Chameleon Changelog for April 2021

Dear Chameleon users,


The summer is almost in full swing, and we're frantically getting stuff done so we can kick back and just rest on our laurels from May until August… just kidding. Expect lots more fun stuff to be announced this summer!


Run your workshops on Chameleon with Jupyter. Chameleon's JupyterHub Trovi artifact makes it simple to set up a multi-user Jupyter Notebook environment on a bare metal node, allowing you to provide a user-friendly interface to Chameleon's computing capabilities. This was deployed successfully for a few workshops (such as the 2020 ANL Big Data Camp and the 2020 CSInParallel Summer Virtual Workshop) and we recently added more features to the Trovi artifact to make the workshop or classroom experience even better, namely, support for pre-configured user environments, easier onboarding of users, and the ability to provide datasets to all users. Read more about the features on Trovi and give it a spin! Keep an eye out later this month for a Tips & Tricks blog post with more details about how to run a workshop on Chameleon.


Support for part-time availability for CHI-in-a-Box associate sites. One common question that site operators ask when evaluating CHI-in-a-Box, the packaging of Chameleon that you can deploy to manage and provide your own hardware via Chameleon, is "can I block out periods of time where the hardware is allocated elsewhere?" The answer is yes, and now it's significantly easier to do this with the Doni enrollment service, which comes pre-packaged with CHI-in-a-Box. Using the Doni CLI client, you can not only enroll your hardware inventory easily, but you can also set availability windows, to indicate temporary maintenance or perhaps mark the entire site as unavailable (not reservable) for a period of time. Check out the CHI-in-a-Box wiki for more information on Doni and the capabilities it offers.


New Jupyter notebook stitching tutorial (create wide-area dedicated layer 2 links!). The Chameleon User Guide: Networking Trovi artifact continues to grow. We have added a tutorial that demonstrates layer 2 networking stitching between Chameleon sites using the Python API. Come learn how to create isolated high bandwidth network connections for your Chameleon experiments!


Improved support for sign in via federated identity. As you know, your host institution can change. Names can change. Life comes at you fast! We've worked on making some changes to the new federated single sign-on to accommodate this reality. Before, it was pretty easy to confuse the system into thinking it needed to make a new account for you, leading to you having multiple Chameleon accounts, which may sound delightful, but was a bit annoying in practice. We sent the login system to do a post-bacc year in pattern recognition and it's doing a better job of tracking down your account now.


Re-home our retired phase 1 hardware! Announcing the Chameleon Legacy Hardware Program! We're offering our retired hardware to members of the community interested in joining the Chameleon family as operators of a Chameleon Associate Site. Learn all about the Chameleon Legacy Hardware Program and apply for servers by submitting a short application. Applications are encouraged by 05/16. Application available by filling out this short Google Form. Applications encouraged by 05/16, when we plan to start initial discussions with future partners.


Seeking comment on IoT hardware investment. Chameleon is expanding to support Edge to Cloud experiments later this year. For this reason we will be purchasing a variety of IoT devices that our users will be able to provision in a manner similar to our nodes. We expect this functionality to be available in late summer/fall on a preview basis for early users only. A rough outline of our current purchase plans is here and we seek your feedback. If you would like to provide feedback on what devices to purchase and/or are interested in being an early user, please fill out this form


Upcoming changes to floating IP usage. So, turns out you all REALLY like Floating IPs. So much so that we're having trouble keeping them in stock. To make this situation a bit better, we're going to start reclaiming IPs that are not reserved in advance as part of a lease. Fun fact: you can reserve IPs, and should in fact do so! In May we'll start automatically reclaiming Floating IPs not tied to a lease with more regularity. Please refer to our documentation for instructions on how to reserve a public IP address via the CLI (or the Python API if you prefer!)

Add a comment

No comments