Integrating Puppet and Gitolite to provide a novel solution for scalable system management at the MPPMU Tier2 centre

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

物理学;计算机科学

Delle Fratte, C.^1 ; Kennedy, J.A.^2 ; Kluth, S.^1 ; Mazzaferro, L.^1

Max Planck Institut für Physik, Föhringer Ring 6, München

80805, Germany^1

Rechenzentrum Garching, Giessenbachstraße 2, Garching

85748, Germany^2

关键词: Best practices;    Computing infrastructures;    Deployed systems;    Max Planck Institute;    Scalable systems;    Scaled-up;    Software deployment;    System administrators;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/664/5/052026/pdf
DOI  :  10.1088/1742-6596/664/5/052026

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

In a grid computing infrastructure tasks such as continuous upgrades, services installations and software deployments are part of an admins daily work. In such an environment tools to help with the management, provisioning and monitoring of the deployed systems and services have become crucial. As experiments such as the LHC increase in scale, the computing infrastructure also becomes larger and more complex. Moreover, today's admins increasingly work within teams that share responsibilities and tasks. Such a scaled up situation requires tools that not only simplify the workload on administrators but also enable them to work seamlessly in teams. In this paper will be presented our experience from managing the Max Planck Institute Tier2 using Puppet and Gitolite in a cooperative way to help the system administrator in their daily work. In addition to describing the Puppet-Gitolite system, best practices and customizations will also be shown.