Events

Nemo will be down for maintenance on 13.03.2018. We will make good use of this externally triggered downtime by applying all pending software upgrades to NEMO. Job scheduling will be suspended during the maintenance. Login and access to your files will not be possible during the maintenance.

You can submit jobs until the downtime starts (13.03.2018, 06:00 AM). Please note that jobs projected to finish after the downtime starts will be blocked. Your jobs should run after the downtime is complete.

Target audience: Scientists (students, PhD students, Postdocs) who need well-defined software environments on the bwForCluster NEMO or in the bwCloud that they can control themselves.

Principal goals of the workshop: 

• Overall: Learn how to create, use and maintain, and share virtual research environments (VRE) on NEMO or bwCloud.
• Clarify distinction between the ability to *distribute/reconstruct* a VRE (ViCE) versus the ability to reference/archive/re-execute of VRE (CiTAR)
• Usage of "conda" for software package management in user space (i.e. no admin required)
• Create and run singularity containers on HPC for maximum independence from given HPC node operating system and software environment
• Handle specific or conflicting Python requirements in separate VREs
• Share VRE images or VRE recipes between users

RSVP not required, but appreciated for planning purposes. If you are interested in attending or have any questions, please send an email to bernd.wiebelt@rz.uni-freiburg.de

The first day of the workshop started with presentations and live demos by Jonathan Bauer on "Current developements in bwLehrpool" and the convergence of VM building and PXE booting in bwLehrpool context as well as the provisioning of the major container technologies Docker and Singularity (Beta) on that platform. Björn Grüning of Prof. Backofen's group in Freiburg and associated to the de.NBI project (German Network for Bioinformatics Infrastructure) demonstrated practical computational reproducibility in the life sciences. The session was followed by Oleg Stobbe from the CiTAR projecton current developments in VRE frameworks. The first session got closed by Christopher Hauser from Ulm University by a presentation of ViCE Registry project.

The afternoon session was dedicated to a hacking sessions on ViCE Registry, ways of VRE building (Container/VM) and the access to research data (Research Data Management). The second day clarified the distinction between provisioning of VRE (ViCE) and re-usability of VRE (CiTAR) as well as defined some (lightweight) guidelines for the creation of VRE and VRE workflows. Several existing usecases (HEP, computational chemistry, bioinformatics, bioconda for HPC) in different disciplines got discussed and evaluated for further uptake. Additionally, external dependencies (services, hardware, data sources and sinks, ...) of VREs were summarized. The workshop produced a good overview on the current state of developments in that field.

This includes integration into existing infrastructure and testing. It is not clear yet how long this will take. Please check here for status updates.

Please check here for status updates.

Registration is not yet open. However, please save the date!