I’m tasked with building an online monitor, and I’m looking for advice. I would like one process to build up histograms, and a second process to display them.
My first thought was to use an often-updated ROOT file, and have the display process re-open the file at intervals. However, proof-of-principle scripts show that this doesn’t work if I write my histograms with the TObject::kOverwrite flag. However, not-overwriting will quickly use disk space.
Does anyone have any advice? I don’t have to use a file to communicate between the processes, but it would greatly simplify the design, by making archived and live data files look the same.