In the upcoming LHC era, databases have become an essential part for the experiments collecting data from LHC, in order to safely store, and consistently retrieve, a wide amount of data, which are produced by different sources. In the CMS experiment at CERN, all this information is stored in ORACLE databases, located on several servers, both inside and outside the CERN network. The task of monitoring different databases is a crucial database administration issue, since different information may be required depending on different users' tasks that involve data transfer, inspection, planning and security. We present here a web application based on a Python web framework and Python modules for data mining purposes. To customize the GUI we record traces of user interactions that are used to build use-case models. In addition the application detects errors in database transactions, such as user errors, application failures, unexpected network shutdown or Structured Query Language (SQL) statement errors, and provides warning messages on the basis of the different categories of users, such as end-user performing database transactions, sub-detector manager, CMS central manager, developer and database administrator. Finally, in order to fullfill the requirements of the CMS experiment community, and to meet the new developments in many Web client tools, our application was further developed, and new features were deployed.
展开▼
