LHCONE R&D and RNT-WG call

Tuesday 8 Apr 2025, 17:00 → 18:00 Europe/Zurich

Edoardo Martelli (CERN), Joe Mambretti (International Center for Advanced Internet Research Northwestern University), Marian Babik (CERN), Shawn Mc Kee (University of Michigan (US))

1 LHCONE

Presents: Edoardo Martelli, Harvey Newman, Phil DeMar, Bruno Hoeft, Luca Bassi, Pablo Collado Soto, Bill Johnston, Garhan Atterbury, Jim Chen, Dale Carder, Tristan Scott, Ettore Tamiro, Marian Babik, Karl Newell, Joe Mambretti

Change of frequency the bi-weekly call:
- Reduce frequency to 4 weeks.  We can resume to 2 weeks closer to Super Computing
- Next meeting on the 6th of May

LHCOPN meeting actions
- Change to AUP to allow temporary resources: no progress
- MultiONE, tag for NOT-AUP : defined in the twiki page https://twiki.cern.ch/twiki/bin/view/LHCONE/MultiOneBGPcommunities ; DONE
- MultiONE tags for opportunistic resources: no progress
- perSONAR: remove IPv4 tests between dual-stack endpoints; DONE
- IPv6 wg to make a proposal to remove IPv4 from LHCOPN: no progress

Next f2f meeting:
- LHCOPN meeting confirmed in KIT on 7-9 of October . It can be possible to host the -SIG-NGN meeting.

 

2 RNTWG

Outreach
- CHEP paper - received comments (https://indico.cern.ch/event/1338689/papers/6010796/)
     - Paper is at https://www.overleaf.com/read/jwwxfcqcpqqs#3ae27b
- Scitags roadshow: Tim gave a talk at GEANT SIG-NOC last week and GNA-G Europe this week; next talk is at GNA-G US on Thursday 
- ISE submission - we agreed to extend the draft before re-submitting
- Plan to update our web page 

Collectors
- Missing Kafka input/output plugin (A: Andy Lake)
- Meeting with GEANT dedicated to collectors will be this Thursday
- Reaching ESnet collector from RAL doesn't seem to work (and is not easy to debug)
      - For now I have offered CERN site collector as intermediate, but having UK collector up and running soon would be great
      - Added monitoring to CERN site-collector to understand the workload (uses logstash exporter/prometheus and grafana)

Registry
- Setup new API at https://scitags.docs.cern.ch/api.json (based at CERN, to be used by WLCG experiments and their storages)
     - This URL is used to configure fallback mechanism in xrootd, so in principle all WLCG storage nodes can access it
- The previous access http://api.scitags.org still works and is in sync with the CERN one (points to Github)
- Will keep running both for the time being

Dashboard (@ESnet)
- Following CMS xrootd configuration campaign we now have some sites sending fireflies but not showing up in the dashboard (CNAF, RAL)

dCache 
- Submitted two PRs - both were accepted
   - Support for sending fireflies, parsing information from xroot/http protocols 
   - Also implemented fallback mechanism (if no information in xroot/http) - based on x509 VO (user certificate)
- Currently working on adding stats: usage (bytes sent/rcvd) and storage technology statistics (read/write bandwidth, read active/idle, write active/idle)

xrootd
- RAL enabled fireflies, but there was an issue with configuration which was found and fixed in collab. with xrootd developers 
    - For fallback mechanism to work we need to have both directory mapping and activity mapping in the config

EOS (CERN)
- Requested to change the configuration, so that we start tracking flows not coming from FTS/Rucio (fallback mechanism based on directory access)
- EOS team is looking into it, the change adds Scitags API URL, which becomes an operational dependency

flowd-go
- flowd-go evaluation/review (A: Marian, Tristan)
   - Packaging in koji in progress (TBD)

3 AOB

4 Next meeting

- Reduce frequency to 4 weeks.  We can resume to 2 weeks closer to Super Computing
- Next meeting on the 6th of May