You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Arnold: I looked at your Python images in detail, and I saw that you are using /home/jovyan/lab as the working directory where the notebooks and data is located. We mount the persistent volume to /home/jovyan/work. Can you switch to using that directory instead? And my guess is that if you set jovyan/work as the --notebook-dir the users should see all files and folders there? That would give them access to project-vol for persistent storage.
Rasool: For python labs I solved the data handling issue in a different way, so that the docker images already contain the required data for the lab. But it is very good to know for future how we should handle it.
Arnold: Hm but then if the container restarts nothing will be saved, they will have to start over. And we cannot control when containers restart, it is up to Kubernetes. It can happen if for example the container itself has issues (e.g. runs out of allocated CPU or there is an error) or another part of the cluster has issues so Kubernetes wants to change the node from which this particular container is running to balance the load better. For this reason we do not recommend having any data that's supposed to be persistent inside the container. If the users keep the data in the dedicated persistent volume folder (project-vol) then their work will be saved.
The text was updated successfully, but these errors were encountered:
Arnold: I looked at your Python images in detail, and I saw that you are using /home/jovyan/lab as the working directory where the notebooks and data is located. We mount the persistent volume to /home/jovyan/work. Can you switch to using that directory instead? And my guess is that if you set jovyan/work as the --notebook-dir the users should see all files and folders there? That would give them access to project-vol for persistent storage.
Rasool: For python labs I solved the data handling issue in a different way, so that the docker images already contain the required data for the lab. But it is very good to know for future how we should handle it.
Arnold: Hm but then if the container restarts nothing will be saved, they will have to start over. And we cannot control when containers restart, it is up to Kubernetes. It can happen if for example the container itself has issues (e.g. runs out of allocated CPU or there is an error) or another part of the cluster has issues so Kubernetes wants to change the node from which this particular container is running to balance the load better. For this reason we do not recommend having any data that's supposed to be persistent inside the container. If the users keep the data in the dedicated persistent volume folder (project-vol) then their work will be saved.
The text was updated successfully, but these errors were encountered: