To install Jupyterhub inside s390x VM I followed a different approach which I am describing in this blog.
This process consists of two main steps:
- Creating ruamel.yaml.clib conda package with the help of feedstock.
- Consuming this package with other dependencies of Jupyterhub during installation.
Creating ruamel.yaml.clib conda package with the help of feedstock
S390x is a Big Endian CPU arch and ruamel.yaml.clib has an open bug related to this hence we need to apply a patch and build the ruamel.yaml.clib package with the help of feedstock.
1. Clone the repo https://github.com/manogy/ruamel.yaml.clib-feedstock.git
git clone https://github.com/manogy/ruamel.yaml.clib-feedstock.git
2. change the directory to recipe
cd ruamel.yaml.clib-feedstock/recipe
3.
Make sure you are in the base environment of the conda unless you want to use a specific conda environment.
If the conda-build package is not installed please install it with the below command.
conda install conda-build
4. To build the conda package run the below command (from the recipe folder).
conda build .
5. The above command will start the build of the package, once it completes it will generate the last lines of the log similar to the below lines,
Here /root/miniconda3/conda-bld is the channel name which conda-build created and created the conda package.
# Automatic uploading is disabled
# If you want to upload package(s) to anaconda.org later, type:
# To have conda build upload to anaconda.org automatically, use
# conda config --set anaconda_upload yes
anaconda upload \
/root/miniconda3/conda-bld/linux-s390x/ruamel.yaml.clib-0.2.7-py310ha92ec59_2.tar.bz2
anaconda_upload is not set. Not uploading wheels: []
INFO :: The inputs making up the hashes for the built packages are as follows:
{
"ruamel.yaml.clib-0.2.7-py310ha92ec59_2.tar.bz2": {
"recipe": {
"c_compiler": "gcc",
"target_platform": "linux-s390x"
}
}
}
####################################################################################
Resource usage summary:
Total time: 0:00:53.9
CPU usage: sys=0:00:00.2, user=0:00:05.3
Maximum memory usage observed: 200.2M
Total disk usage observed (not including envs): 2.7
Consuming this package with other dependencies of Jupyterhub during installation.
Previous Steps will create a conda package that we need to install during our Jupyterhub installation.
Here I am creating a conda environment with the name "jupyterhub-test".As a Prerequisite make sure Nodejs is available on the VM.
In my VM nodejs version 16 was installed. I installed some other packages as well on the base VM / conda environment.
- Install configurable-http-proxy on the VM globally by running the below command:
npm install -g configurable-http-proxy
- Create a new conda env with the desired name
conda create -n jupyterhub-test
- Activate the conda env and install the required packages
conda activate jupyterhub-test
conda install python=3.10.13 cryptography=41.0.7 notebook jupyterlab ipykernel libsodium ruamel.yaml
Install the Jupyterhub with ruamel.yaml.clib package which we created with Feedstock
- Install the ruamel.yaml.clib package like below, here /root/miniconda3/conda-bld is the channel name that we get during the build process.
conda install ruamel.yaml.clib -c /root/miniconda3/conda-bld
- Install Jupyterhub with the help of the PIP command
pip install jupyterhub==4.0.2 --no-cache-dir
Now Jupyter Hub has been installed, which can be used by the user.
### Example:
09:06:59 |jupyterhub-test|root@manogya1 ~ → uname
Linux
09:07:03 |jupyterhub-test|root@manogya1 ~ → uname -m
s390x
09:07:09 |jupyterhub-test|root@manogya1 ~ → jupyterhub
[I 2024-02-07 09:07:18.364 JupyterHub app:2859] Running JupyterHub version 4.0.2
[I 2024-02-07 09:07:18.364 JupyterHub app:2889] Using Authenticator: jupyterhub.auth.PAMAuthenticator-4.0.2
[I 2024-02-07 09:07:18.364 JupyterHub app:2889] Using Spawner: jupyterhub.spawner.LocalProcessSpawner-4.0.2
[I 2024-02-07 09:07:18.364 JupyterHub app:2889] Using Proxy: jupyterhub.proxy.ConfigurableHTTPProxy-4.0.2
[I 2024-02-07 09:07:18.369 JupyterHub app:1664] Loading cookie_secret from /root/jupyterhub_cookie_secret
[I 2024-02-07 09:07:18.423 JupyterHub proxy:556] Generating new CONFIGPROXY_AUTH_TOKEN
[I 2024-02-07 09:07:18.451 JupyterHub app:1984] Not using allowed_users. Any authenticated user will be allowed.
[I 2024-02-07 09:07:18.463 JupyterHub app:2928] Initialized 0 spawners in 0.001 seconds
[I 2024-02-07 09:07:18.467 JupyterHub metrics:278] Found 0 active users in the last ActiveUserPeriods.twenty_four_hours
[I 2024-02-07 09:07:18.468 JupyterHub metrics:278] Found 0 active users in the last ActiveUserPeriods.seven_days
[I 2024-02-07 09:07:18.468 JupyterHub metrics:278] Found 0 active users in the last ActiveUserPeriods.thirty_days
[W 2024-02-07 09:07:18.468 JupyterHub proxy:625] Found proxy pid file: /root/jupyterhub-proxy.pid
[W 2024-02-07 09:07:18.469 JupyterHub proxy:637] Proxy no longer running at pid=4381
[W 2024-02-07 09:07:18.470 JupyterHub proxy:746] Running JupyterHub without SSL. I hope there is SSL termination happening somewhere else...
[I 2024-02-07 09:07:18.470 JupyterHub proxy:750] Starting proxy @ http://:8000
09:07:18.741 [ConfigProxy] info: Proxying http://*:8000 to (no default)
09:07:18.746 [ConfigProxy] info: Proxy API at http://127.0.0.1:8001/api/routes
Other Ways to Run Jupyter Notebook on LinuxOne VM.
- Cloud Pak for Data provided Jupyter Notebooks with a much more advanced layer of security and data connectivity.
Notebooks provided by CPD on IBM Z systems and LinuxOne are tightly knit with Project, which provides many capabilities like loading of data with drag and drop
and auto-generation of code snippets to load the data. A Quick Introduction to Notebooks on CPD as a Service ( hosted on IBM cloud ) is available in a Demo video.
For detailed information refer to the Documents available on IBM's official website.
- The notebook can be run on the LinuxOne Community Cloud. Details are available at the jupyter-lab-ml Github.