I installed TLJH on my lab’s main computer and it worked like charm (my students loved it!). It’s fantastic ! However, as always in this life, after some updates, installations, errors and successes, I see that I probably need to delete everything and start from scratch.
After some upgrades, my TLJH showed “Spawn failed”, and after an attempt to upgrade conda, and others, I now have a systematic error with “The ‘ruamel.yaml==0.15.*’ distribution was not found”, and even with an attempt at internal installation via “/opt/tljh/hub/bin/python3 -m pip install --upgrade”, I started getting lost in one error after another (eg: legacy-install-failure).
Anyway, I think the best solution is to reset and start over.
Any tutorials ready for this?
Thank you in advance.
If TLJH is the only thing running on your server the easiest option is probably to re-install the OS, or if it’s a virtual machine delete it and spin up a new one.
This is a bit too brutal (let’s say…) but I’m tending to do just that, erase and reinstall the entire OS. As I’ve migrated almost all my operating to the cloud, I think it’s time to ghost my HD just in case it gets lost. In that case, I should opt for it myself - reinstall the OS from scratch.
I’ve been reading the documentation you indicated but I confess that sometimes it feels like I’m going around in circles. An “uninstall” option would really be useful in this case.
Thanks for your comments.
TLJH isn’t a self-contained application, so an automated uninstall has a high risk of deleting user data or customisations. For example, the user-environment is designed to be customised, so some admins might expect it to be removed as part of an “uninstall”, whereas others would expect it to be kept since it’s been modified.
The only almost-safe option is to delete just the JupyterHub application, but that’s pretty minimal compared to everything else that’s setup by TLJH.