[Pycon] [new paper] "Marco Pernigotti" - Pythonic management of the ATLAS computing farm

[info a pycon.it]

Title: Pythonic management of the ATLAS computing farm
Duration: 45 (includes Q&A)
Q&A Session: 15
Language: en
Type: Talk

Abstract: 
The online farm of the ATLAS experiment at the LHC (CERN), consisting of ~4000 PCs with various characteristics, provides configuration and control of the detector and performs the collection, processing, selection, and conveyance of event data from the front-end electronics to mass storage.
Python has been chosen to write, improve and extend some of the tools used to manage the farm.
The monitoring system, based on Icinga2 and Ganglia, handles the status and health of each host using also Python scripts to complement this information querying the host data sources, such as SNMP and IPMI, to provide system and hardware health information.
An in-house configuration database is used to gather specific system administration information with data coming from different sources, such as the CERN central network database. Its main function is to ease the management of the network configurations (e.g. DHCP deployment) and the PXE booting. Python is used for a large fraction of its functionalities.
In this talk, we will give you an overview of the farm management, focusing on how we use Python to automate the day-to-day activities, and also to address more complex and dedicated tasks, such as securing the access to the experimental computing facility.

Tags: [u'packaging', u'REST', u'linux', u'integration', u'sysadmin', u'sql', u'networking', u'physics', u'postgres', u'mysql', u'monitoring', u'scientific-computing', u'SOAP']