Skip to content

Implement robust configuration

Summary

At the moment any FE instability makes the "Configure", "Start" and "Stop" state transitions fail.

What is the expected correct behavior?

A robust configuration mode that ignores/bypassed/workarounds the communication instabilities should exist.

Possible fixes

Implement the following actions during state transitions:

  • Initial:
    • All "hard" configuration loaded, i.e. xDAQ XML configuration, hardware layout tree,...
    • No interaction with other software components
    • No interaction with the hardware
  • Initialize:
    • Establish the communication between the different relevant applications
    • No interaction with the hardware, i.e. no recovering of the back-end board
    • Duration goal: < 10 s
  • Configure:
    • Recover the back-end boards
    • Prepare the front-end communication
    • Configure the back-end boards
    • No non-robust interaction with the FE
    • Provides a list of the chambers that can potentially be power-cycled, i.e. the ones that we already know will be disabled
    • No timeout
    • Duration goal: < 15 s
  • Start:
    • Pushes the configuration to the FE:
      • Exists in both robust and non-robust mode
      • Non-robust mode fails for the first error encountered
      • Robust mode disables chambers for which an error is encountered (finer granularity to come later)
    • Reset the counters
    • Enable the events building and trigger
    • Duration goal: < 15 s
  • Stop
    • No non-robust interaction with the FE
Edited by Laurent Petre