ADC Weekly

Tuesday 24 Jan 2012, 15:40 → 17:40 Europe/Zurich

I Ueda (Department of Particle Physics-University of Tokyo)

Meeting Access Information: ATLAS ADC meeting room

• Meeting URL : http://evo.caltech.edu/evoNext/koala.jnlp?meeting=eneueMvnveasaDa2In
• Phone Bridge ID: 55767  (Phone numbers for phone bridge : http://evo.caltech.edu/evoGate/telephone.jsp)

+New share for production/analysis at T1s
- Implemented everywhere (Correction for CNAF to be included)

+ADC downtime

* General introduction
- Apparently, offline did not propagate to all pilot factories. Looks like new
jobs started or kept in 'submitted' state in TW/ND after 7:00 (cloud set offline). Should be followed-up.

* Database

Situation is better : Will not rollback to 10g
Security patch just came and will be done at the same time as the parameter change.

Upgrade client 10g to 11g : 10g client should be OK and no 11g will be provided soon

DDM : Can increase the number of sockets but to be discussed.


* Panda : Jobs finishing during downtime failed. Heartbeat was retried for running jobs.

* PandaMon : Caching functionnality was not tested in december.

+SL6:
 * CERN : T0 runs over different clusters -> big migration to SL6 (probably 2013)
 * DDM for Nosql: Python 2.6 for SLC6

+ DDM Ops :
 - List of T2Ds : Rules updated relying on statistic over 4 weeks
 - Next LFC migration
 - SSB sonar page : Should be updated to match the new rules
 - Cloud should have a look at the T2D classification. New classification will be included for next ABCD grouping

15:40 → 15:45 Hot topics

15:40 News from the run coordination

Speaker: Jaroslava Schovancova (Acad. of Sciences of the Czech Rep. (CZ))

15:41 WLCG-TEG-OPS F2F meeting

agenda https://indico.cern.ch/conferenceDisplay.py?confId=161833

15:43 T1 computing shares

CREM decision: /atlas/Role=production (95) : /atlas/Role=pilot (5) instead of historical 80:20 The ATLAS "Computing Model" tells 80% of T1 computing resources to the central production and 20% to "group+user" activities. The current system uses Role=production for central+group production, and Role=pilot for group+user analysis, i.e. Role=production needs more share. We need to complete the ongoing MC production well in advance before the winter conferences. without them, not much sense to run analysis jobs. Extra allocations made at T1s (more than 95% of the pledge) are very much appreciated.

more information https://indico.cern.ch/materialDisplay.py?contribId=6&sessionId=0&materialId=minutes&confId=172126

15:45 → 16:20 ADCR+ATLR downtime 17 Jan - post-mortem

slides ADCweekly-20120124-iueda.pdf

15:55 DBA

from DB point of view, especially load per node

Speaker: Gancho Dimitrov (CERN)

Slides Gancho_ADC_weekly_24thJan2012.pdf Gancho_ADC_weekly_24thJan2012.ppt

16:05 Panda server

it kept running, so would be nice to see how it went with automatic recovery with retries

Speaker: Tadashi Maeno (Brookhaven National Laboratory (US))

Slides adcruppsrv.pdf

16:10 Panda monitor

Speaker: Dr Valeri Fayn (Brookhaven National Laboratory (US))

text From Sergey: The CERN IT haven't plans to make new "oracle-instantclient" at all at least in coming years. The Oracle 11g is full compartible with 10 and CERN IT didn't see reason for update oracle client for 11g.

16:20 → 16:30 AMOD report

Speakers: Alessandro Di Girolamo (CERN), Jaroslava Schovancova (Acad. of Sciences of the Czech Rep. (CZ)), Stephane Jezequel (Centre National de la Recherche Scientifique (FR))

16:30 → 16:50 ATLAS and SL6

Convener: Dr Alexei Klimentov (Brookhaven National Laboratory (US))

16:30 ATLAS software and SL6

Speaker: Dr Alexei Klimentov (Brookhaven National Laboratory (US))

Slides ADCweekly_Jan24_2012_SLC6.pdf ADCweekly_Jan24_2012_SLC6.pptx

16:40 Points for Tier-0

Speakers: Alessandro Di Girolamo (CERN), Armin Nairz (CERN)

Slides ADCWeekly_Tier0_25-1-2012.pdf

16:45 SL5 environment on SL6

Presentation at the GDB January 2011, "Mixing SL5 and SL6 with ‘chroot’" by Brian Bockelman

more information https://indico.cern.ch/sessionDisplay.py?sessionId=5&confId=155064

16:50 → 17:05 DDM Operations

Speaker: Cedric Serfon (Ludwig-Maximilians-Univ. Muenchen (DE))

Slides 20120124DDMreport.pdf

A question on the EVO (not noticed during the meeting):

• Leslie Groer: Does a T2D become disqualified if it starts failing the criteria?
• we aim so

17:05 → 17:15 Site availability for analysis and data acceptance status

Speaker: I Ueda (University of Tokyo (JP))

more information https://twiki.cern.ch/twiki/bin/viewauth/Atlas/DataDistributionSharesForT2#Shares_2012_Jan

17:15 → 17:20 AOB