Messages from CommandableFragmentGenerator should be associated with the fragment generator in question
A common occurrence in a running DAQ system is for a pathology associated with a particular fragment generator to be identifiable by a warning or error message which originates in its CommandableFragmentGenerator base class. E.g., on protoDUNE, if an SSP isn't sending data (or is sending it delayed), a message like the following will appear (artdaq 854943d1010efb55594a29324f98ed68e667237d, from March 15, is used here):
2018-03-23 18:15:35 +0100: %MSG-w CommandableFragmentGenerator: Early pre-events 2018-03-23 18:15:35 +0100: A timeout occurred waiting for data to close the request window (max=70293078933504, buffer=0 (if no buffer in memory, this is shown as a 0)). Time waiting: 2002729 us (> 2000000 us).
In a run with, say, four protoDUNE SSPs where only one SSP is causing problems, it would be very helpful if this message could be associated with the problematic SSP. One could imagine that the label associated with the CommandableFragmentGenerator's messages, rather than being simply "CommandableFragmentGenerator", could be the value in the CommandableFragmentGenerator member
instance_name_for_metrics_, or that there could be a new
instance_name_for_messages_<code> member which users could set.
#2 Updated by Eric Flumerfelt over 2 years ago
- Category set to Additional Functionality
- Status changed from New to Closed
- Assignee set to Eric Flumerfelt
- Target version set to artdaq v3_01_00
As part of the TRACE updates, the TRACE_NAME for CommandableFragmentGenerator is now set to <app_name>_CommandableFragmentGenerator.