Speakers:
Alessandro Thea
(STFC Rutherford Appleton Laboratory), Asher Kaboth
(RHUL), Roland Sipos
(CERN), Wesley Ketchum
(Fermi National Accelerator Laboratory)
08:20
→
08:30
Run Coordination news10m
Speaker:
Wesley Ketchum
(Fermi National Accelerator Laboratory)
08:30
→
08:45
Activity coordination round table15m
Release Coordination v5
Iceberg
ND test setups
SW coordination
Speakers:
Bonnie King
(FNAL), Eric Flumerfelt
(Fermilab), John Freeman, Kurt Biery
(Fermilab), Wesley Ketchum
(Fermi National Accelerator Laboratory)
Software Coordination, John:
At the last v5.2 release meeting I mentioned that it was hard running unit tests, etc. in containers. But: not so on bare metal. So Andrew's working on creating a GitHub Workflow which can run nightly unit tests and clang-format checks. Linting would also be nice, but takes several minutes for a given package, so perhaps this last should be a weekly thing.
We also have a new nightly workflow which tests various aspects of daq-buildtools; this was done after Kurt ran into problems (since fixed) which dbt-clang-format.sh
I've also been looking into resurrecting use of the lcov code-coverage tool which Eric worked on all the way back in 2021. The question it answers: how much of our codebase do our unit tests hit. I've been installing lcov as a local Spack package which takes about 5 minutes and 50 MB, bring in over 50 Perl packages. I think we can agree that we don't want this in our image, but having it on /cvmfs would be convenient
Related to lcov: unlike with Spack 0.20.0, with Spack 0.22.0 which we switched to over the summer, the incredibly annoying "feature" where build-only dependencies of a package got loaded in by spack load is *gone*. This makes the idea of a build environent vs. a run environment feasible again. Meaning, e.g., we could have lcov, CMake, etc., loaded into a local work area, but not by dbt-setup-release
v5.2 Release, Eric:
Candidate created on Thursday, testing started on Friday.
Using Google Forms for reporting tests
One more RC needed (updated trigger tag)
Testing should be largely complete by the end of the week
08:45
→
09:00
Working groups round table15m
Speakers:
Adam Barcock
(UKRI STFC), Alec Habig
(Univ. of Minnesota Duluth), Alexander Tapper
(Imperial College London), Artur Sztuc
(University College London), Bonnie King
(FNAL), David Cussans
(University of Bristol), Jonathan Hays
(Queen Mary University of London), Joshua Klein, Kurt Biery
(Fermilab), Pierre Lasorak
(Imperial College London), Roland Sipos
(CERN), Stoyan Trilov, Wesley Ketchum
(Fermi National Accelerator Laboratory)
Timing config schema refinement: e.g.: full GIB/Gaia support
Full custom timng command integration with drunc: implement full command list
CCM:
Control:
In v5.2:
A lot of bug fixes (RTE, Zombie)
FSM prefix removed, enabling batch mode
Log path in configuration
SSH validator script (only runs ssh to check that one is able to ssh on each host specified in the configuration)
Shell formating
Trigger rate changes
Dynamic control ports
v5.2.0 delayed:
Opmon
ERS
Daemon Process Manager
Post v5.2 (Note: we merged new features in develop_post_v5.2.0, which will be merged in develop today).
Unit tests (started)
Recursive statuses (removing `get_children_status`, and `ls`): A small kerfuffle here, should have been merged in develop_post_v5.2.0, but was merged in develop and also broke everything on Friday, so reverted.
Daemon Process Manager
--option rather than argument (users will have to enter “start --run-number 1234” rather than implied “start 1234").
FSM schema upgrade on-going
Opmon/ERS
Still need to prioritise the issues.
In parallel, UI progressing:
First version of the Process Manager presented
Need to discuss packaging and how to make it available for everyone
Started controller UI. Discussion on-going
Configuration:
In v5.2:
Initial version of TUI utility to modify configuration
Configuration inspector (shell-based)
generate.py in daqconf from Eric
Dynamic control ports
Connectivity server configuration
v5.2.0. delayed:
Session V System renaming (meeting John on Tuesday to resume this work).
Connectivity server configuration schema rework
In parallel:
Configuration workshop (this week!!)
Monitoring
In v5.2:
Slight update to opmon configuration
Opmon and ERS python bindings.
v5.2.0 delayed:
None
Post v5.2.0
?
Core Software
The recent two meetings of the Core Software WG have focussed on presentations and discussion of the issues with translating DataRequests into the correct number and time window of data frames, particularly with WIBEth data. A first round of unit tests and fixes to the code are included in the first fddaq-v5.2.0 release candidate. There are additional discussions that need to happen and work that needs to be done to properly address all of the issues that have been uncovered, and we plan to move ahead on those in the coming weeks.
Trigger
Tuning of the Ground-Shake configuration & offline ground-shake study done by Michal.
All the deliverables from Trigger for v5.2 done, looking out for any bugs to fix.
Looking beyond v5.2:
New PR that enables all the triggering algorithms to output any TC-type
Michal working on the replay application. Porting to v5 first, rewriting second.
Planning some maintenance work: 10s of issues on github (some still valid), known ways to simplify trigger configuration.
Planning work ahead: rewriting MLT, decoupling TC types from TC algorithms etc.
09:00
→
09:05
Actions5m
- Alessandro: request 2 parallel sessions at the CERN Collaboration Week.
- Pierre/Stoyan: report on early findings about the CCM/timing-as-a-service interface requirements.
- Roland, Artur, Josh: organise discussion about TP format update.