Shut Down All Docker Containers Based on Internal Analysis – JupyterHub Example

I manage a few decent sized Jupyter Hub environments based on the docker spawner.  Each frequently has more than 50 users, sometimes much more…. and recently, one of the servers ran out of memory.

I have some read-only notebooks inside the user containers… so I figured that if a user only had those read only notebooks, I could shut down their docker containers.  They weren’t doing any work that could be lost.

So, I wrote this script to:

  1. List all docker containers.
  2. Get their names.
  3. Exec a bash command in them.
  4. Shut them down based on the result.

I hope it helps you with a similar docker-related issue! 🙂

CONTAINERS=`docker container ls | awk '{print $14}'`
  #echo $name
  COUNT=`docker exec ${NAME} ls -a | grep .ipynb | grep -v checkpoints | wc -l`
  if [[ $COUNT = 1 ]];
    echo "Stopping $NAME with COUNT = $COUNT."
    docker container stop $NAME;

No Private Variables / Methods in Python (Jupyter & iPython)!?

Having coded for a long time and in a relatively large number of languages, I was a little panicky to realize that Python doesn’t have private variables and/or methods.

Some Context

When I came across this fact, I was trying to write an iPython notebook for use by others.  You really can’t secure things in iPython as users have access to execute arbitrary code, but I thought that you could at least make it relatively hard to break by storing things in a module and loading them.  But even this doesn’t work because nothing is private in Python – a user could just interrogate my classes and get right to the connection information variable (even if they had little to no knowledge of programming).

In Java, you can technically get around private variables by cracking open classes with reflection… but non-technical people wouldn’t know that and most programmers wouldn’t bother.  The entry bar in Python is a lot lower unfortunately.

What Does Python Do Instead?

This stack overflow post says the following which helps shed some light on the situation.

It’s cultural. In Python, you don’t write to other classes’ instance or class variables. In Java, nothing prevents you from doing the same if you really want to – after all, you can always edit the source of the class itself to achieve the same effect. Python drops that pretence of security and encourages programmers to be responsible. In practice, this works very nicely.

If you want to emulate private variables for some reason, you can always use the __ prefix from PEP 8. Python mangles the names of variables like __foo so that they’re not easily visible to code outside the class that contains them (although you can get around it if you’re determined enough, just like you can get around Java’s protections if you work at it).

By the same convention, the _ prefix means stay away even if you’re not technically prevented from doing so. You don’t play around with another class’s variables that look like __foo or _bar.

So… basically, python’s style guide, PEP-8, suggests using “_xyz” for internal identifiers and “__xyz ” (with 2 underscores) for private identifiers.  Private identifiers will be name-mangled outside the module, so people won’t know what they’re using… but they’re not really private.  People can still probe around, find them, and use them if they are determined.

Again, in Java you could go use reflection to crack open a private class… so while I’m a little annoyed at this in Python, it is true that it’s not really terribly different from a real security standpoint.

Final Thoughts

So… it seems that if you want to use real secrets (like database connection details), you have to put them in a separate application on a separate server behind an API.  That way the user (especially in an iPython context) is completely decoupled from the information as a whole.

Jupyter Contributor Extensions – Auto Run + Hide Code

My Problem

I’ve been messing around with Jupyter quite a bit trying to make a nice notebook for people that are not necessarily coders.  So, it would be nice to give them a mostly graphical notebook with widgets and just let they play with the results at the end.

Unfortunately, you cannot auto-run things in Jupyter notebooks properly, and the hacks are brittle.  You also cannot hide code easily, etc.

The Solution?

Thankfully, while these features are not built into Jupyter for some reason, there are a ton of contributor extensions to Jupyter it turns out!  For example, if you need to reliably auto-run cells on start-up, you can install the init_cell plugin and the hide_input plugin.

The installation is actually very easy and can me done in a few bash commands as shown below, and there are a ton of other plugins around that you can use as well.  You can even manage the plugins from within the UI after you set them up.

pip install jupyter_contrib_nbextensions
jupyter contrib nbextension install --system
jupyter nbextension enable init_cell/main
jupyter nbextension enable hide_input/main

To use these, just read the links (or the documentation inside the UI at Edit > NB Extensions Config).  You’ll find that there is just a cell metadata attribute you can add for each cell you want to affect.  You can enable and disable the plugins in the UI as you like too.

Jupyter/Hub – Export Data to User (Not Local) PC

While building a mostly widget-based notebook for some other people, I came across a situation where I needed to allow them to export data from a pandas data frame to CSV.  This seemed trivial, but it actually was not.

What’s the Problem!?

I was building a notebook that was intended to run on Jupyter Hub… so, not on the same PC as the person using it. So, when I just saved the file, it was on the notebook server and the user could not access it.


