zirias (on snac) on Nostr: I think terms like "modern features" and "modern demands" belong on your typical ...
I think terms like "modern features" and "modern demands" belong on your typical bullshit-bingo card.
By that, I don't mean to completely disagree with your writing as a whole. A thing like live migration of virtual machines is a very valid requirement, for example, it's immensely useful in operations.
Looking at this whole "containerization" topic, first of all, jails are not only one of the oldest solutions, but also one of the best in terms of reliability and security. If there's anything missing, it's tooling around that. Of course, you can profit a lot from features like standardized mechanisms to easily integrate creation and deployment of a container with some CI/CD pipelines, like running a container in some "cluster" with fully automated migrations, load management, and so on. That's all still assuming you build, deploy and operate your very own containers (whether running your own or third-party software or a mix of both), which is IMHO the only sane way to do things.
But then, you mention just running pre-built "images", IOW, the "docker style". I'd say that's a potentially dangerous hype. In a nutshell, it's a mechanism to deliver utter crap with lots of horrible duct-tape inside in a way the user installing/deploying that doesn't get aware of it. Among other things (like wasted resources), this has security implications. There are already tools helping you to "scan" your containers, so you have at least a chance to identify risks and vulnerabilities (another great way to make money).
That said, I certainly don't mind when FreeBSD can host and run a Linux container. But I'd never consider that a crucial feature either. FreeBSD should make sure to keep changing carefully and thoughtfully, which includes NOT to jump on every hype. The argument that you need to convince people to run FreeBSD is partially valid, the project certainly needs some relevance to attract contributors and donators and ultimately "stay alive". But that wouldn't help if it would sacrifice some core strengths that are important to its current users, among them its evolutionary, no-bullshit, no-hype approach to change.
#FreeBSD #development #features #containers
By that, I don't mean to completely disagree with your writing as a whole. A thing like live migration of virtual machines is a very valid requirement, for example, it's immensely useful in operations.
Looking at this whole "containerization" topic, first of all, jails are not only one of the oldest solutions, but also one of the best in terms of reliability and security. If there's anything missing, it's tooling around that. Of course, you can profit a lot from features like standardized mechanisms to easily integrate creation and deployment of a container with some CI/CD pipelines, like running a container in some "cluster" with fully automated migrations, load management, and so on. That's all still assuming you build, deploy and operate your very own containers (whether running your own or third-party software or a mix of both), which is IMHO the only sane way to do things.
But then, you mention just running pre-built "images", IOW, the "docker style". I'd say that's a potentially dangerous hype. In a nutshell, it's a mechanism to deliver utter crap with lots of horrible duct-tape inside in a way the user installing/deploying that doesn't get aware of it. Among other things (like wasted resources), this has security implications. There are already tools helping you to "scan" your containers, so you have at least a chance to identify risks and vulnerabilities (another great way to make money).
That said, I certainly don't mind when FreeBSD can host and run a Linux container. But I'd never consider that a crucial feature either. FreeBSD should make sure to keep changing carefully and thoughtfully, which includes NOT to jump on every hype. The argument that you need to convince people to run FreeBSD is partially valid, the project certainly needs some relevance to attract contributors and donators and ultimately "stay alive". But that wouldn't help if it would sacrifice some core strengths that are important to its current users, among them its evolutionary, no-bullshit, no-hype approach to change.
#FreeBSD #development #features #containers