Speaker
Georgiana Lavinia Darlea
(Polytechnic University of Bucharest (RO))
Description
The ATLAS Online farm is a non-homogeneous cluster of more than 3000 PCs which run the data acquisition, trigger and control of the ATLAS detector. The systems are configured and monitored by a combination of open-source tools, such as Quattor and Nagios, and tools developed in-house, such as ConfDB.
We report on the ongoing introduction of new provisioning and configuration tools, Puppet and ConfDB v2 which are more flexible and allow automation for previously uncovered needs, and on the upgrade and integration of the monitoring and alerting tools, including the interfacing of these with the TDAQ Shifter Assistant software and their integration with configuration tools.
We discuss the selection of the tools and the assessment of their functionality and performance, and how they enabled the introduction of virtualization for selected services.
Author
Sergio Ballestrero
(University of Johannesburg (ZA))
Co-authors
Mr
Alexandr Zaytsev
(Budker Institute of Nuclear Physics (RU))
Diana Scannicchio
(University of California Irvine (US))
Franco Brasolin
(Universita e INFN (IT))
Georgiana Lavinia Darlea
(Polytechnic University of Bucharest (RO))
Irina Dumitru
(University of Bucharest (RO))
Liviu Valsan
(University of Bucharest (RO))
Matthew Shaun Twomey
(University of Washington (US))
