Abstract
The Grid resources are distributed among hundreds of independent Grid sites, requiring a higher level Workload Management System (WMS) to be used efficiently. Pilot jobs have been used for this purpose by many communities, bringing increased reliability, global fair share and just in time resource matching. glideinWMS is a WMS based on the Condor glidein concept, i.e. a regular Condor pool, with the Condor daemons (startds) being started by pilot jobs, and real jobs being vanilla, standard or MPI universe jobs. The glideinWMS is composed of a set of Glidein Factories, handling the submission of pilot jobs to a set of Grid sites, and a set of VO Frontends, requesting pilot submission based on the status of user jobs. This paper contains the structural overview of glideinWMS as well as a detailed description of the current implementation and the current scalability limits.
Export citation and abstract BibTeX RIS