High-Energy Physics Applications on the Grid - Grid Computing: Infrastructure, Service, and Applications - page 438

FIGURE 17.8 The Site Efi ciency Dashboard application at work for VL-eMed. Job
attempts are identii ed and the grid failures are categorized and associated to computing
resources of the sites. The application permits the very quickly identii cation of a specii c
error pattern.
attempt is taken into account to test all available grid sites. The main
difference with the Job Monitor application ( Figures 17.5 through 17.7 )
is that in that case only the i nal execution of a job is considered. Site
Efi ciency permits very quick identii cation of error patterns, typically
connected to a site misconi guration. In the case of common errors the
tool points to a list of explanations/solutions that are accessible via the
drill-down functionality of the tool.
The future of this activity is that it will continue to grow. The availabil-
ity of more data allows more sophisticated studies. Very important devel-
opments are going on to propose a unii ed mechanism to exchange data
(e.g., using ActiveMQ http://activemq.apache.org/) and to better interface
with the different systems used in the grid computer center (e.g., using
Nagios http://www.nagios.org/). Here, the idea is to feedback monitoring
data (like grid efi ciency at a site) into the monitoring system of the site
itself, allowing seamless integration between local established operational
procedures and the newly available information.