ClusterAlive

21st International Conference on Computing in High Energy and Nuclear Physics

物理学;计算机科学

Caruso, G.^1 ; Arezzini, S.^1 ; Ciampa, A.^1 ; Formuso, A.^1 ; Mazzoni, E.^1

INFN Sezione di Pisa, Pisa, Italy^1

关键词: HPC clusters;    Maintenance operations;    Monitoring tools;    Pro-active approach;    Single platform;    Standard tools;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/664/9/092005/pdf
DOI  :  10.1088/1742-6596/664/9/092005

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

INFN-Pisa Scientific Computing Center is working from many years both in GRID and HPC computing. The monitoring and managing tools have been key components of the center's successful operation. The lessons learned from the use of standard tools, such as Ganglia, have been starting points for the development of new tools specific for our infrastructure. In this note we will illustrate the integration of many different monitoring tools in one single platform called ClusterAlive. Aim of ClusterAlive is to increase the HPC cluster performance and simplify maintenance operations, possibly in a proactive approach.