Report about the ATLAS TDAQ Data Challenge Support by IT-FIO/FS

From 6th of June 2005 till 22nd of July 2005, ATLAS has been running its TDAQ Large Scale Tests on part of the IT LXBATCH cluster at CERN. The tests were successfully terminated after 6 weeks of running on a stepwise increasing farm size from 50 to up to 700 dual CPU nodes. Functionality and verification tests of the ATLAS DAQ and HLT parts of the system were carried out and selected performance measurements were taken. Individual sub-system tests were performed for the online system, the Event Building system, and for the Level2 Trigger system and Event Filter system. On the IT side, the cluster was configured and maintained by the FIO/FS section according to the ATLAS specific requirements, with the help of the ELFms tools. Requirements were on many aspects: hardware specifications, network layout, operating system and specific ATLAS software deployment and upgrades, machine monitoring and problem solving. This document describes how the ATLAS requirements were satisfied by IT.