TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost TRIUMF SITE REPORT Corrie Kost Update since Hepix Fall 2005

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Devolving Server Functions Windows print server cluster - 2 Dell PowerEdge SC1425 machines sharing external SCSI disk holding printers data Windows Print Server Windows Domain controller OLD NEW - 2 Dell PowerEdge SC primary & secondary Windows domain controllers NEW

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Update since Hepix Fall 2005

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Update since Hepix Fall 2005 Waiting for 10Gb/sec DWDM XFP

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Update since Hepix Fall km DWDM 64 wavelengths/fibre CH34=193.4THz nm ~ $10kUS each

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Update since Hepix Fall 2005

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Update since Hepix Fall 2005

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Servers / Data Centre GPS TIME TRPRINT CMMS TRWEB TGATE DOCUMENTS CONDORG TRSHARE TRMAIL TRSERV LCG WORKER NODES IBM CLUSTER 1GB TEMP TRSHARE KOPIODOC LCG STORAGE IBM ~ 2TB STORAGE TNT2K3 RH-FC-SL MIRROR TRWINDATA TRSWINAPPS WINPRINT1/2 PROMISE STORAGE

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Update since Hepix Fall 2005 TNT2K3 TRWINDATA TRSWINAPPS PROMISE STORAGE TRPRINT TRWEB TGATE CONDORG TRSHARE TRMAIL TRSERV DOCUMENTS CMMS GPS TIME WINPRINT1 WINPRINT2

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost TRIUMF-CERN ATLAS Lightpath - International Grid Testbed (CA*Net IGT) Equipment Amanda Backup ATLAS Worker nodes: (evaluation units) Blades, Dual/Dual 64-bit 3GHz Xeons 4 GB RAM, 80GB Sata VOBOX: 2GB, 3GHz 64-bit Xeon, 2*160GB SATA LFC: 2GB, 3GHz 64-bit Xeon, 2*160GB SATA FTS: 2GB, 3GHz 64-bit Xeon, 3*73GB SCSI SRM Head node 2GB, 64-bit Opteron 2*232GB RAID1 sc1-sc3 dCache Storage Elements 2GB, 3GHz 64-bit Xeon, 8*232GB RAID5 2 SDLT 160GB drives / 26 Cart 2 SDLT 300GB drives / 26 Cart TIER1 prototype (Service Challenge)

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost ATLAS/CERN TRIUMF

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost SiteDisk-DiskDisk-Tape ASGC10075 TRIUMF50 BNL20075 FNAL20075 NDGF50 PIC60*60 RAL15075 SARA15075 IN2P FZK20075 CNAF20075 Tier0 Tier1 Tests Apr 3 30 Any MB/sec rates below 90% nominal needs explanation and compensation in days following. Maintain rates unattended over Easter weekend (April 14-16) Tape tests April Experiment-driven transfers April * The nominal rate for PIC is 100MB/s, but will be limited by the WAN until ~November

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost ATLAS SC4 Plans – Extracted from Mumbai Workshop 17 Feb/2006 (1) March-April (pre-SC4): 3-4 weeks in for internal Tier-0 tests (Phase 0) April-May (pre-SC4): Tests of distributed operations on a small testbed (the pre-production system) Last 3 weeks of June: Tier-0 test (Phase 1) with data distribution to Tier-1s (720MB/s + full ESD to BNL) 3 weeks in July: Distributed processing tests (Part 1) 2 weeks in July-August: Distributed analysis tests (Part 1) 3-4 weeks in September-October: Tier-0 test (Phase 2) with data to Tier-2s 3 weeks in October: Distributed processing tests (Part 2) 3-4 weeks in November: Distributed analysis tests (Part 2) (1) Update from Hepix Fall 2005

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Repeated reads on same set of (typically 16) files (at ~ 600MB/sec) – during ~ 150 days ~ 7 PB (total since started ~13PB to March 30 – no reboot for 134 days)

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Repeated reads on same set of (typically 16) files (at ~ 600MB/sec) – during ~ 150 days ~ 7 PB (total since started ~13PB to March 30 – no reboot for 134 days)

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Keeping it Cool Central Computing Room isolation fixed – Combined two 11-Ton air-conditioners to even out load Adding heating coil to improve stability Blades for Atlas! – 30% less heat, 20% less TCO 100 W/sq-ft 200 W/sq-ft 400 W/sq-ft means cooling costs are a significant cost factor Note: Electrical/Cooling costs estimated at $Can150k/yr Water cooled systems for (multicore/multicpu) blade systems?

