Issues:

• Pilot failures due to xrootd-set nproc limit
  • Fully mitigated now
  • Turned out xrootd client also sets max open files limit when imported :-(
    • Has not affected anything (yet)
• A lot of rescheduled and failed jobs at all T1 sites, due to a problem with EOS-pilot SE at CERN
  • Not our fault
• Lots of failed transfers to SARA from all T1 sites
  • Again, not our fault

Data Challenge (sorry, can not edit appropriate section):

• Very good throughput on Monday and Tuesday, well above the target
  • Much better than during the DC24 
  • the rate reduce a bit during the OPN cut, but still was close to the target
• Not so good rate on Wednesday, where gateways were struggling
• Read test on Thurdsay (RAL -> CERN), failed completely due to storage problems at CERN
• Checksum tests on Friday, worked well (at least from client perspective)
• Simultaneously with the challenge, LHCb started new Sprucing workflow, where jobs running at RAL download data directly from CERN (eospilot SE), that contributed to the OPN load (and LHCONE a bit, due to old WN gens).