jupyter_client: Runtime error raised by kernel dying in wait_for_ready

I’ve seen this error with conda builds on RTD several times. The latest was a recent build of the notebook docs.

Documenting here if not related to random kernel deaths and more relevant here (or nbconvert) than notebook issues.

Error occurs:

after converting several Sphinx documents
when preprocessing a notebook file with nbconvert
in jupyter_client wait_for_ready

Relevant lines from Traceback

nbconvert/preprocessors/execute.py", line 141, in preprocess
jupyter_client/manager.py", line 433, in start_new_kernel
kc.wait_for_ready(timeout=startup_timeout)
jupyter_client/blocking/client.py", line 59, in wait_for_ready
    raise RuntimeError('Kernel died before replying to kernel_info')
RuntimeError: Kernel died before replying to kernel_info

Full Traceback

reading sources... [ 22%] examples/Notebook/Connecting with the Qt Console

Traceback (most recent call last):
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/sphinx/cmdline.py", line 244, in main
    app.build(opts.force_all, filenames)
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/sphinx/application.py", line 267, in build
    self.builder.build_update()
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/sphinx/builders/__init__.py", line 251, in build_update
    'out of date' % len(to_build))
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/sphinx/builders/__init__.py", line 265, in build
    self.doctreedir, self.app))
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/sphinx/environment.py", line 547, in update
    self._read_serial(docnames, app)
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/sphinx/environment.py", line 567, in _read_serial
    self.read_doc(docname, app)
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/sphinx/environment.py", line 720, in read_doc
    pub.publish()
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/docutils/core.py", line 217, in publish
    self.settings)
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/sphinx/io.py", line 46, in read
    self.parse()
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/docutils/readers/__init__.py", line 78, in parse
    self.parser.parse(self.input, document)
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/nbsphinx.py", line 406, in parse
    rststring, resources = exporter.from_notebook_node(nb, resources)
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/nbsphinx.py", line 360, in from_notebook_node
    nb, resources = pp.preprocess(nb, resources)
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/nbconvert/preprocessors/execute.py", line 141, in preprocess
    cwd=path)
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/jupyter_client/manager.py", line 433, in start_new_kernel
    kc.wait_for_ready(timeout=startup_timeout)
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/jupyter_client/blocking/client.py", line 59, in wait_for_ready
    raise RuntimeError('Kernel died before replying to kernel_info')
RuntimeError: Kernel died before replying to kernel_info

Exception occurred:
  File "/home/docs/checkouts/readthedocs.org/user_builds/jupyter-notebook/conda/latest/lib/python3.5/site-packages/jupyter_client/blocking/client.py", line 59, in wait_for_ready
    raise RuntimeError('Kernel died before replying to kernel_info')
RuntimeError: Kernel died before replying to kernel_info
The full traceback has been saved in /tmp/sphinx-err-u6hd8ohp.log, if you want to report the issue to the developers.
Please also report this if it was a user error, so that a better error message can be provided next time.
A bug report can be filed in the tracker at <https://github.com/sphinx-doc/sphinx/issues>. Thanks!

About this issue

Original URL
State: open
Created 8 years ago
Comments: 30 (14 by maintainers)

Commits related to this issue

Update Dockerfile wrap jupyter nbconvert in sh -c (https://github.com/jupyter/jupyter_client/issues/154) — committed to floriangeigl/docker-DataScience by floriangeigl 5 years ago

Most upvoted comments

We were getting this error on papermill when we tried running a tests in a AWS Lambda

uvlavi on Feb 4, 2020

We were getting this error on papermill when we tried running the tests in a virtual environment using venv. The only thing that solved it was installing ipykernel then running python -m ipykernel install --user before executing a notebook

ammarasmro on Jan 30, 2020

Hi @mlucool - you’re right, this is a race condition. Ideally, the kernel would create/bind to the ports and communicate back to the launching application (jupyter_client), which is what the kernel handshaking pattern proposes. Another possibility would be to write a kernel provisioner that does what’s necessary (kernel handshake?) since the connection info now comes from there. Moving port management kernel-side seems like a good goal.

kevin-bates on Nov 11, 2021

I believe I understand this now. This code finds free ports and then writes the set of ports to the connection file so that a kernel knows where to join. This is a race condition as the ports can be used up before the kernel tries to bind them. I’m not sure of a good way to fix this.

mlucool on Nov 10, 2021

It looks like I’m still using versions older than those. I’ll update and see if I still see the issue. Thanks for the reply. On Tue, Jul 11, 2017 at 3:15 PM Carol Willing notifications@github.com wrote:

Hi @desilinguist https://github.com/desilinguist,

I thought that the latest stable versions of jupyter_client and nbconvert had resolved this issue. However, as you are still seeing intermittent issues, there may be a timing issue particularly if it is happening in the CI builds. Can you share what versions you are running? Also if you have a CI error log that would be helpful too. Thanks!

nbconvert 5.2.1 https://github.com/jupyter/nbconvert/releases/tag/5.2.1 is the latest stable version

This release did allow ExecutePreprocessor.iopub_timeout : Int to be configurable by the user. The default is 4s for timeout. You could try configuring it to be larger and see if that helps when running CI too.

jupyter_client 5.1 https://github.com/jupyter/jupyter_client/releases/tag/5.1.0 was released a few weeks ago.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/jupyter/jupyter_client/issues/154#issuecomment-314543988, or mute the thread https://github.com/notifications/unsubscribe-auth/AAJ3kP7AGL4aa_NEcOBcBZizAAuDBDMEks5sM8nNgaJpZM4IDgsm .

desilinguist on Jul 11, 2017

Hi @desilinguist,

I thought that the latest stable versions of jupyter_client and nbconvert had resolved this issue. However, as you are still seeing intermittent issues, there may be a timing issue particularly if it is happening in the CI builds. Can you share what versions you are running? Also if you have a CI error log that would be helpful too. Thanks!

nbconvert 5.2.1 is the latest stable version
- This release did allow ExecutePreprocessor.iopub_timeout : Int to be configurable by the user. The default is 4s for timeout. You could try configuring it to be larger and see if that helps when running CI too.
jupyter_client 5.1 was released a few weeks ago.

cc/ @minrk @mpacer

willingc on Jul 11, 2017

Is there any resolution for this issue yet? I also intermittently encounter this error in the CI builds for my python application that uses nbconvert.exporters.HTMLExporter to programmatically render a jupyter notebook to HTML.

desilinguist on Jul 11, 2017

@jdanbrown thank you! that worked (when I got around to testing it)

michael-erasmus on Sep 7, 2016