Operational issues:

• Upload failures on RAL (WNs) -> CERN channel
  • no updates.
• Issues on WNs last Sunday
  • Both Job and transfer failures were present;
  • The most meaningfull error message came from one of the jobs -- runtime/cgo: pthread_create failed: Resource temporarily unavailable;
  • The issue disappeared on Monday;
  • Might be some PID limit excess, Tom would like to upgrade it to the newer version (after the break, obviously);
  • Ticket GSTSM-277 is opened.
• LHCb pilots are killed at RAL due to memory excess
  
  • That looks suspicious since all types of jobs are suffering from it
  • Could it be some wrong memory accounting?
• Some WNs have >0 values in cvmfs IO error counter, it causes warnings in ETF tests.

Other:

• Writeable WN GW sandbox deployed to a few WNs on the preprod farm last week
  
  • Did not work
  • A few bugs found, to be fixed (in the new year)