My first thought was to have the notebook servers automatically set up a file server and just to save the files there.  Then the notebook could give users the URL to the file via the file server. I’m sure this would work, but it requires extra components and would require some clean-up of old files now and then.

While searching online, I found this solution which is much more elegant (though it will take a little extra memory). 

It base64 encodes the content and provides a link to it that way (the link actually contains all the data).  You can find the original article from medium by clicking here.  It has some other options as well.  I changed the display line and added an extra import and some altered CSV generation arguments; aside from that it is theirs.

from IPython.display import HTML
import base64

def create_download_link( df, title = "Download CSV file", filename = "data.csv"):  
    csv = df.to_csv(index=False, line_terminator='\r\n')
    b64 = base64.b64encode(csv.encode())
    payload = b64.decode()
    html = '<a download="{filename}" href="data:text/csv;base64,{payload}" target="_blank">{title}</a>'
    html = html.format(payload=payload,title=title,filename=filename)
    return HTML(html)

display(create_download_link, your_data_frame)

I hope it helps you!

What is Jupyter Hub?

First Things First… What is Jupyter?

Lately, I’ve been moving into the Python world where I quickly encountered Jupyter notebooks.  They seem like a pretty dominant technology that lets you script python block-by-block and render the results.  You can also render data into charts, manage user-interface widgets, and do most anything else.

What is the Problem With Jupyter?

But Jupyter really just runs on a single machine.  What about when you want to share this information to say, teach a class, or work with a team of data scientists?

So… We Have Jupyter Hub!

Jupyter Hub is a multi-user version of Jupyter… so it fixes our problems! Here I’ll paraphrase content and use images from a wonderful video I watched on YouTube – you can watch it at the bottom of this post if you like.

Basically, Jupyter Hub just provides a higher level service to the standard Jupyter notebooks.  It contains:

  1. A proxy server to route requests.
  2. A “hub” which handles authentication, user details, and spawning new notebooks.  Authentication is flexible and can most likely tie in your corporate authentication system.
  3. Any number of spawned Jupyter processes to run notebooks for the given users.  A variety of spawning techniques exist (e.g. spawning to Docker).

You can see this architecture below.

Image result for jupyter hub

So, if you need multi-user Jupyter, I suggest you look into installing and trying Jupyter hub, and I highly recommend the video below as a starting point!

Jupyter Auto-Run Cells on Load

Why Do We Need This?

If you are making a Jupyter notebook that heavily uses widgets and conceals the code used to make them, you’ll quickly run into an issue. Another person coming to this notebook would basically just see this message for all of your widgets:

“A Jupyter widget could not be displayed because the widget state could not be found. This could happen if the kernel storing the widget is no longer available, or if the widget state was not saved in the notebook. You may be able to create the widget by running the appropriate cells.”.

You can simulate this for yourself by pressing the “restart the kernel (with dialog)” button and then force refreshing your browser (ctrl + shift + r in chrome).

How Do We Do It?

I came across this stack-overflow post which gives a good solution (especially if you are already hiding code in other areas to make it look neater, like I noted in this blog).

Just paste this in its own cell at the top of your notebook:

        ['base/js/namespace', 'jquery'], 
        function(jupyter, $) {
            $("kernel_ready.Kernel", function () {
                console.log("Auto-running all cells-below...");

Then all your cells will run on load and all of your widgets will show up nice and neat the first time around.

Read-only / Protected Jupyter Notebooks

Jupyter notebooks are fantastic, but they’re really geared at developers.  I had to lock one down so it could be used by non-developers (without damaging it).  It took quite a lot of googling!

I figure that a lot of people must need this.  If I were a university instructor, I’d like to send students to a server, let them play, but prevent them from breaking my example.

Locking Things Down

Here are the different things I did to mitigate damage:

  1. You can actually make the entire notebook read-only by setting the file permissions on command line or in properties (works for Windows and Linux).  Jupyter will detect this (after you reload the page) and show that you can’t save anything.
  2. You can make individual cells un-deletable and un-editable (so they don’t mess up the top cells that the lower down cells they’re working with depend on):
    • Run a cell.
    • Click “Edit Metadata” in its banner.
    • Add:
      • “deletable”: false,
      • “edittable”: false,
  3. You can actually hide the code for a range of cells like this code which hides the first four (which is very useful if you’re using iPython UI widgets and just want to show widgets and now how they were made) – disclaimer – I got this off stack overflow but am having trouble finding it to reference it currently:
from IPython.display import HTML

    function code_toggle_and_hide_move_buttons() {
        if (code_show){
        else {
        code_show = !code_show



Further Recommendations

I would also suggest making sure you disable the terminal in your Jupyter config file and that you set a known location for your notebooks to be loaded from so that you can add the read-only attribute.

Also, you can disable various hotkeys in the UI, and you can use a CSS selector similar to the one in my “hide code” example above to hide the move-up/move-down cell buttons to help prevent errors cropping in that way.