https://indico.cern.ch/event/307176/
minutes
------>>>>> MC and Group production status and news 2
Andrej: 400M FS is single core mc11
Emmanoueli: the 400M FS is full chain, new MC, not only simul.
good to have estimate (even with the disclaimer!) ! can we try also the diskspace? Jose said he will.
------>>>>> DPD prod:
need to finalize the cache, failed today.
2 trains, one slow (not really data reduction but similar to this) one fast (similar how it will be the xAOD) .
mainly one/two week of workload. but important to get it done when is submitted.
------>>>>> Rucio migration - status update
DRChecker: we will evaluate with AA if it's easy to migrate or if we just drop it. we have to update the info in the monitoring ticket.
https://its.cern.ch/jira/browse/ADCMONITOR-316
------>>>>> Pilot
FAX are calculated in popularity , but are they calculated properly? the storage is remote, is not the "local" storage… to be checked by Paul.
--> agreed Paul will release the pilot today.
------>>>>> Exporting system libraries for ATLAS on special facilities
slide3:
1option what's the difference respect to today?
2option 3 bullet: which CVMFS release? will it be fixed?
3option: will you link to specific versions?
cernVM with all the lib installed, the best option would be to re-use this approach with parrot.
Andrej: slide6 very heavy solutions, in sys like ubuntu. slide4 is the direction to go.
AleDS: indeed. we all agree.
AleDS will send email to ADC dev to discuss this in a dev meeting to be ready for a proposal for the SW week.
------>>>>> ADCoS report
data loss in MPPMU:
instructions to clean dark data: we will make an email and a presentation next week to re-explain
- BNL DDM TEST needs to be updated in the DQ2 DB
------>>>>> Reprocessing
Alastair: the overlay is on data or mc? most probably data, but Alastair will ask Bruno/Till
------>>>>>
AOB:
IN2P3-CC noticed too much pressure from pilot factory when no activity-- to be discussed with Jose/John/Peter .
Ranking of sites: could it be that if a site is not running because of lack of jobs then the ones that are running are preferred to run also the "few" jobs in the pipeline/submitted just after?
not clear from https://twiki.cern.ch/twiki/bin/view/PanDA/PandaBrokerage . We will ask Tadashi
DAST: Farida many complains from users because we can access log files.
Ale: could it be that it was the blue pandamon? the blue doesn't work, the pink does… if not, please send us an email.
Frontier: intervention on both BNL and IN2P3-CC the 6th. IN2P3-CC can postpone, thanks.