The consumer DAQErrMon receives events from the CSL based on the Level 3 triggers MONITOR_DAQ_ERROR and MONITOR_L3TAG_PS200. The former trigger examines the event record and checks for data synchronization across the various readout data banks and for the persistent presence of a list of central banks. The latter trigger is a random sample of Level 2 accepted events without any Level 3 selection. The source code for the MONITOR_DAQ_ERROR trigger can be found at Level3Filters/src/L3DaqErrorFilter.cc

The MONITOR_DAQ_ERROR trigger path is present primarily in physics trigger tables; for cosmics and other decoupled triggers tables, the MB_XING_PS1M random crossing trigger is used. When DAQErrMon receives events, it reruns the MONITOR_DAQ_ERROR algorithm; in case of an error is found, the DAQErrMon records it in the main error histogram, as shown below. Bins:

• EventsIn How many events have been processed for this run so far − should increase steadily during data taking.
• AnyError Sum of DataSyncErr and DataMissing bins.
• DataSyncErr Event count of data sychronization errors − typically one or two errors per hour − a higher fraction indicates a serious problem. If the rate exceeds 20 errors per hour, contact an expert. Note that a burst of errors will show up persistently in the histogram during a run since the histogram is cummulative.
• DataMissing Event count of data missing errors − entries in this bin indicate a bank was missing from the event record − normally this bin is zero

  When a data missing error is encountered, an error message is sent to RunControl via the ConsumerErrorReceiver process. DAQ Ace in this case should immediately issue a Halt-Recover-Run sequence; if this fails to clear the problem, a new DAQ run should be started; if that fails to clear the problem, experts should be contacted.

Back to hardware page