| |
For each control system of the Large Hadron Collider (LHC) experiments, there will be many processes spread over many computers. All together, they will form a PVSS distributed system with around 150 computers organized in a hierarchical fashion. A centralized tool has been developed for supervising, error identification and troubleshooting in such a large system. A quick response to abnormal situations will be crucial to maximize the physics usage. The tool gathers data from all the systems via several paths (e.g., process monitors, internal database) and, after some processing, presents it in different views: hierarchy of systems, host view and process view. The relations between the views are added to help to understand complex problems that involve more than one system. It is also possible to filter the information presented to the shift operator according to several criteria (e.g. node, process type, process state). Alarms are raised when undesired situations are found. The data gathered is stored in the historical archive for further analysis. Extensions of the tool are under development to integrate information coming from other sources (e.g., operating system, hardware).
|
|