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Keeping it Cool2 HP offers Modular Cooling System (MCS) Used when rack > 10-15Kw US$30K Chilled (5-10C) water Max load 30Kw/rack (17GPM / 5C 20C air) Water cannot reach servers Door open? - Cold air out front, hot out back Significantly less noise with doors closed HWD 1999x909x1295mm (79x36x51) 513Kg/1130lbs (empty) Not certified for Seismic or Zone 4

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost Amanda Backup at TRIUMF Details by Steve McDonald Thursday ~ 4:30pm

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost End of Presentation Extra Slides on SC4 plans for reference…

TRIUMF Site Report for HEPiX, CASPUR, April 3-7, 2006 – Corrie Kost fts: FTS Server FTS= File Transfer Service homepage: Oracle database used 64-bit Intel Xeon 3 GHz 73 GB SCSI disks (3) 2 GB RAM IBM 4560-SLX Tape Library attached (will have 2 SDLT-II drives attached when they arrive, probably next week) SDLT-II does 300 GB native, 600 compressed. Running SL bit lfc: LFC Server LFC= LCG File Catalog info page: MySQL database used 64-bit Intel Xeon 3 GHz 160 GB SATA disks (2), software raid-1 2 GB RAM Running SL bit vobox: VO Box Virtual Organization Box info page: 64-bit Intel Xeon 3 GHz 160 GB SATA disks (2), software raid-1 2 GB RAM Running SL bit sc1-sc3: dCache Storage Elements 64-bit Intel Xeons 3 GHz 3ware Raid Controller 8x 232 GB disks in H/W Raid-5 giving 1.8 TB storage 2 GB RAM Running SL bit sc4: SRM endpoint, dCache Admin node and Storage Element 64-bit Opteron 246 3ware Raid Controller 2x 232 GB disks in H/W Raid-1 giving 250 GB storage 2 GB RAM Running SL bit IBM 4560-SLX Tape Library attached. We are moving both SDLT-I drives to this unit. SDLT-I does 160 GB native, 300 compressed. Service Challenge Servers Details

Dario Barberis: ATLAS SC4 Plans WLCG SC4 Workshop - Mumbai, 12 February ATLAS SC4 Tests l Complete Tier-0 test nInternal data transfer from Event Filter farm to Castor disk pool, Castor tape, CPU farm nCalibration loop and handling of conditions data Including distribution of conditions data to Tier-1s (and Tier-2s) nTransfer of RAW, ESD, AOD and TAG data to Tier-1s nTransfer of AOD and TAG data to Tier-2s nData and dataset registration in DB (add meta-data information to meta-data DB) l Distributed production nFull simulation chain run at Tier-2s (and Tier-1s) Data distribution to Tier-1s, other Tier-2s and CAF nReprocessing raw data at Tier-1s Data distribution to other Tier-1s, Tier-2s and CAF l Distributed analysis nRandom job submission accessing data at Tier-1s (some) and Tier-2s (mostly) nTests of performance of job submission, distribution and output retrieval

Dario Barberis: ATLAS SC4 Plans WLCG SC4 Workshop - Mumbai, 12 February ATLAS SC4 Plans (1) l Tier-0 data flow tests: nPhase 0: 3-4 weeks in March-April for internal Tier-0 tests Explore limitations of current setup Run real algorithmic code Establish infrastructure for calib/align loop and conditions DB access Study models for event streaming and file merging Get input from SFO simulator placed at Point 1 (ATLAS pit) Implement system monitoring infrastructure nPhase 1: last 3 weeks of June with data distribution to Tier-1s Run integrated data flow tests using the SC4 infrastructure for data distribution Send AODs to (at least) a few Tier-2s Automatic operation for O(1 week) First version of shifters interface tools Treatment of error conditions nPhase 2: 3-4 weeks in September-October Extend data distribution to all (most) Tier-2s Use 3D tools to distribute calibration data l The ATLAS TDAQ Large Scale Test in October-November prevents further Tier-0 tests in 2006… n… but is not incompatible with other distributed operations No external data transfer during this phase(?)

Dario Barberis: ATLAS SC4 Plans WLCG SC4 Workshop - Mumbai, 12 February ATLAS SC4 Plans (2) l ATLAS CSC includes continuous distributed simulation productions: nWe will continue running distributed simulation productions all the time Using all Grid computing resources we have available for ATLAS The aim is to produce ~2M fully simulated (and reconstructed) events/week from April onwards, both for physics users and to build the datasets for later tests lWe can currently manage ~1M events/week; ramping up gradually l SC4: distributed reprocessing tests: nTest of the computing model using the SC4 data management infrastructure Needs file transfer capabilities between Tier-1s and back to CERN CAF Also distribution of conditions data to Tier-1s (3D) Storage management is also an issue nCould use 3 weeks in July and 3 weeks in October l SC4: distributed simulation intensive tests: nOnce reprocessing tests are OK, we can use the same infrastructure to implement our computing model for simulation productions As they would use the same setup both from our ProdSys and the SC4 side nFirst separately, then concurrently

Dario Barberis: ATLAS SC4 Plans WLCG SC4 Workshop - Mumbai, 12 February ATLAS SC4 Plans (3) l Distributed analysis tests: nRandom job submission accessing data at Tier-1s (some) and Tier-2s (mostly) Generate groups of jobs and simulate analysis job submission by users at home sites Direct jobs needing only AODs as input to Tier-2s Direct jobs needing ESDs or RAW as input to Tier-1s Make preferential use of ESD and RAW samples available on disk at Tier-2s Tests of performance of job submission, distribution and output retrieval Test job priority and site policy schemes for many user groups and roles Distributed data and dataset discovery and access through metadata, tags, data catalogues. nNeed same SC4 infrastructure as needed by distributed productions Storage of job outputs for private or group-level analysis may be an issue nTests can be run during Q First a couple of weeks in July-August (after distributed production tests) Then another longer period of 3-4 weeks in November