CCM Meeting

Wednesday 9 Oct 2024, 08:00 → 09:00 US/Central

Zoom: https://cern.zoom.us/j/66569029743 (usual password)

08:00 → 08:05 News

Speakers: Alexander Tapper (Imperial College London), Pierre Lasorak (Imperial College London)

08:05 → 08:10 Round table: control

Speakers: Pawel Plesniak (Imperial College London), Pierre Lasorak (Imperial College London), Tiago Alves (Imperial College London), Zbynek Kral (CTU Prague)

drunc gantt chart

drunc issues prioritisation

Pierre

• Env variables merged from last week created havoc on the thread pinning:
  • Reverted a PR from appfwk and confmodel to restore functionality
  • Decided that CLI is going to be the only way the app name and session name is going to be passed, so this is going to require a bit of changes in logging - Marco preparing some PRs.
• Connectivity server: almost done, last item for discussion:
  • Right now each application has a list of Services in exposes_service that is used in the following way:
    • The application has a "_control" service, that is used to find the control service. For example, the root-controller has a Service called: root-controller_control, the dfo-01 has a Service called: dfo-01_control etc...
    • These _control endpoints are used to specify the rest/grpc API.
    • Mapping is 1 "_control" service -> 1 application. This isn't very portable because this means we need to define a Service for each application.
  • Proposal (idea isn't Pierre's, can't remember who suggested it)
    • Define a "ControlService" schema that inherits from "Service" (identical in all ways to the Service)
    • Each application/controller needs to have one (and only one) ControlService (in exposes_services or in a direct relation)
    • We use that to get the control port/service.
    • This means we can now have 1 ControlService for all controllers and 1 ControlService for all daq_applications, if the ports are 0 and if the connectivity service is used.

Pawel

• Wrapping up OpMon work with Marco.
• Starting to work on ERS afterwards.

 

Last remaining for the release:

Hiding developper feature from shifter

 

08:10 → 08:15 Round table: config

Speakers: Alessandro Thea (STFC Rutherford Appleton Laboratory), Giovanna Lehmann (CERN), Gordon Crone (University College London), Henry Wallace (Royal Holloway, University of London), John Freeman (Fermi National Accelerator Laboratory), John Freeman, Marco Roda (University of Liverpool)

Alessandro:

• CLI inspector, adding incremental changes, responding to feedback

John:

• Working on the big move from Session -> System 
• Config plotter doesn't handle more than 1 session. Working on fixing that.

Alex

• Heads-up that we plan a longer meeting dedicated to configuration, tentative is 6th or 7th November

08:15 → 08:20 Round table: k8s etc.

Speakers: Alessandro Thea (STFC Rutherford Appleton Laboratory), Bonnie King (FNAL), Jonathan Hancock (University of Birmingham), Patrick Riehecky (Fermilab), Pierre Lasorak (Imperial College London), Tiago Alves (Imperial College London), Tiago Alves (Imperial College London)

08:20 → 08:25 Round table: monitoring

Speaker: Marco Roda (University of Liverpool)

Grafana
 - Some minor changes requested from trigger people. They are changing the units of what they are publishing

Command line and logging
 - DUNEDAQ_SESSION and DUNEDAQ_APPLICATION_NAME will not be created by RC
   They will not be created by appfwk
   They will be created by logging, changes in logging::setup() required to pass the information from command line
 - PRs in preparation

Working with Pawel on the Opmon binding

08:25 → 08:30 Round table: others