jupyter-zfssa-analysis

Jupyter Notebook to analyse my zfssa (ZFS Storage Appliance).

This notebooks needs the output from my zfs script explorer:

https://github.com/aldenso/zfssa-scripts/tree/master/explorer

After generating the explorer you can adjust the path in the notebooks to your unzipped explorers (default path: "/tmp/datazfssa").

Using this repo with pip

Create a new virtualenv and activate it.

python -m venv jupytervenv
source jupytervenv/bin/activate

Clone the repo, change to the new directory and install the dependencies with the requirements file.

git clone https://github.com/aldenso/jupyter-zfssa-analysis
cd jupyter-zfssa-analysis
pip install -r requirements.txt

Enable the widgets.

jupyter nbextension enable --py widgetsnbextension --sys-prefix # use --sys-prefix when using virtualenvs

Now you can launch your jupyter notebook.

jupyter-notebook

Using this repo with pipenv

If you don't have pipenv already installed, then proceed to install it with pip.

sudo pip install pipenv

With pipenv just clone the repo and change to the new directory, then run the install option.

git clone https://github.com/aldenso/jupyter-zfssa-analysis
cd jupyter-zfssa-analysis
pipenv install

Once all dependencies are installed then activate the virtualenv.

pipenv shell

Enable the widgets.

jupyter nbextension enable --py widgetsnbextension --sys-prefix # use --sys-prefix when using virtualenvs

Now you can launch your jupyter notebook.

jupyter-notebook

Use this notebooks with a docker container if you prefer it

• step 1:

Clone this repo and change your working directory to it.

git clone https://github.com/aldenso/jupyter-zfssa-analysis
cd jupyter-zfssa-analysis

• step 2:

With Ipython, create a hashed password (if you want to change the password).

In [1]: from notebook.auth import passwd
In [2]: passwd('password')
Out[2]: 'sha1:5b6fcd2f3bcc:acd703754a0dfaff49853f3233f395f50838ea7d'

• step 3:

If you decided to set a new password then change the hash in the Dockerfile.

# hashed password generated previously (check README)
ENV PASSWORD="'sha1:5b6fcd2f3bcc:acd703754a0dfaff49853f3233f395f50838ea7d'"

• step 4:

Build the image (it may take awhile) and validate if the new image is available.

$ sudo docker build -t yourname/archjupyter .

$ sudo docker images
REPOSITORY                 TAG                 IMAGE ID            CREATED             SIZE
aldenso/archjupyter        latest              95c1aa739366        About an hour ago   1.21GB
.
archlinux/base             latest              71415bf73a2b        3 weeks ago         394MB
.

• step 5:

Run a container and put some unzipped zfsssa explorer in the path shared to the container.

sudo docker run -d -v /tmp/datazfssa:/tmp/datazfssa -p 8888:8888 yourname/archjupyter

Now you can open your browser and test the notebooks (Remember to adjust the paths for the csv files).

http://youripaddress:8888

Auto-save .html and .py versions of the notebooks

For convenience we are adding auto-saving .py (not .html, but we add the commented line for it) for the notebooks, making a more readable output for code review.

You just have to add the next lines of code to your profile (~/.jupyter/jupyter_notebook_config.py).

import os
from subprocess import check_call

def post_save(model, os_path, contents_manager):
    """post-save hook for converting notebooks to .py scripts"""
    if model['type'] != 'notebook':
        return # only do this for notebooks
    d, fname = os.path.split(os_path)
    check_call(['jupyter', 'nbconvert', '--to', 'script', fname], cwd=d)
    #check_call(['jupyter', 'nbconvert', '--to', 'html', fname], cwd=d)

c.FileContentsManager.post_save_hook = post_save

If you don't have a config file already, you can create one.

jupyter notebook --generate-config