Modernizing the Monitoring of Mass Storage systems

Project Specification The work is a project of IT-ST group, dealing with mass storage systems at CERN and responsible for more that 150PB of data stored on disks and tapes. The task consists of modernizing the logging system of one of of the storage systems namely CASTOR, in particular by integrating it into common monitoring framework provided by the IT department and by moving to new display technologies. The main tools used are Hadoop/HBase, Spark, Python, and web technologies like Django, Graphite and Grafana. Abstract The LHC has entered a new regime by producing more than 25PB of data since the beginning of this year. Thus the storage and monitoring of log messages is of crucial importance for the researchers. The CERN Advanced STORage system (CASTOR) is a storage system developed at CERN and is used by LHC experiments to store physics data files. Monitoring of this complicated distributed system is mandatory in order to assure it’s stable operation and improve its future performance. This project presents an infrastructure meant to collect and display monitoring information about the system by the means of Graphite and Grafana data visualization platform