Weekly Joint PPD/Off/Trig/Comp Ops Meeting

  • Last pp-ref fill was started on Sunday night dumped at 7:30 this morning;
    • 28 pb-1 collected in total for ppRef
  • On Sunday, first tests with heavy ion configuration, a few Pb-Pb bunches circulated in LHC

  • today-tomorrow plans:
    • LHC today: cryo recovery, loss maps
    • LHC tomorrow (Tuesday Nov 24): go for Pb-Pb stable beam ; no stable beams before tomorrow
    • CMS: magnet ramping down this morning (Monday) for re-generation. The goal is to have it back before the Pb-Pb stable beams.


  • 2 modes, one for ppRef (74X) and one for HI (75X)
  • Differentiate Cosmics and HICosmics
  • Only HICollisions get tracker in Virgin Raw mode, otherwise it is zero-suppressed


  • one problem emerged with AlCaLumiPixels primary dataset in replay exercise:
    • contains events w/o FEDRawDataCollection
    • has been followed up by Zhenyu HLT Doc: from run 262307 the issue should be fixed. T0 : can we run an additional replay using with the same Pb-Pb configuration as already done with 7_5_7 except using one of the runs with unstable Pb-Pb collisions
  • note: AlCa streams have a customized event content which is why FEDRawDataCollection went missing from a specific random path; after the fix, there still are random triggers in the AlCaLumiPixels which produce FEDRawDataCollection
  • action on HLT doc: once new HLT ready, take a short run with the test menu, communicate to orm, for further replay

  • (documentation) Final menu for HI:


  • For Vdm and low PU fills it was finally decided to use the ppRun2 scenario (for 25 ns reco) as for the baseline ppRef run
    • although physics performances are not expected to be optimal in ECAL
    • for Vdm what matters is the pixels behaviour, and this is unchanged between scenarios

  • CMSSW_7_5_7 built, reverting the frontier change
    • new replay requested and ongoing
  • replay for Pb-Pb configuration with CMSSW_7_5_6_patch1
    • one problem emerged: AlCaLumiPixels primary dataset contains events w/o FEDRawDataCollection -> to be followed up by Zhenyu HLT Doc (see HLT section)
    • all good things went fine for the rest

Primary Datasets and Skims

  • PDs and skims communicated to T0 (see HLT section)
  • Run2015E prompt reco datasets are not currently subscribed at CERN (/eos/cms/store/data)

All of the HI skims data stream shows up in the CMSSW_7_5_6_patch1 replay. The maps seems working correctly. The skim rates to be verified.

1) Photon / HIPhoton40AndZ / AOD /HIPhoton40AndZ/Tier0_Test_SUPERBUNNIES_vocms047-Photon-PromptReco-v104/AOD

2) ZEE / HIPhoton40AndZ / AOD /HIPhoton40AndZ/Tier0_Test_SUPERBUNNIES_vocms047-ZEE-PromptReco-v104/AOD

3) ZMM / HIEWQExo / RAW-RECO /HIEWQExo/Tier0_Test_SUPERBUNNIES_vocms047-ZMM-PromptReco-v104/RAW-RECO

4) BJet / HIHardProbes / AOD /HIHardProbes/Tier0_Test_SUPERBUNNIES_vocms047-BJet-PromptReco-v104/AOD

5) D0Meson / HIHardProbes / RECO /HIHardProbes/Tier0_Test_SUPERBUNNIES_vocms047-D0Meson-PromptReco-v104/RECO

6) HighPtJet / HIHardProbes / RAW-RECO /HIHardProbes/Tier0_Test_SUPERBUNNIES_vocms047-HighPtJet-PromptReco-v104/RAW-RECO

7) OniaCentral / HIOniaCentral30L2L3 / RECO /HIOniaCentral30L2L3/Tier0_Test_SUPERBUNNIES_vocms047-OniaCentral-PromptReco-v104/RECO

8) OniaPeripheral / HIOniaPeripheral30100 / RECO /HIOniaPeripheral30100/Tier0_Test_SUPERBUNNIES_vocms047-OniaPeripheral-PromptReco-v104/RECO

9) SingleTrack / HIHardProbes / AOD /HIHardProbes/Tier0_Test_SUPERBUNNIES_vocms047-SingleTrack-PromptReco-v104/AOD

