Weekly operations meeting for sites doing DUNE computing
DUNE computing ops meeting
6/28/2021
Tickets —SURFSARA (GGUS 151908). The test jobs work production jobs run but don’t finish.
QMUL (151364) haven’t made progress as yet
CERN 151401—no recent progress
Will discuss at cern / it protodune meeting
NP02 running
No news
Working on DAQ infrastructure—coldbox will be in EHN1.
NP04 —no news
Production news
Nothing out of the ordinary
Still haven’t heard anything from UNICAMP
Some weirdness at Nebraska also
NERSC — are we done?
Not yet—some are taking much longer and we don’t know why
Regular stuff—is FTS doing the right thing? We think yes
Discussion of data tiers. Reco-recalibrated tier
Data management ops
Ticket over the weekend.. many thousands of small files getting written to tape
Default dune file family is SFA now.
In production we control, should be putting in files on SFA or not
For users, have to watch
Discussions project.py is putting garbagoose in the metadata and we have to deal with it sometime.
Need to set a meeting with J. Hendry and co.
Site report
GridPP
Everything broadly OK
Fermilab—persistent DCache remains almost full
BNL
DUNE allocation temporarily lower at BNL due to STAR
Will DUNE have access to Perlmutter GPU’s at the beginning? Yes
CC-IN2P3
Nothing
FZU—petr no report