I updated my local renku, then ran renku init in a folder and pushed the folder to gitlab.renkulab.io. After logging into renkulab, I am presented with recently visited projects, but no link to find my other projects. Only after manually adding my username to https://renkulab.io/ I am presented with the option to see all my projects. Navigated to the newly created project (soil_CO2) but there, I was told that the project is very outdated and automatic update is not possible. Could it be that the renku version on renkulab is behind the one on pypi? Locally, I am on 2.9.4, but this is what I see on renkulab:
Renku version
(2.9.2 avaliable)
The project is very outdated. An upgrade is necessary to support the project from the RenkuLab UI.
This happens because the underlying Renku project metadata is still on version 1 while the latest version is 10.
Automatic updates are not available; please use the migrate command in a session. More info
Template version
Error
Older project versions might not expose some template details and trigger an error. This usually disappears after updating the Renku version.
Does anyone have an idea how to fix this and avoid similar problems in the future?
Hello @schymans , I couldn’t check your project, but I suspect your local template is outdated.
The renku version is disentangled from the template version. The Renku template repo is fetched and stored locally. You can check it by using renku template ls. Try pick a template and see the version, E.G. renku template show python-minimal. For reference, this is the repo GitHub - SwissDataScienceCenter/renku-project-template: A repository for default Renku project templates.
In older versions, the Dockerfile was very different and we stored Renku metadata in a way that is not supported by the latest version of the renku service, hence the “very outdated” message.
To fix that, you might need to manually update the project from the console. In general, you can update the local template by using renku template update.
In the future, you can also reference a specific repo/version when creating a new project (see renku init --help for all the parameters).
P.S. Keep in mind that we are working on Renku 2.0 and it will have a fundamentally different way of creating projects. That’s already available as an early access at Reproducible Data Science | Open Research | Renku
You might want to check it out for new projects.