WLCG DOMA general meeting
NONE
-
- 16:00 → 16:10
-
16:10
→
16:30
Third Party Copy 20mSpeakers: Alessandra Forti (University of Manchester (GB)), Brian Paul Bockelman (University of Nebraska Lincoln (US))
Current status
Sites are deploying the most recent version of the software pushed by WLCG Ops Coordination. Should be done by spring for DPM and dcache.
- dcache sites are upgrading but not configuring the new features for xrootd this needs to be corrected
- DPM sites more generally the TF was setup to do both
- StoRM and plain xrootd sites don't have a TF will need to tackle site by site
- EOS and echo are working on their implementation
Stress tests
DOMA stress tests continuing with few sites not all production as we wanted. There are still several long threads ongoing to iron out problems created by the increase in load. Both for xrootd and for http. RAL should start stress testing echo soon.
ATLAS is going to start stress testing a handful of production sites next week. At the moment these will only be DPM sites (Lancaster, Manchester, Prague) because they are the sites of the people involved in looking after the testing. More sites will be added at a later stage. Both protocols will be tested. This will allow not only to test larger transfers but will help understanding what parameters will need to be tweaked also on the experiement side and in rucio. See slides attached for details.
Authzwork on Authz is continuing in parallel with the organisation of another hackaton at the WLCG/HSF workshop in Lund (dates still being discussed). There is token authz work ongoing beyond the storage, particularly related to HtCondor-CE and WMSs in ATLAS and CMS so we thought to create also a communication channel on the discourse instance we setup for wlcg (originally for k8s but now slightly extended to wlcg new technologies acitivities) https://wlcg-discourse.web.cern.ch/c/wlcg-token-authorization
HL-LHC review document
There will be a document hackaton likely this week with the aim to be ready by the rucio workshop at FNAL. - 16:30 → 16:50
-
16:50
→
17:10
Content delivery and caching 20mSpeakers: Frank Wuerthwein (UCSD), Ilija Vukotic (University of Chicago (US)), Markus Schulz (CERN), Dr Stephane Jezequel (LAPP-Annecy CNRS/USMB (FR)), Xavier Espinal (CERN)
ACCESS WG activity
- Discussions on the future of compact analysis datasets and analysis models evolution. Triggered by the presentation form Andrea Rizzi on CMS's nano-AODs
- XCache deployments in US and EU with substantial different approaches and XCache developments. Need to focus on introduce costs estimation with the collaboration of the cost model WG.
- Focus on the HL-LHC review. This is seen as the basis for a white paper to conclude the activity of the WG, towards the end of the year. The WG was created for addressing the analysis use case but the scope most probably should be enlarged to cover full data distribution and data access challenges.