Attendees : Lisa Paspalaki, Artur Darovic Gottmann, David Mason, Alessandra Forti, Benedikt Maier, Christophe Haen, David South, Diego Davila, Enrico Fattibene, Eric Fede, Eric Lancon, Haykuhi, James Walder, Julien Leduc, Katy Ellis, Mario Lassnig, Oliver Keeble, Shigeki, Xin Zhao
ATLAS plan : almost final
Details in slides
Tape read test will be using production requests, so will not read back the (fake) data written during DT test.
FTS throughput setting will be done by DDM experts, default to 3.5GB/s. Sites should explicitly tell DDM if a different number is preferred.
CMS plan : more details to come
see slide for updates
In the second half of DT week, will still run data export tests to T1s, using accumulated data from DAQ.
Sites ask about exact start time of the tests, and expected throughput etc. Please update ggdoc with more details as they become available.
Test requests will be submitted to sites in a similar way to the real production cases. Approval mechanism for sites to approve tape (write) requests.
LHCb plan :
Expect some small adjustments to the target rates.
Performed a test with CERN this week.
Not yet confirmed P8 will be joining the test, but enough data already accumulated on CERN disk from the CERN test.
Srm+https will be used.
LHCb tape read plot to be refilled on the central dashboard (LHCb didn't do tape read test last time)
Did a test with LHCb, resources in place to meet the target rates.
Gave ATLAS DDM the info of files to be cleaned up from previous tests.
Having dCache/GPFS performance issue right now, trying to fix before the CMS A-DT test next week. Will keep updated.
Reminder of downtime on March 15th (12h)
Migrating to the new tape system (Antares), which will be used in the tape challenge for both tape read and tape write tests.
VOs and sites : please keep ggdoc updated with any new updates and more details of the tests (timeline, data volume, etc). Let me know if you would like to have the editor role.