan available at You do not have access to view this node and overall status of what is done and not done is available on the Production dataset Inventory page.
    Update from Lidia:
    • Run 16
      • Au+Au 200 GeV st_physics stream data production is completed. 3200 M events are produced in total (AuAu_200_production_2016 and AuAu200_production2_2016)
      • Au+Au 200GeV st_nosst stream data production is also completed, 147M events produced, dataset can be retreived with the trigger setup name = AuAu200_production2_2016.
      • Au+Au 200GeV st_sst stream data production is ongoing, currently produced 348M events - 77% of total, will be done in ~10-12 days with 50% of farm usage.
    • Run 15
      • p+p 200GeV st_ssdmb sream data are processing now. production_pp20trans_2015 dataset is completed, 472M events are produced and running now is under production_pp20long2_2015 dataset, 134M events are produced, which is 28% of total. This will be done in ~9-10 days with 50% of farm usage.
      • production_pp20long_2015 dataset next, 183M events to be processed, will be done in ~6-7 days with 50% of farm usage