Interactive access and optimization of a CMS computing farm

CMS data taking at LHC is going on and its computing activity is increasing day after day. The need of efficient and reliable computing centers to run analysis and MC production is getting more and more pressing. The requirements, in terms of functionalities and performance, coming from the physics local community on a typical Tier2 for the everyday work will be described. A report will be given on the tests performed and the results achieved while attempting to improve a typical Tier2. Important improvements on the storage performance have been achieved through a patient optimization of several layers starting from the hardware to the storage manager software. In particular the performance and the functionalities obtained using two different storage manager solutions, dCache and StoRM/Lustre, will be presented. Particular attention has been taken regarding the batch system configuration and its scheduling policy to better fulfill the requirements of the different use cases. Several optimization tests have been performed to provide dynamic resources to the interactive work in order to accommodate both the batch typical Tier2 activity and the interactive activity as it is required in a Tier3 site. This result was obtained by means of an advanced configuration of the farm and in particular of the batch system used on the Tier2 batch farm. As a consequence of the optimization a great improvement in the efficiency of resource usage has been observed and a substantial reduction of the manpower required to provide an efficient and reliable interactive service has been achieved.