Fighting against the instability : Debian Science at the synchrotron SOLEIL
summary
Facility produces about 3 PB/year of raw synchrotron data across 29 beamlines, serves ~4,000 visitors, and faces a dependency forest: 54 core apps pull ~2,000 packages and ~16 GB of system install. The chosen remedy is Debian-based distributions with in-house packaging, a web VM service (Darts / QMU web desktop), JupyterHub, and contributions of ~400 packages pushed into Debian so the stack gains a stable two-year upgrade cycle and practical reproducibility.
AI-written summary · about
chapters (8)
- 0:00SOLEIL synchrotron and Python
- 3:14Scientific data challenges
- 8:40SOLEIL’s computing limits
- 11:06Instability in scientific software
- 16:02Traditional packaging failures
- 20:23Debian Science for stability
- 24:42Data sovereignty and open-source
- 25:24Q&A
related talks
27mDevSecOps - Why Debian repositories are Mission-Critical!
27mE4S: Extreme-scale Scientific Software Stack
31mScience Reliability Engineering for High Performance Computing
49mDeveloping Research Infrastructures the DevOps way
42mThe Long Road to Reproducible Builds
46mThe Reproducible Build Zoo
16mlightning talk, CentOS Dojo, DevConf.us
44mEconomics of Volunteer Labor: Three stories from Debian