Config option kernel_spec_manager_class not recognized by DisableLabExtensionsApp.

[nicornk]

Platform identifier
Amazon Linux 2, Jupyter Lab 3
(notebook-al2-v2)

setup-codeserver.sh outputs before the systemctl restart jupyter-server command is executed:

Config option `kernel_spec_manager_class` not recognized by `DisableLabExtensionsApp`.
[W 2022-11-15 07:38:31.512 LabApp] Config option `kernel_spec_manager_class` not recognized by `LabApp`.

Afterwards, jupyterlab restarts and the instance is returning a 503. When I restart the notebook it is working again, but code-server is not activated.

Here is the full log:

sh-4.2$ sudo ./setup-codeserver.sh
+ CODE_SERVER_VERSION=4.5.2
+ CODE_SERVER_INSTALL_LOC=/home/ec2-user/SageMaker/.cs
+ XDG_DATA_HOME=/home/ec2-user/SageMaker/.xdg/data
+ XDG_CONFIG_HOME=/home/ec2-user/SageMaker/.xdg/config
+ CREATE_NEW_CONDA_ENV=1
+ CONDA_ENV_LOCATION=/home/ec2-user/SageMaker/.cs/conda/envs/codeserver_py39
+ USE_CUSTOM_EXTENSION_GALLERY=0
+ EXTENSION_GALLERY_CONFIG='{{\"serviceUrl\":\"\",\"cacheUrl\":\"\",\"itemUrl\":\"\",\"controlUrl\":\"\",\"recommendationsUrl\":\"\"}}'
+ LAUNCHER_ENTRY_TITLE='Code Server'
+ PROXY_PATH=codeserver
+ LAB_3_EXTENSION_DOWNLOAD_URL=https://github.com/aws-samples/amazon-sagemaker-codeserver/releases/download/v0.1.5/sagemaker-jproxy-launcher-ext-0.1.3.tar.gz
+ INSTALL_LAB1_EXTENSION=1
+ LAB_1_EXTENSION_DOWNLOAD_URL=https://github.com/aws-samples/amazon-sagemaker-codeserver/releases/download/v0.1.5/amzn-sagemaker-jproxy-launcher-ext-jl1-0.1.4.tgz
+ export XDG_DATA_HOME=/home/ec2-user/SageMaker/.xdg/data
+ XDG_DATA_HOME=/home/ec2-user/SageMaker/.xdg/data
+ export XDG_CONFIG_HOME=/home/ec2-user/SageMaker/.xdg/config
+ XDG_CONFIG_HOME=/home/ec2-user/SageMaker/.xdg/config
+ export PATH=/home/ec2-user/SageMaker/.cs/bin/:/sbin:/bin:/usr/sbin:/usr/bin
+ PATH=/home/ec2-user/SageMaker/.cs/bin/:/sbin:/bin:/usr/sbin:/usr/bin
+ EXT_GALLERY_JSON=
+ '[' 0 -eq 1 ']'
+ JUPYTER_CONFIG_FILE=/home/ec2-user/.jupyter/jupyter_notebook_config.py
+ grep -q /home/ec2-user/SageMaker/.cs/bin /home/ec2-user/.jupyter/jupyter_notebook_config.py
+ cat
++ /home/ec2-user/anaconda3/envs/JupyterSystemEnv/bin/jupyter-lab --version
+ JUPYTER_LAB_VERSION=3.5.0
+ sudo -u ec2-user -i
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
100 19651  100 19651    0     0  25306      0 --:--:-- --:--:-- --:--:--  223k
Looking in indexes: https://pypi.org/simple, https://pip.repos.neuron.amazonaws.com
Processing ./SageMaker/.cs/lab_ext/sagemaker-jproxy-launcher-ext.tar.gz
  Installing build dependencies ... done
  Getting requirements to build wheel ... done
  Preparing metadata (pyproject.toml) ... done
Building wheels for collected packages: sagemaker-jproxy-launcher-ext
  Building wheel for sagemaker-jproxy-launcher-ext (pyproject.toml) ... done
  Created wheel for sagemaker-jproxy-launcher-ext: filename=sagemaker_jproxy_launcher_ext-0.1.3-py3-none-any.whl size=28044 sha256=07c6e1b86f57d64e26b94a21c11d9871db08f9263b51d97af802f31c13ac7705
  Stored in directory: /home/ec2-user/.cache/pip/wheels/b1/ae/23/e402630d56afaedbf7083d851b13612dbb641f84efdc4523b5
Successfully built sagemaker-jproxy-launcher-ext
Installing collected packages: sagemaker-jproxy-launcher-ext
Successfully installed sagemaker-jproxy-launcher-ext-0.1.3
Config option `kernel_spec_manager_class` not recognized by `DisableLabExtensionsApp`.
[W 2022-11-15 07:38:31.512 LabApp] Config option `kernel_spec_manager_class` not recognized by `LabApp`.
+ systemctl restart jupyter-server

[giuseppeporcelli]

Hi @nicornk thanks for reaching out.
The kernel_spec_manager_class one is a warning that does not affect the install procedure.

After the Jupyter server restarts, you might have to wait a few seconds before reloading the browser window.

If you stop and restart the notebook, the setup_codeserver.sh script should be executed again, as documented.

Hope this helps.

[nicornk]

Good to know that the warning is expected.

however, my jupyter server seems to not start up again - the 503 continues until I manually stop and start it from the Sagemaker console.