Hi, I realize the authentication/access token is a security feature and that’s great.
Now say I’ve got a long-running job, that I nohupped and I come back to it later, but I accidentally forgot to copy-and-paste the access token for future reference.
…Am I SOL or is there still a way to reconnect? (I don’t see any kind of “forgot password?” link.)
For context:
- I’m running on my own personal, dedicated remote machines, for which I have superuser access. Though it’d be good to know if there are ways for non-su users to do the same (e.g. my students).
- There may be more than 1
jupyter notebook
process running on the machine. I often run two separate ones on two different GPUs, viaCUDA_VISIBLE_DEVICES=<N> jupyter notebook --no-browser --port=<port>
. So I’m not sure how to target the job I want viajupyter notebook password
(cf. forgotten Jupyter notebook password · Issue #2971 · jupyter/notebook · GitHub). If there’s a better practice for separating GPUs on kernels given just one jupyter process (…e.g., not all the other-people’s codes I run include the Pytorch specification for which GPU to use…), feel free to instruct me on that too! …But/and also more than one user might each be running their ownjupyter notebook
process because we share the machine.
Thanks folks!