10) MinBias / HIMinimumBias1 / AOD /HIMinimumBias1/Tier0_Test_SUPERBUNNIES_vocms047-MinBias-PromptReco-v104/AOD

11) OniaUPC / HIForward / RAW-RECO /HIForward/Tier0_Test_SUPERBUNNIES_vocms047-OniaUPC-PromptReco-v104/RAW-RECO


  • Latest GT announcement to Tier0
  • HI express: 75X_dataRun2_ExpressHI_v2
  • HI prompt: 75X_dataRun2_PromptHI_v3
    • new GT to be announced for HLT

  • A few conditions updates have been measured on the ppRef runs, updates proposed; full-validation procedure is starting now, updates to GT's may follow upon green-light

DQM & Visualization

* fixed by CMSSW_7_5_6. * fixed by CMSSW_7_5_6 ?
  • The work-around for frontier works for DQM. 7_5_7 will be tried out for DQM, to assess the frontier revert for a possible update of the Pb-Pb configuration


Certification of pp reference run will start this week. Only first few runs are currently available for certification.
It should be clarified if Castor status should be included in the Golden JSON file. A discussion with HI experts will be done to clarify this point for pp and for HI runs.

Monte Carlo

Computing Operations (Tier-0)

  • large size of streamer files (from the Storage Manager streams) to be repacked generated jobs at T0 taking anomalously long time
  • Partial re-organization of the Primary Dataset => Stream to be discussed to mitigate the duration of the repacking (and start of prompt reco) for PbPb running:
    • splitting streams will reduce size of the files mitigating the T0 problem
    • too much splitting will generate excessive load on the Storage Manager file system
    • proposal for discusison: split streamHIPhysicsMinBiasUPC in three, keep the other streams the same

  • Streams and Primary datasets:
HIPhysicsMuon: -> 18 GB/lumi
* HIOniaCentral30L2L3: ~~5 GB/lumi
* HIOniaL1DoubleMu0: ~5 GB/lumi
* HIOniaPeripheral30100: ~2 GB/lumi
* HIEWQExo: ~40 Hz; 2.2 GB/lumi
* HIOniaTnP: ~50 Hz; 5.9 GB/lumi

HIPhysicsHardProbes: -> 14 GB/lumi
* HIHardProbesPeripheral: ~60 Hz 2 GB/lumi
* HIPhoton40AndZ: ~20 Hz; 3.4 GB/lumi
* HIFlowCorr: ~50 Hz; 4 GB/lumi
* HIHardProbes: ~40 Hz 6 GB/lumi

HIPhysicsMinimumBiasUPC: -> 15 GB/lumi
* HIMinimumBias1: ~300 Hz (fluctuates a lot with prescale columns); 8 GB/lumi
* HIMinimumBias2: ~500 Hz (peripheral-dominated PD, fluctuates a lot with; prescale columns); 5 GB/lumi
* HIForward: ~100 Hz (wild guess);

  • agreed actions on large streamer size of files:
    • HLT doc: split HIPhysicsMuon and HIPhysicsHardProbes in two, with minimal overlap possible
    • no actions on SM; no action on T0
    • HLT doc: once new HLT ready, take a short run with the test menu, communicate to orm, for further replay

  • switch to HIRun2015 acquisition - plan agreed upon by T0
    • must check that all ppRef runs have been transferred to T0, before the switch
    • planned to today, Monday
    • action on ORM to communicate to T0 when switch must happen

  • switch T0 configuration for express ( see also this) to heavy ion, first - plan agreed upon by T0:
    • once repacking has finished all ppRef runs
    • once express has finished over all ppRef runs (Express settings change take effect immediately)
  • switch T0 configuration for Prompt to heavy ion ( see also this), with these steps (proposal) - plan agreed upon by T0:
    • leave ppRef configuration till Wednesday morning
    • stop the T0
    • then manually release all the outstanding ppRef runs for PromptReco, so they pick the correct configuration
    • switch the configuration of T0 the heavy ion
    • un-stop the T0, which from that point onwards will create jobs on new runs only using heavy ion configuration


-- FabioCossutti - 2015-11-22-- FabioCossutti - 2015-11-22

Edit | Attach | Watch | Print version | History: r8 < r7 < r6 < r5 < r4 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r8 - 2015-11-23 - GiovanniFranzoni
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    CMSPublic 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