Recent searches
Search options
"Anyone can read #Python!"
missing_users=",".join(sorted([user for user in user_list if user not in helm_values["hub"]["config"]["Authenticator"].get("admin_users") or []])) if 'helm_values["hub"]["config"]["Authenticator"]' in local()
@vwbusguy i see you're hacking on jupyterhub :D
@yuvipanda Indeed - and trying to *not* make my test code look like this :-)
@yuvipanda Also, one of the hub endpoints we're testing returns a 202 response code. You see those about as often as a $2 bill.
@vwbusguy now very curious what you are building / testing :)
Also I made one of the other hub end points return 424, which is pretty rare too!
@yuvipanda Lots of JupyterHub environments. Some running notebooks, some scipy, some Rstudio. Retooling the deployments to replace terraform and moving a number of things to baremetal Kubernetes from GKE in the process.
@vwbusguy nice nice! Always feels good to see things I started out there in the wild :D
@yuvipanda Basically, trying to get as turn key as possible with it. Hoping to get down to the level where everything just magically happens wherever it should happen all kicked off from an Ansible AWX template, and not too far off from the reality as of this week. Things are coming together 
. And some parts of what we've done can be reused to automate/modernize other operational processes, like I finally got one of our DNS stacks fully automated.
@vwbusguy wonderful!!! This is kinda what we do for running our infra too, with github.com/2i2c-org/infrastructure.
@yuvipanda Very cool! We're using Zero 2 JupyterHub. Since I work at UC, it was a natural choice, but a good one!
https://z2jh.jupyter.org/en/stable/
Our images are also public, open source, and rebuilt/tested/published at least weekly (library list on this README is very likely out of date).
@vwbusguy wonderful!!! I helped run all the hubs at UC Berkeley for close to 6 years (https://github.com/berkeley-dsep-infra/datahub/ is what zero to jupyterhub gre wout of!). Great to see other parts of UC using it too!
@yuvipanda That's awesome! I'm doing likewise now at UCSB! We're migrating things to on-prem #Rancher RKE2 and k3s environments and between that and moving off of Terraform, our full stack will be open source at that point.
https://help.lsit.ucsb.edu/hc/en-us/categories/360005255312-Jupyter
@yuvipanda As a followup, you might find this interesting. My AlmaLinux post was published yesterday.
@yuvipanda Looks like we have a common upstream since yours is based on the z2jh chart 
@vwbusguy oh yeah, absolutely! upstream first to the extent possible :)
