SAS Viya for Learners 4 + Jupyter: Where did my files go?!
File this post in the “Who Moved My Cheese?” section:
Global Academic Programs is very pleased to announce the release of our latest version of software for teaching and learning: SAS Viya for Learners 4. Yay! You can learn why I’m excited about this new software for academics, here.
In this SAS Communities Library Article, I’ll share more details on how the Jupyter environment is configured in SAS Viya for Learners 4, with the ultimate eye on three questions:
- I logged out of Jupyter and lost all my files… where did they go?
- Why can’t I permanently save files and data to Jupyter in SAS Viya for Learners 4?
- What’s the work around?
Let’s get started on demystifying the environment a bit.
Part 1: I logged out of Jupyter and lost all my files… where did they go?
The simple answer is that Jupyter access in SAS Viya for Learners is an impermanent environment. Now, to me, the word “impermanent” makes me get all philosophical and think about the meaning of life. Because nothing is permanent, right? And I guess that’s somewhat accurate here.
Impermanence in the SAS Viya for Learners + Jupyter context means that the Jupyter environment will reset itself after every session ends. And you’ll start from scratch at this home screen each time you log back in:
- Yes, this will make some users sad – because your files are gone after the session ends. But if I get the word out now, I’m hoping that I can minimize the sadness with knowledge.
Part 2: Why can’t I permanently save files and data to Jupyter in SAS Viya for Learners 4?
I’ll start this section with a question that you’re already thinking. Why on earth did we deploy an impermanent environment?
Well, the short answer is twofold: (1) Kubernetes and (2) security.
For the former, SAS Viya 4 is deployed using Kubernetes. Utilizing the Kubernetes platform with SAS Viya 4 ensures that the product is (1) highly flexible and scalable, (2) highly reliable, and (3) more resource efficient. To end users, this simply means that your analytics experience should be more consistent and that the data intensive processes should run fast, because the system is able to handle a much larger volume of usage.
Now a few more details on architecture. In deploying SAS Viya 4 using Kubernetes and our Engage software for fleet management, we also had to consider the security of the system – meaning either open or closed to all users within Jupyter. The challenge here was that there wasn’t a technical way to partition the folders so that users could only see their content (at least not yet). Other deployments can be configured to partition the folders (think: commercial enterprise), but we couldn’t do that in SAS Viya for Learners 4, given existing out-of-the-box R&D tools. So, to adhere to privacy standards, like GDPR, we need to lock down the user folder so everyone wouldn’t see everyone else’s email address.
Where am I going with that last paragraph?
Well, this is all building to our solution: an impermanent environment that lasts just for the session. Where that session is localized to just the user. Also, technically everyone uses the same generic jovyan account in that local session.
What is a jovyan? Well, let’s hear from Jupyter directly:
Much like the planet Jupiter and our solar system, the Jupyter community is large, distributed, and nebulous. We like to use the word Jovyan to describe members of this community. Jovyans are fellow open enthusiasts that use, develop, promote, teach, learn, and otherwise enjoy tools in Jupyter’s orbit.
So, you lose your data because Jupyter in SAS Viya for Learners 4 is an impermanent environment where everyone uses the same jovyan account in a session that is localized to them. And that’s just the way we need to configure the SAS Viya for Learners 4 environment – for now – to support thousands of worldwide users and keep their data safe. And to keep the system performant and top-notched.
Part 3: What’s the work around?
GitHub. GitHub. GitHub.
Confession: I’m a huge fan of GitHub! The solution here is to simply become best friends with the industry standard for developer collaboration: GitHub. Integration becomes second nature – and I dare say it, reflexive – after you learn the process. And if you’re not quite there, please see my previous SAS Communities Article, aptly titled GitHub + Jupyter Integration in SAS Viya for Learners 4 (VFL4).
