Hacker News Reader

Comment by onezerozeroone | Hacker News Reader

Wouldn't the point of running in Docker still be to guarantee version, environment, and config compatibility? Just because you run an Elixir app on your local machine doesn't mean it will work worry-free on Prod. Without Docker, you might have a different version of BEAM installed (or none at all), etc.

View on HN

Isn't that the purpose of tools like https://erlang.mk/guide/getting_started.html#_getting_starte... ?

Edit: and for cross-system dev (e.g. debian to windows) -> https://erlang.mk/guide/cross_compiling.html

Correct, but you still need to compile the release on a similar system.

For example if your dev machine is Debian it won't work if the deploy machine is Alpine (similarly Windows builds won't work on Linux or Mac).

I think OTP Releases would still depend on the target host being configured correctly and having all dependencies installed that your application needs to run.

Containers are all about consistency of the environment. For example, maybe your Elixir app wants to log out to something like Splunk or DataDog...if those agents aren't installed on the host, it doesn't matter what facilities BEAM provides. Or maybe your app depends on certain configs, libraries, or files being available at specific paths on the system. Things like that would be captured in the image your containers are running and guarantee consistency across machines and environments.

Right, they do, that is why containers are still useful to releases for deployment.

My bad, I sould have specified, I was only addressing "Without Docker, you might have a different version of BEAM installed (or none at all)"

Ah, makes sense!

you still need to compile your system on the same combination of OS/libc

libc builds don't work on Alpine

This is a question that I see hotly debated a lot in Elixir circles. BEAM and the Elixir toolset do have a lot of tools that overlap with Docker, so there's a lot of people that advocate for "not bothering" with Docker, while there's other people that are familiar with Docker (and perhaps not as familiar with BEAM) that will tend to use Docker. They both have more or less the same capability to deliver the same functionality.

IMO, in most use cases it makes sense to just use whichever you're more familiar with and will be more productive with. After all, that is the entire point of said tools...

They really do not overlap at all. I don't understand why the comparison with docker and Kubernetes continues.

Running BEAM on docker in kubernetes is great, a lot is added and operations smoothed by doing so.

Elixir Releases/Distillery definitely do overlap with Docker deployment techniques. The BEAM also has scaling capabilities that are tantamount to K8s horizontal scaling.

Deployment doesn't overlap. Erlang releases bundle up your code but you still need libraries like openssl, to have a nice bundle that can actually just run anywhere Docker is the way to go.

And horizontal scaling between the two do not overlap either, not sure what you mean. Packing containers (processes) efficiently into a cluster of nodes is not something Erlang does.

Edit: Erlang/OTP does offer fail over for what it calls "distributed applications" but based on a static set of nodes -- not horizontal scaling, anymore than other languages/frameworks do by letting you spawn new instances...

>Deployment doesn't overlap. Erlang releases bundle up your code but you still need libraries like openssl, to have a nice bundle that can actually just run anywhere Docker is the way to go.

And with Docker you still need libraries like, y'know Docker for your code to run.

>Packing containers (processes) efficiently into a cluster of nodes is not something Erlang does.

How much have you used BEAM? It definitely does do that. Scalability is one of Erlang/Elixir's primary benefits, and node clustering is exactly how it does that.

> How much have you used BEAM?

Over a decade professionally.

The closest thing Erlang has to it is pool http://erlang.org/doc/man/pool.html

BEAM doesn't do everything containers or a container orchestration system can, but in view of my relatively short time as a regular Erlang programmer (definitely less than 10 years), the spirit of tidepod's point accords with my experience (with a weaker interpretation of his term "node clustering"). BEAM and OTP don't do automatic load distribution between nodes, but can be fashioned to do so easily with application support.

We never used pool. The nodes were mapped onto heterogenous machines sharing the host with a 3rd-party daemon. It's configuration changes even took place through a module update hook written in Erlang itself. We both deployed new code and distributed work "manually" across them entirely on OTP.

[NOTE] It it surprising, or was to me, that there are problems with having a fairly small number of nodes fully connected. I'm lucky enough to have avoided learning this the hard way, but imagine this could serve as a painful backbone to an "Erlang deployment war story".

Sure, you can write it yourself, and probably it is even easier to write in Erlang -- up to a limit due to the issues with distributed Erlang discussed above.

I worry about, and have seen this both in the first hype phase of Erlang a number of years ago, the misconception about what Erlang offers and the resulting frustration, blaming it on the tool and quitting.

Hoping in the chapters I'm working on for https://adoptingerlang.org/docs/production/ I can better explain the benefits of running on k8s (or similar), while also making clear it certainly isn't the right choice in all cases.

- Tristan

Nice synopsis in the docs. I’d suggest that instead of saying k8s and OTP don’t overlap that it’s more that "they don’t completely overlap" Fenn diagram style. That’d be a bit more nuanced answer and help people judge which category they fall into.

There are many cases where having an Erlang cluster and doing ‘naive pooling’ at the application level would get you 80% of what k8s + routing layer would get you. I’m assuming more of an an all Erlang/Elixir environment which many small startups could get by with. Even then Docker + Fargate or whatnot would still simplify deployment. Personally I harbor a secret desire to see if I could replicate part of the k8s interface using Nerves images + some otp tooling. Probably better things to do with my life though. ;)

Also by ‘heart’ are you referring to how ‘epmd’ works or Erlang distribution?

For 'heart' here I'm referring to the heart program that'll start if you run erl with `-heart` http://erlang.org/doc/man/heart.html -- it will restart a down node.

I think part of the confusion is you in general don't need k8s, but when your size and requirements get to a point that k8s makes sense (which arguably lowers as hosted k8s becomes better) it is not in conflict with your also use of Erlang/Elixir.

I find distributed Erlang much nicer in k8s env (I don't have to maintain it obviously) where I get an IP per node, can add a k8s service making it possible to use DNS SRV queries to find all other nodes and letting k8s worry about where pods run and keeping them up.

Plus there is configuration management and consistent storage (resources in etcd).

BEAM tools are kubernetes done right for a single environment.

I say done right because kubernetes doesn't allow hot code reload without restarts.

Reply on news.ycombinator.com