--+ Tier1 and Tier-2 Weekly Report 2015-12-16
General News
- ADC and ATLAS Computing News [20151216]
- ADC Weekly meeting
- held on Wednesday this week due to LHC Seminars
- MC Production
- MC15b - bulk sample of 770M almost done, another ~150M going on now
- Xmas Break
- Expect O(500M) fullsim - still being prepared
- Sherpa 2.2 and Alpgen V+jets > 400M
- Sherpa ttbar 30M
- Powheg dibson 30M
- At least 100M will be ready (probably 200M, hopefully more)
- Work on task submission up to Mon 21 Dec, then best effort Dec 22-Jan 3
- Will send update at end of week
- Derivation Production during break
- Built a bug-fix cache for few derivations, AtlasDerivation-20.1.9.2, being distributed on the grid.
- Will do “data” reprocessing with this cache for EXOT2, EXOT7, HIGG4* and HIGG3* derivations, there will be also some MC processing. Validation tasks are about to be submitted.
- AtlasDerivation-20.1.9.3 has just been requested by TOP and SM groups for additional variables (small change). Plan to build tomorrow. MC done by group contacts (small production)
- Reprocessing cache - AtlasDerivation-20.7.3.5 requested for build today. Requests are still being collected from groups, so far we have: HIGG1, HIGG2, HIGG4, HIGG5, HIGG6, SUSY, EGAM, TOP, JETM.
- Will run derivations on the reprocessed samples unless otherwise is planned
- Bulk reprocessing
- Page describing the full 2015 data reprocessing (112 runs), scheduled for Dec 2015 / Jan 2016. Scheduled to start today
- https://twiki.cern.ch/twiki/bin/viewauth/Atlas/December2015Run2Reprocessing
- Release signed off on yesterday. Tests run overnight. Data available for checking
- Merges will be forced to run at T1's
- Still force 25ns by hand
- Jobs still running 21-24GB per job, 2.5-3GB/core. If use 2500MB/core then will eliminate a lot of sites in brokerage.
- Will be a little but more testing of tuning brokerage before submit full bulk to ensure hit the most sites possible
- Still some validation on express stream necessary so may have to abort (a bug found yesterday) and continue next year
- ADC Coverage over break
- A normal level of Grid activity is foreseen over the holidays
- All IT services and ADC experts are under best-effort support from 19 Dec to 3 Jan
- No CRC but there is quite good ADCoS coverage
- No ADC morning meetings
- If central services go down there is no guarantee anyone can restart them
- Sites can expect GGUS tickets in case of problems, we understand that support levels vary
- If a serious problem comes up please follow ADC emergency procedures
- In case of security issue contact atlas-adc-csirt@cern.ch
- Experts should log significant events in the Daily Reports twiki
- Experience tells us that best-effort is usually good enough
- See you on 5th Jan for the first ADC weekly of 2016
- CRC/ADCoS report
- Running 200k slots most of last week
- Last week was taking Heavy Ion physics data
- LHC/ATLAS now in technical stop, next Run Meeting on January 19th 9:30am
- FTS-BNL had CRL problem preventing transfer requests
- NGDF dCache upgrade, rumours of nasty bug so perhaps some comment in WLCG meeting should be requested
- Staging from Tape 1.7PB mostly successful, some bringonline timeouts but performance OK
- Decision on bulk reprocessing...
- Electricity outage tomorrow (Thurs.) from 6am for 10min. Service check first thing in morning.
- Sim@P1 Profiling Test
- see slides
- Short summary of ADC TIM meeting in Stiges Dec 2-4
- see slides
- CA Cloud Squad Actions
- Open Issues and Ongoing Discussions
- WLCG and General Computing News
- Security News
- gLite/EMI News
- ROC_Canada News
ABCD Determination and Monitoring
SSB Results
SAM Results (based on VO tests)
APEL Accounting
- ACTION ITEM FOR SITES: Sites please regularly check and confirm that your historical APEL data has been published and looks correct when comparing with batch system utilization
APEL Synchronization Test Links
Production and Usage Status
--
Leslie Groer - 2015-12-16
Comments