Finding associations in Grid monitoring data

Abstract

Error handling is a crucial task in infrastructures as complex as grids. Today, there are several monitoring tools which can be used to report failing grid jobs including corresponding error codes. However, the error codes do not always indicate the actual fault which originally caused the job failure. Human time and expertise is required to manually trace errors back to the real fault underlying an error. We perform Association Rule Mining on grid job monitoring data to automatically retrieve knowledge about the grid components' behaviour by taking dependencies between grid job characteristics into account. Therewith, problematic grid components are located automatically and this information -expressed by association rules -is visualised in a web interface. This work achieves a decrease in time for fault recovery and consequently yields an improvement of a grid's reliability.

Efficient Resubmission Strategies to Design Robust Grid Production Environments

“…Beyond simple traces collection, many recent works have focused on post-analysis of the data archived [10], [11], [12]. In the AMon monitoring system [10], most relevant information on jobs submitted to EGEE is filtered out of the traces in order to help users to monitor experiments yielding large amounts of jobs.…”

“…Their study is based on CONDOR and experiments have been made both on a local grid and on the Open Science Grid 3 . Maier et al [12] pointed out the fact that error codes returned by systems do not always properly identify the real cause of failure. They are using data mining techniques on EGEE traces in order to determine the root cause for faults.…”