"Large container deployments" usecase - where are we at?


#1

On the Ubuntu Core website there is a mention of both IoT and “large container deployments” as valid usecases for Ubuntu Core, but all the documentation, and the current state of snapcraft store, seem to be heavily focused on IoT, with server usecase being put in the back seat.

Given how there are multiple alternatives in that space (RedHat’s Atomic, SuSE’s MicroOS, Rancher OS, Photon OS just to name a few) is lack of a coherent story and documentation aimed at servers a signal that Canonical is currently not interested in pursuing that angle?

To give some examples of what I’m looking for: “cloud” and bare metal images for a quick ride, documentation on how to build own images, with custom kernels and “base OS” as part of the existing CD pipeline, on-premise snaps server software with multiple subscription channels. Right now I can piece some of that informations by browsing random repositories, and sifting through discussions, but that does not give me enough confidence to keep going.


#2

There’s indeed more focus on IoT devices in the online material as the characteristics of Ubuntu Core make it a great fit for the use case. That said, Ubuntu Core is a really nice runtime environment for cloud deployments too. We’ve just not put so much emphasis on it so far, mainly because classic Ubuntu images are so popular and readily available, and people know how to handle them already.

That last point is important, and one of our concerns while introducing Ubuntu Core images. We want to make sure people understand that Ubuntu Core is not just another flavor of Ubuntu in the traditional sense, but rather a special kind of environment where all software is made available via snaps, it is transactional, and the system itself is mostly read-only.

These differences make it attractive in the first place, if you know what you’re looking for and why, but it might be frustrating for someone that expects a classic Linux system there with apt, etc. That’s why we haven’t put so much emphasis on pushing it like that so far.

We’ll get there, but we need to make sure these details are clearly explained first.


#3

It’s great that you think it is.a nice environment for cloud deployments. However, it’s hard to agree with you given how little on that is available publicly - and some of that does not paint a good picture, like Disabling automatic refresh for snap from store or External repositories

I think you are selling the community short, if you think separate “cloud” images of the core would be too confusing, but even dealing with some confusion would be, in my opinion, worth it, if it allowed more people to a look at what is being made, and be able to give you their input.

Sure, you’ll get there - but so far you’ve missed two, or even three amazing opportunities - CoreOS being bought by RedHat, CoreOS being more or less dismantled by RedHat, and RedHat being bought by IBM. I know that each of those happened, I took a look at where Ubuntu is at, and each time the answer was, “I have no idea”. There is such thing as being too late to a party ;).


#4

To be clear, this wasn’t an attempt of convincing you. In fact, my point was precisely that we haven’t been trying to convince anyone of that, because we haven’t managed to focus on it yet. There were major changes towards Ubuntu Core 18, which was released just last week, and we have an event specific to Ubuntu Core in three weeks where details like this will be discussed.

Ubuntu Core is a snap platform. We have literally millions of users looking at what’s being made, and I’m proud to say that pretty much every day we have feedback from that community, and that feedback is going back into the platform. That includes cloud environments.

If you want to engage and use Ubuntu Core images in a cloud environment and have good experience, my suggestion is to get started and engage in a more productive way by asking specific questions about issues you have while doing that. That’s healthy collaboration and we’d be happy to engage.

It’s certainly worth taking a moment to ponder why these events happened.