How to persist jupyterhub scheduling in an event of DR failure

JupyterHub
, , ,
1

How do we persist JupyterHub scheduling metadata if the setup is on two kubernetes clusters (PROD and DR). With a failover or DR-flip, how can we ensure that the metadata or scheduler is persisted.

  • Is there an existing mechanism to do this today

  • If not, then is there a way to do this ?

2

JupyterHub doesn’t have a scheduler, it only starts/stops singleuser servers such as JupyterLab. Are you using a Jupyter-server/lab extension such as
https://jupyter-scheduler.readthedocs.io/ ?