Support #10622

Kurt's notes on issues, October 2015

Added by Kurt Biery over 5 years ago. Updated over 5 years ago.

Start date:
Due date:
% Done:


Estimated time:


To help me keep track of the issues that I've seen, heard about, and thought of, I want to write them down in this Issue. Of course, it might be good to create individual Issues for each topic, but I'm not sure if everyone would be open to that option.

Observed errors:
  • "failed to connect" complaints from artdaq processes when the msgviewer is not running (should have better log message, and maybe shouldn't be error)
  • "unsufficient buffers" errors from RCE readout code at end-run, most often when there is persistent backpressure
Larger-scale issues:
  • synchronizing the data across the detector (RCEs and SSPs) at begin-run and end-run
  • automatically closing data files after they reach a configurable size, number of events, or duration


#1 Updated by Kurt Biery over 5 years ago

  • Yesterday, Ricardo tried taking data with SSP08 at PC4, but I thought that SSP08 was still at DAB. How can we "ping" SSPs and RCEs just to confirm which ones are physically connected to a particular DAQ?

#2 Updated by Kurt Biery over 5 years ago

  1. It would be nice to be able to select a new configuration in the Ready start and run "init" from there (at least from the user perspective).
  2. why do we get "Unknown exception '<Fault 1: "<class 'psycopg2.InternalError'>:current transaction is aborted, commands ignored until end of transaction block\n">'" sometimes when we run lbnecmd commands?

#3 Updated by Kurt Biery over 5 years ago

  1. Mark events somehow when they are flushed from the EventStore, but don't have all of their fragments.

#4 Updated by Kurt Biery over 5 years ago

At the 35-ton meeting on 28-Oct, Alex Himmel reported that there are "missing SSP" events. That is, 100 triggers are requested and only 30 events show up on disk.

Also available in: Atom PDF