Supervision des clusters de surveillance et de traitement de données dans l'expérience LHCb

The LHC at CERN is the project that looks for answers to different questions that the physics has made itself, such as the presence of the Higgs boson, the origin of mass in particles, the existence of supersymmetry, and many others. In the LHCb experiment, analysis corresponding to the asymmetry between matter ans antimatter will be made.The collisions will produce ephemeral information that will be obtained by a readout system that will send the data to the processing and monitoring farms. Once preprocessed, the relevant information will selected to be permanently stored. My work was to develop tools that permit the supervision of the monitoring and processing systems, find their errors and notify the workgroup. The processing farm has 50 clusters with approximately 1000-2000 nodes and the monitoring farm has a cluster with 50 nodes. The applications I made are the Tasksupervisor and the ClusterMonitor.The TaskSupervisor finds the errors en each node of a cluster and publishes the results while the ClusterMonitor collects the information of the TaskSupervisors and concentrates them in a single place.