Hi @AshwinKAshok - Since Kernel Gateway only launches local kernels (to itself) this answer would be similar to Notebook relating kernel_ids to processes since JKG is essentially the same implementation (just headless, except that the kernel management takes place on the JKG server).
I don’t know of a way to do this with Notebook and nowhere in the framework is the kernel-pid surfaced AFAIK. Do you know of a way to accomplish this using standard Notebook?
Of course, kernel management knows this relationship (jupyter_client in particular), but the pid is rarely used internally (the Popen instance is used for signals and polling).
/api/kernels and /api/kernelspecs endpoints.)I guess the answer is the same as what I responded with on this issue: jupyter/kernel_gateway#336.
Had the metrics of NBresuse been implemented to extend
/api/kernels/{kernel-id} and the gathering been done in MappingKernelManager where activity recording occurs, this would just work because that api is handled by gateway servers (JKG and EG).I also see nbreuse using the kernel process’s command line to make the kernel-id association based on the fact that the kernel-id happens to be present in the connection-file name. This would not be necessary if also associated in
MappingKernelManager - which has direct access (via the KernelManager instance) to the kernel process.So, there are some things we can do, not so much in notebook, but perhaps in jupyter_server and I’d be happy to help (time permitting).
@/all - [ANN] Notebook 6.1.0 is available!
The Notebook 6.1.0 release is now available! You can catch up on the changes here.
This release culminates in a substantial community effort with nearly 60 different contributors. Thank you to all those that contributed, raised issues, and participated in its success!
To get 6.1.0:
For pip-managed installations:
new: pip install notebook
existing: pip install --upgrade notebook
For conda-managed installations:
new: conda install -c conda-forge notebook
existing: conda update -c conda-forge notebook
https://github.com/jupyter/notebook/blob/9cc562d33922a7996975d04a75074298b662793a/notebook/static/notebook/js/notebook.js#L3265-L3287
So for the past few weeks, I have been building a Jupyter Notebook extension that creates a Pyspark cluster locally or on Kubernetes on the click on a few buttons.
But I am facing an issue.
I have made a modal, and inside the modal we have a few input boxe. But when I click on the input boxe and try to type in values the focus of the keyboard does not change. Whatever I try to type in the input boxe inside the modal, seems to go to the Jupyter Notebook Cells.
I am also attaching a screencast that will help you understand the issue: https://www.loom.com/share/342abfc3c34143b39bf9d98462cd9548
IPython.display.HTML(...), so I am thinking maybe we can also support that in Jupyter notebook by detecting quoted string inside HTML() in a code block. Could you please give me some hints?
Hi, I am having a very weird problem trying to run Jupyter notebook from a server via SSH port forwarding.
- Notebook hangs after I try to open, create or copy a notebook using the Jupyter interface.
- Once Jupyter starts hanging, it stops responding to Ctrl-C on the server
- Opening, creating or copying text files on the same Jupyter interface works, so it doesn't seem to be a file permission issue.
- Renaming notebooks using the Jupyter interface works, so it doesn't seem to be a file permission issue.
- The issue isn't with my browser, since Jupyter works from another server as well as when run from my laptop
- I've tried deleting my browser's cache, using different browsers as well as incognito without success, so doesn't seem to be a browser issue
- The problem persists even when I run Jupyter from a fresh Conda installation, so doesn't seem to be a Jupyter version issue. I've also verified that the exact same version of Jupyter notebook works when I run it from the other server
- I've tried a range of different SSH ports. I am always able to connect to the main Jupyter interface, and Jupyter starts hanging only when I try to open, create or copy a notebook
- I get the same problem whether I try to create an R, Python or Bash notebook, so doesn't seem to relate to the Kernel used
Anybody with any idea on what I could try next would be really great...
Have you tried operations across the tunnel but unrelated to Jupyter? i.e., try troubleshooting at a lower level.
ssh -v -v -v:debug2: channel 2: read<=0 rfd 9 len 0
debug2: channel 2: read failed
debug2: channel 2: chan_shutdown_read (i0 o0 sock 9 wfd 9 efd -1 [closed])
debug2: channel 2: input open -> drain
debug2: channel 3: read<=0 rfd 10 len 0
debug2: channel 3: read failed
debug2: channel 3: chan_shutdown_read (i0 o0 sock 10 wfd 10 efd -1 [closed])
debug2: channel 3: input open -> drain
debug2: channel 2: ibuf empty
debug2: channel 2: send eof
debug3: send packet: type 96
debug2: channel 2: input drain -> closed
debug2: channel 3: ibuf empty
debug2: channel 3: send eof
debug3: send packet: type 96
debug2: channel 3: input drain -> closed
...
jupyter notebook I'm getting a KeyError on 'color', has anyone seen this?--- Logging error ---
Traceback (most recent call last):
File "/usr/lib/python3.6/logging/__init__.py", line 994, in emit
msg = self.format(record)
File "/usr/lib/python3.6/logging/__init__.py", line 840, in format
return fmt.format(record)
File "/home/osboxes/jupyter/notebook/.tox/lib/python3.6/site-packages/traitlets/config/application.py", line 117, in format
return super(LevelFormatter, self).format(record)
File "/usr/lib/python3.6/logging/__init__.py", line 580, in format
s = self.formatMessage(record)
File "/usr/lib/python3.6/logging/__init__.py", line 549, in formatMessage
return self._style.format(record)
File "/usr/lib/python3.6/logging/__init__.py", line 391, in format
return self._fmt % record.__dict__
KeyError: 'color'
8 replies
Hi, I'm having difficulty switching conda environments while running a jupyter notebook. I first created a 2nd env, installed ipykernel in that environment, and created a 2nd kernel.
Shut down all kernels, open my notebook and choose the 2nd kernel (env2), but running the notebook, i can see that the old the commands are being run under the previous env, as demonstrated with a !conda env list.
as a test, i've ran it as a list. It seems like every subsequent command is run from a default shell.
!conda env list!source /opt/conda/etc/profile.d/conda.sh && conda activate testy && conda env list!conda env list
will show:
conda environments:
#
base * /opt/conda
testy /opt/conda/envs/testy
conda environments:
#
base /opt/conda
testy * /opt/conda/envs/testy
conda environments:
#
base * /opt/conda
testy /opt/conda/envs/testy
Preparing transaction: done
Verifying transaction: done
Executing transaction: failed
ERROR conda.core.link:_execute(502): An error occurred while uninstalling package 'conda-forge::notebook-4.2.3-py35_0'.
OSError(16, 'Device or resource busy')
Attempting to roll back.
Rolling back transaction: done
OSError(16, 'Device or resource busy')