Analysis Systems
Carry over from 2 weeks ago:
- Uproot, coffea, ATLAS PHYSLITE
- Full reading of PHYSLITE latest schema with uproot/coffea 2024 still blocked by https://gitlab.cern.ch/atlas-asg/columnar-analysis-benchmarks/-/issues/1
- ATLAS OpenData
- Draft of open data files available now (in final prep)
- $ rucio list-datasets-rse CERN-PROD_OPENDATA > CERN_PROD_OPENDATA.txt
- Currently not enough space for the ~60 TB at MWT2_UC_SCRATCHDISK
while IFS= read -r did; do rucio add-rule "${did}" 1 MWT2_UC_SCRATCHDISK; done < CERN_PROD_OPENDATA.txt
- Would be useful to have ATLAS open data at AFs for demos and testing of workflows
- N.B.: Obviously more useful to use more recent ATLAS data, but open data is still useful especially for collaborating on problems with Analysis Systems teams and Scikit-HEP developers
Software
Reminder: OSG 3.6 is in critical bug/security fix only mode
Release
- HTCondor 23.6.0 (this week)
- XRootD 5.6.9 + IO time gstream monitoring patch (soon, https://opensciencegrid.atlassian.net/issues/SOFTWARE-5850)
KAPEL + Gratia integration: we've taken a first stab at the Gratia specific container and are testing it against the GRACC ITB
Security
WLCG SOC Hackathon outcomes:
- Further improvement/refinement to the docker-compose configuration for MISP
- https://github.com/JiscCTI/misp-docker
- In-progress updates to add support for OIDC nearly complete
- Goal is to have a straightforward deployment model for WLCG sites
- Discussion about long-term support for pDNSSOC
- Currently not production ready (no RPMs etc)
- Some question as to who will provide support/development
- Plan is to deploy it at some small subset of WLCG sites along with MISP to see if this is viable/useful prior to solving those problems
- FNAL hosting a MISP instance?
- Some discussion with FNAL about hosting a MISP instance that could sync from the WLCG instances (CERN, etc)
- Could be used to provide threat intel to US WLCG sites
- Some logistical hurdles remain to be solved, no definite timeline as of yet
