Jupyter hub 403 : Forbidden

I am trying to make my way through the jupyterhub notebook tutorials one can link to from the documentation. However, if / when I stop and then want to come back, I get this 403:Forbidden error over and over and over again.

The only thing I’ve found that sort of looks like a solution is here: python 3.x - Jupyterhub service unavailable error and http :403 forbidden - Stack Overflow

The idea is that I have an expired API token, which causes me to get the 403. If so, why not give those an automatic - and much shorter - time out / self destruct switch / whatever?

The problem with this solution is that the pid seems to change at the blink of an eye:

[email protected]:~$ ps aux | grep configurable-http-proxy
malikar+    6836  0.0  0.0   9040   712 pts/0    S+   12:38   0:00 grep --color=auto configurable-http-proxy
[email protected]:~$ ps aux | grep configurable-http-proxy |awk '{print "kill -9 " $2}'|sh
[email protected]:~$ ps aux | grep configurable-http-proxy
malikar+    7071  0.0  0.0   9040   652 pts/0    S+   12:42   0:00 grep --color=auto configurable-http-proxy
[email protected]:~$ ps aux | grep configurable-http-proxy kill
grep: kill: No such file or directory
[email protected]:~$ ps aux | grep configurable-http-proxy | kill
kill: usage: kill [-s sigspec | -n signum | -sigspec] pid | jobspec ... or kill -l [sigspec]
[email protected]:~$ 
[email protected]:~$ kill configurable-http-proxy
bash: kill: configurable-http-proxy: arguments must be process or job IDs
[email protected]:~$ kill 7071
bash: kill: (7071) - No such process

I even switched from Chrome to Firefox, but I still got the 403. Despite showing up in the terminal, none of these pid have ever shown up in my system monitor.

Does anyone have a solution, please? Thank you.

@malikarumi These are really Binder / Jupyterhub issues. We don’t control anything about their session timeouts or API tokens. You would probably be better served asking these questions on some support forum for those projects.

I guess I would also ask: do you actually need JupyterHub, specifically? JupyterHub is a large project for supporting distributed, multi-tenant notebook environments for entire organizations. If you don’t need all that, then using it will add a huge amount of complication for no reason.

If you just want to run notebooks locally then all you would need is classic Jupyter Notebook (or newer JupyterLab). These (not JupyterHub) are the appropriate choice for the vast majority of regular users.

xref: error without message · Issue #98 · bokeh/bokeh-notebooks · GitHub

1 Like

Yea, I was able to download all the tutorial notebooks and I am working through them at home this weekend. I used jupyterhub because that’s where the online tutorial took me. No other reason.

The linked version is on JupyerHub because Binder is a public service for serving notebooks to tens of thousands of users at once on the open internet. If you are just running locally for yourself you should use JupyterLab:

1 Like