80 likes | 175 Views
Focus on 3 Global items. Muon Week, CERN 17-21 February 2014 Nicoletta Garelli ( SLAC). Fragment Sampling. Presented at DAQ-Detectors meeting on 27 Jan 2014 by Wainer (https:// indico.cern.ch/event/296275) DAQ and Monitoring ( gnam ) experts should decide what to do
E N D
Focus on 3 Global items Muon Week, CERN 17-21 February 2014 NicolettaGarelli (SLAC)
Fragment Sampling • Presented at DAQ-Detectors meeting on 27 Jan 2014 by Wainer (https://indico.cern.ch/event/296275) • DAQ and Monitoring (gnam) experts should decide what to do • RUN I: muon systems sample events at ROS • Which type of data? • Which rate? • RUN II: Is the sampling at ROS really needed? Why? • Note: it is not sufficient to try to compile against tdaq-05-03-00 and discover what is new, some planning and reading of release notes is necessary
Wainer’s slide • Is ROS fragment sampling needed (in physics datataking or DAQslice mode)? • sampling of full events at the HLT farm is possible • a 'Monitoring' stream type has been introduced • dedicated HLT algorithms can accept events for monitoring → events will be fully built and made available over emon but not saved to disk • If ROS fragment sampling is needed, is it acceptable to receive partial data? • not containing all ROL fragments for a given events • of course ROS can select requests containing data from all ROLs • no guarantees on the rate at which this will happen → depends on trigger menu
Enable/Disable Resources • If a part of the detector is described in the data taking database (OKS) as a Resource, it can be automatically enabled/disabled from the data taking • status of the resource saved in COOL under /TDAQ/EnabledResources/… • We do not want to use DCS for storing this info in COOL (PVSS2COOL) anymore • We need to: • Define a reasonable granularity for each subsystem • Define if/when to issue LB change • Modify OKS • Send the proper command to CHIP • Test
Some technical notes from CHIP twiki • Light-weight procedure to disable/enable modules during a run that have smaller granularity than a ROL and for which the BUSY and ECR setting are handled internally, if needed. • The only thing that the CHIP does is notifying the ResInfoProvider of the change of status of some resources. Sequence: • Detector specific software sends a ERS issue of type daq::rc::ModulesDisabled or daq::rc::ModulesEnabled. The parameters contain a comma separated list of modules and a flag indicating whether the luminosity block shall be changed or not (NEW!). • the parameters are validated (they shall be valid resources in OKS) and the change in enabled/disabled resources is notified to the ResourcesInfoProvider application using a USER command with arguments that are encoded using a dedicated utility (class Utils). • if appropriate the MasterTrigger is asked to change luminosity block.
Round-Table • MDT • Run I: chambers and mezzanines enabled/disabled by RCD, but info stored in COOL via DCS • Run II: Describe chambers as Resources, but keep mezzanines in DCS only for the time being • TGC • Run I: sector granularity. • Run II: Daniel would like to implement the star switch granularity (~10 per sector). Today, when a star switch times out the ROD drops it, but not listed as Resources • RPC • Run II: at tower and trigger sector level • CSC • Run II: at ROD and chamber level
DAQ-DCS Communication • DT: Data Transfer - CT: Command Transfer - MT: Message Transfer • CSC • Run I: DT for sending enabled/disabled list at the beginning of the run + MT for checking HV status • Run II: ? Drop it? • MDT • Run I: DT for sending enabled/disabled list + MT for sending ‘end of initialitation’ • Run II: ? • TGC • Run I: ? • Run II: ? • RPC • Run I: ? • Run II: ?