search

art-framework-suite / art

The implementation of the art physics event processing framework
Other
2 stars 7 forks source link

Issues with events coming from different art histories in artdaq #90

Closed knoepfel closed 3 years ago

knoepfel commented 3 years ago

This issue has been migrated from https://cdcvs.fnal.gov/redmine/issues/19846 (FNAL account required) Originally created by @eflumerf on 2018-05-01 20:42:10

In artdaq, we have situations where the output from more than one process is combined in another art process. We have noticed that this process only works as expected when the first set of art processes have identical configurations. This is an issue for protoDUNE, as they would like to have for DataLoggers writing data to disk (with different fileNames defined for their RootOutput_modules), and reporting events to one Dispatcher for online monitoring.

Our current minimum working example for this behavior is the artdaq-demo: wget https://cdcvs.fnal.gov/redmine/projects/artdaq-demo/repository/revisions/develop/raw/tools/quick-mrb-start.sh chmod +x quick-mrb-start.sh ./quick-mrb-start.sh --run-demo

Change the configuration in artdaq-utilities-daqinterface/simple_test_configs/demo/EventBuilder1.fcl to be different from /EventBuilder2.fcl, and only one set of events will appear correctly in the output file/online monitor.

The issue may come from the fact that only one art process' metadata is loaded in the downstream art process, and the input source therefore cannot process events from the other upstream art processes.

This is just an outline, let me know what questions you might have...

knoepfel commented 3 years ago

Comment by @knoepfel on 2018-05-07 15:50:28

Based on discussions with Paul, the designer of the transport system between the event-builders and aggregators, the intention was for all configurations to be consistent. This was due to constraints imposed by art's internals. To adjust this behavior will take some analysis. We are changing the tracker value from "Bug" to "Support". We will have to schedule this analysis for after art 3.0 is released.

knoepfel commented 3 years ago

Comment by @knoepfel on 2018-05-14 15:21:47

We will contact you for a meeting after art 3 is out.

eflumerf commented 3 years ago

We have worked around this issue by overwriting event histories in the artdaq input sources. Since the DAQ is the progenitor of all events we use (so far, at least), we can safely discard histories as needed.

knoepfel commented 3 years ago

Closed based on information from @eflumerf.