-
Notifications
You must be signed in to change notification settings - Fork 1
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Changes to avoid warning messages when tardy TPs arrive at the TPStreamWriter #358
Conversation
…ce to support our desire to not complain too loudly when late-arriving TPs can't be added to existing TimeSlices in the TPStreamWriter.
…mWriter and to better track such a condition with additional metrics.
Questions to be discussed in meetings on Wednesday...
|
Results from a 30-second emulated-data test run with 2 ReadoutApps, each with 3 links. In this run, the TPs from the two RUs were close in time (so no warnings should have been produced). The flag for allowing the warning messages was set to false.
|
Results from a 30-second emulated-data test run with 2 ReadoutApps, each with 3 links. In this run, the TPs from the two RUs were close in time (so no warnings should have been produced). The flag for allowing the warning messages was set to true. (but, no warning messages were produced, as expected)
|
Results from a 30-second emulated-data test run with 2 ReadoutApps, each with 3 links. In this run, the TPs from the two RUs had a significant amount of time between them (~2.5 seconds), so TPs should have been dropped. The flag for allowing the warning messages was set to false, so no warning message should have been produced, but the loss of TPs should be visible in the metrics.
To avoid warning messages about TPs not being available in the RU latency buffers, the data request timeout was increased in this run, and extra time was allowed at the start and end of the run without triggers.
|
Results from a 30-second emulated-data test run with 2 ReadoutApps, each with 3 links. In this run, the TPs from the two RUs had a significant amount of time between them (~2.5 seconds), so TPs should have been dropped. The flag for allowing the warning messages was set to true, so warning message should have been produced, and the loss of TPs should be visible in the metrics.
|
|
… the TimeSlice class where it is used in the TPStreamWriter code.
…discussion at 03-July Core SW meeting
…tpsw_ignore_late_tps
Here are samples of the latest metrics (shown in collated form from the disk-output opmon mode):
and here is a sample warning message:
|
This PR depends on ones in daqdataformats and hdf5libs.
These changes include a flag to control whether warning messages are produced when tardy TPs arrive at the TPStreamWriter. They also include additional metrics to provide insight into lost TPs when the warning messages are disabled.