Bonn 2020 – scientific programme

The DPG Spring Meeting in Bonn had to be cancelled! Read more ...

Parts | Days | Selection | Search | Updates | Downloads | Help

T: Fachverband Teilchenphysik

T 25: Grid Computing

T 25.2: Talk

Tuesday, March 31, 2020, 17:15–17:30, H-HS V

XCache for ATLAS production — •Nikolai Hartmann, Günter Duckeck, Rodney Walker, and Christoph Anton Mitterer — Ludwig-Maximilians-Universität München

High-level caching provides a promising and lightweight storage solution that can replace or complement permanent storage clusters. The XRootD project includes a proxy caching server (XCache). Integration with the ATLAS data management system Rucio is added through a plugin.

We have setup such an XCache server at the LRZ-LMU Tier-2 and tested several ATLAS workflows. In particular, we ran an ATLAS production queue that processes all input via the XCache server. In our initial setup several problems showed up when operated longer-term under heavy load in production conditions. The latest XCache version proved to be more stable. We also evaluated different hardware configurations of the caching server, comparing the performance of individual disks and RAID 6.