Request deployment of XNAT via DataHub

[vemonet]

Describe the application

The XNAT data server is hosted on DataHub servers.

1. Download data from XNAT on the DSRI
2. Analyze on the DSRI (e.g. JupyterLab / RStudio)
3. Store back in the DSRI

Connect to XNAT REST API using https://xnat.readthedocs.io/en/latest/

Going further: XNAT containerized (not a current priority)

XNAT have documentation about this: https://wiki.xnat.org/container-service/container-scheduling-122978877.html
Note they mention that:

Disadvantage: Configuration is difficult.

This documentation for XNAT ML mention they are currently working on a Kubernetes package (which would be easier to deploy on the DSRI): https://wiki.xnat.org/ml/components-of-the-xnat-ml-release

Currently it can be deployed with docker-compose (which might be harder to deploy on Kubernetes): https://wiki.xnat.org/ml/devops-quick-reference-for-containerized-xnats

This branch describe full stack deployment: https://github.com/NrgXnat/xnat-docker-compose/tree/features/dependency-mgmt

Additionally, they have a presentation about using Docker containers for computing that could help understanding their system: https://www.youtube.com/watch?v=m9kFIEg8ZlM

It seems like XNAT has been built to work in containerized environment, but the Kubernetes deployment seems to still be a work in progress (not sure of how much work it needs)