Speakers
Description
INFN with its DATAcloud infrastructure provides a scalable network of federated cloud sites. The ICSC project (National Research Center in High-Performance Computing, Big Data, and Quantum Computing), funded by the PNRR (National Recovery and Resilience Plan), was established to drive R&D efforts focused on advancing high-performance computing, simulations, and big data analytics innovation. As part of the ICSC initiative, the INFN Milano computing center has expanded its capacity by deploying a bare metal Kubernetes cluster.
This contribution describes how to deploy an HTCondor cluster on Kubernetes to run jobs in a Container Universe, both with Docker and Apptainer as container runtimes. This has been achieved via the virtualization of the Condor execute node in a Kubernetes Deployment, to add scalability. These worker nodes have been joined to the existing baremetal HTCondor cluster.
As a representative use case, a workload generating production events for the LHC ATLAS experiment has been tested with three configurations: Apptainer jobs on baremetal, Apptainer and Docker jobs inside Kubernetes. The added virtualization layer has not hindered the job's performance.
| Desired slot length | 10 minutes |
|---|---|
| Speaker release | Yes |